Star CCM+ User Guide

CCM USER GUIDE STAR-CD VERSION 4.02

CONFIDENTIAL — FOR AUTHORISED USERS ONLY

© 2006 CD-adapco

TABLE OF CONTENTS OVERVIEW 1

2

COMPUTATIONAL ANALYSIS PRINCIPLES Introduction ............................................................................................................... 1-1 The Basic Modelling Process .................................................................................... 1-1 Spatial description and volume discretisation ........................................................... 1-2 Solution domain definition .............................................................................. 1-3 Mesh definition ................................................................................................ 1-4 Mesh distortion ................................................................................................ 1-5 Mesh distribution and density ......................................................................... 1-6 Mesh distribution near walls ........................................................................... 1-7 Moving mesh features ..................................................................................... 1-8 Problem characterisation and material property definition ....................................... 1-8 Nature of the flow ............................................................................................ 1-9 Physical properties ........................................................................................... 1-9 Force fields and energy sources ...................................................................... 1-9 Initial conditions ............................................................................................ 1-10 Boundary description .............................................................................................. 1-10 Boundary location ......................................................................................... 1-11 Boundary conditions ...................................................................................... 1-11 Numerical solution control ..................................................................................... 1-13 Selection of solution procedure ..................................................................... 1-13 Transient flow calculations with PISO .......................................................... 1-13 Steady-state flow calculations with PISO ..................................................... 1-15 Steady-state flow calculations with SIMPLE ................................................ 1-16 Transient flow calculations with SIMPLE .................................................... 1-17 Effect of round-off errors .............................................................................. 1-18 Choice of the linear equation solver .............................................................. 1-19 Monitoring the calculations .................................................................................... 1-19 Model evaluation .................................................................................................... 1-20 BASIC STAR-CD FEATURES Introduction ............................................................................................................... 2-1 Running a STAR-CD Analysis ................................................................................. 2-2 Using the script-based procedure .................................................................... 2-3 Using STAR-Launch ....................................................................................... 2-8 pro-STAR Initialisation .......................................................................................... 2-12 Input/output window ..................................................................................... 2-13 Main window ................................................................................................. 2-15

Version 4.02

i

3

The menu bar .................................................................................................2-16 General Housekeeping and Session Control ...........................................................2-18 Basic set-up ....................................................................................................2-18 Screen display control ....................................................................................2-18 Error messages ...............................................................................................2-19 Error recovery ................................................................................................2-20 Session termination ........................................................................................2-21 Set Manipulation .....................................................................................................2-21 Table Manipulation .................................................................................................2-24 Basic functionality .........................................................................................2-24 The table editor ..............................................................................................2-26 Useful points ..................................................................................................2-31 Plotting Functions ....................................................................................................2-31 Basic set-up ....................................................................................................2-31 Advanced screen control ................................................................................2-32 Screen capture ................................................................................................2-33 The Users Tool ........................................................................................................2-35 Getting On-line Help ...............................................................................................2-35 The STAR GUIde Environment ..............................................................................2-38 Panel navigation system .................................................................................2-40 STAR GUIde usage .......................................................................................2-41 General Guidelines ..................................................................................................2-41 MATERIAL PROPERTY AND PROBLEM CHARACTERISATION Introduction ...............................................................................................................3-1 The Cell Table ...........................................................................................................3-1 Cell indexing ....................................................................................................3-3 Multi-Domain Property Setting .................................................................................3-5 Setting up models .............................................................................................3-6 Compressible Flow ....................................................................................................3-9 Setting up compressible flow models ..............................................................3-9 Useful points on compressible flow ...............................................................3-10 Non-Newtonian Flow ..............................................................................................3-11 Setting up non-Newtonian models .................................................................3-11 Useful points on non-Newtonian flow ...........................................................3-11 Turbulence Modelling .............................................................................................3-12 Wall functions ................................................................................................3-13 Two-layer models ..........................................................................................3-13 Low Re models ..............................................................................................3-14 Hybrid wall boundary condition ....................................................................3-14

ii

Version 4.02

4

Reynolds Stress models ................................................................................. 3-15 DES models ................................................................................................... 3-15 LES models ................................................................................................... 3-15 Changing the turbulence model in use .......................................................... 3-16 Heat Transfer In Solid-Fluid Systems ..................................................................... 3-16 Setting up solid-fluid heat transfer models .................................................... 3-17 Heat transfer in baffles .................................................................................. 3-18 Useful points on solid-fluid heat transfer ...................................................... 3-19 Buoyancy-driven Flows and Natural Convection ................................................... 3-20 Setting up buoyancy-driven models .............................................................. 3-20 Useful points on buoyancy-driven flow ........................................................ 3-20 Fluid Injection ......................................................................................................... 3-21 Setting up fluid injection models ................................................................... 3-22 BOUNDARY AND INITIAL CONDITIONS Introduction ............................................................................................................... 4-1 Boundary Location .................................................................................................... 4-1 Command-driven facilities .............................................................................. 4-2 Boundary set selection facilities ...................................................................... 4-3 Boundary listing .............................................................................................. 4-3 Boundary Region Definition ..................................................................................... 4-5 Inlet Boundaries ........................................................................................................ 4-9 Introduction ..................................................................................................... 4-9 Useful points .................................................................................................. 4-10 Outlet Boundaries ................................................................................................... 4-11 Introduction ................................................................................................... 4-11 Useful points .................................................................................................. 4-12 Pressure Boundaries ................................................................................................ 4-12 Introduction ................................................................................................... 4-12 Useful points .................................................................................................. 4-13 Stagnation Boundaries ............................................................................................ 4-14 Introduction ................................................................................................... 4-14 Useful points .................................................................................................. 4-15 Non-reflective Pressure and Stagnation Boundaries ............................................... 4-16 Introduction ................................................................................................... 4-16 Useful points .................................................................................................. 4-18 Wall Boundaries ...................................................................................................... 4-19 Introduction ................................................................................................... 4-19 Thermal radiation properties ......................................................................... 4-20 Solar radiation properties .............................................................................. 4-20

Version 4.02

iii

5

iv

Other radiation modelling considerations ......................................................4-21 Useful points ..................................................................................................4-22 Baffle Boundaries ....................................................................................................4-23 Introduction ....................................................................................................4-23 Setting up models ...........................................................................................4-24 Thermal radiation properties ..........................................................................4-25 Solar radiation properties ...............................................................................4-26 Other radiation modelling considerations ......................................................4-26 Useful points ..................................................................................................4-27 Symmetry Plane Boundaries ...................................................................................4-27 Cyclic Boundaries ...................................................................................................4-27 Introduction ....................................................................................................4-27 Setting up models ...........................................................................................4-28 Useful points ..................................................................................................4-30 Cyclic set manipulation ..................................................................................4-31 Free-stream Transmissive Boundaries ....................................................................4-32 Introduction ....................................................................................................4-32 Useful points ..................................................................................................4-33 Transient-wave Transmissive Boundaries ...............................................................4-34 Introduction ....................................................................................................4-34 Useful points ..................................................................................................4-35 Riemann Boundaries ...............................................................................................4-36 Introduction ....................................................................................................4-36 Useful points ..................................................................................................4-37 Attachment Boundaries ...........................................................................................4-38 Useful points ..................................................................................................4-39 Radiation Boundaries ..............................................................................................4-39 Useful points ..................................................................................................4-40 Phase-Escape (Degassing) Boundaries ...................................................................4-40 Monitoring Regions .................................................................................................4-40 Boundary Visualisation ...........................................................................................4-41 Solution Domain Initialisation ................................................................................4-42 Steady-state problems ....................................................................................4-42 Transient problems .........................................................................................4-42 CONTROL FUNCTIONS Introduction ...............................................................................................................5-1 Analysis Controls for Steady-State Problems ...........................................................5-1 Analysis Controls for Transient Problems ................................................................5-4 Default (single-transient) solution mode .........................................................5-4 Version 4.02

6

7

8

Load-step based solution mode ....................................................................... 5-6 Load step characteristics .................................................................................. 5-6 Load step definition ......................................................................................... 5-8 Solution procedure outline .............................................................................. 5-9 Other transient functions ............................................................................... 5-14 Solution Control with Mesh Changes ..................................................................... 5-15 Mesh-changing procedure ............................................................................. 5-15 Solution-Adapted Mesh Changes ........................................................................... 5-17 POROUS MEDIA FLOW Setting Up Porous Media Models ............................................................................. 6-1 Useful Points ............................................................................................................. 6-4 THERMAL AND SOLAR RADIATION Radiation Modelling for Surface Exchanges ............................................................ 7-1 Radiation Modelling for Participating Media ........................................................... 7-3 Capabilities and Limitations of the DTRM Method ................................................. 7-5 Capabilities and Limitations of the DORM Method ................................................. 7-7 Radiation Sub-domains ............................................................................................. 7-8 CHEMICAL REACTION AND COMBUSTION Introduction ............................................................................................................... 8-1 Local Source Models ................................................................................................ 8-2 Presumed Probability Density Function (PPDF) Models ......................................... 8-3 Single-fuel PPDF ............................................................................................. 8-3 Multiple-fuel PPDF ......................................................................................... 8-9 Regress Variable Models ........................................................................................ 8-10 Complex Chemistry Models ................................................................................... 8-11 Setting Up Chemical Reaction Schemes ................................................................. 8-14 Useful general points for local source and regress variable schemes ........... 8-16 Chemical Reaction Conventions ................................................................... 8-18 Useful points for PPDF schemes ................................................................... 8-18 Useful points for complex chemistry models ................................................ 8-21 Useful points for ignition models .................................................................. 8-21 Setting Up Advanced I.C. Engine Models .............................................................. 8-22 Coherent Flame model (CFM) ...................................................................... 8-24 Extended Coherent Flame model (ECFM) .................................................... 8-26 Extended Coherent Flame model 3Z (ECFM-3Z) — spark ignition ............ 8-28 Extended Coherent Flame model 3Z (ECFM-3Z) — compression ignition . 8-29 Useful points for ECFM models .................................................................... 8-30 Level Set model ............................................................................................. 8-31 Write Data sub-panel ..................................................................................... 8-32

Version 4.02

v

9

The Arc and Kernel Tracking ignition model (AKTIM) ...............................8-33 Useful points for the AKTIM model .............................................................8-35 The Double-Delay autoignition model ..........................................................8-37 NOx Modelling ........................................................................................................8-39 Soot Modelling ........................................................................................................8-39 Coal Combustion Modelling ...................................................................................8-41 Stage 1 ............................................................................................................8-41 Stage 2 ............................................................................................................8-42 Useful notes ...................................................................................................8-44 Switches and constants for coal modelling ....................................................8-45 Special settings for the Mixed-is-Burnt and Eddy Break-Up models ............8-46 LAGRANGIAN MULTI-PHASE FLOW

Setting Up Lagrangian Multi-Phase Models .............................................................9-1 Data Post-Processing .................................................................................................9-4 Static displays ..................................................................................................9-5 Trajectory displays ...........................................................................................9-8 Engine Combustion Data Files ..................................................................................9-9 Useful Points ...........................................................................................................9-10 10 EULERIAN MULTI-PHASE FLOW Introduction .............................................................................................................10-1 Setting up multi-phase models ................................................................................10-1 Useful points on Eulerian multi-phase flow ..................................................10-4 11 FREE SURFACE AND CAVITATION Free Surface Flows ..................................................................................................11-1 Setting up free surface cases ..........................................................................11-1 Cavitating Flows ......................................................................................................11-5 Setting up cavitation cases .............................................................................11-5 12 ROTATING AND MOVING MESHES Rotating Reference Frames .....................................................................................12-1 Models for a single rotating reference frame .................................................12-1 Useful points on single rotating frame problems ...........................................12-1 Models for multiple rotating reference frames (implicit treatment) ..............12-2 Useful points on multiple implicit rotating frame problems ..........................12-4 Models for multiple rotating reference frames (explicit treatment) ...............12-5 Useful points on multiple explicit rotating frame problems ..........................12-8 Moving Meshes .......................................................................................................12-9 Basic concepts ................................................................................................12-9 Setting up models .........................................................................................12-10 Useful points ................................................................................................12-13 vi

Version 4.02

Automatic Event Generation for Moving Piston Problems ......................... 12-13 Cell-layer Removal/Addition ................................................................................ 12-14 Basic concepts ............................................................................................. 12-14 Setting up models ........................................................................................ 12-15 Useful points ................................................................................................ 12-18 Sliding Meshes ...................................................................................................... 12-18 Regular sliding interfaces ............................................................................ 12-18 Cell Attachment and Change of Fluid Type ......................................................... 12-22 Basic concepts ............................................................................................. 12-22 Setting up models ........................................................................................ 12-23 Useful points ................................................................................................ 12-27 Mesh Region Exclusion ........................................................................................ 12-28 Basic concepts ............................................................................................. 12-28 Moving Mesh Pre- and Post-processing ............................................................... 12-28 Introduction ................................................................................................. 12-28 Action commands ........................................................................................ 12-29 Status setting commands ............................................................................. 12-30 13 OTHER PROBLEM TYPES Multi-component Mixing ........................................................................................ 13-1 Setting up multi-component models .............................................................. 13-1 Useful points on multi-component mixing .................................................... 13-3 Aeroacoustic Analysis ............................................................................................ 13-3 Setting up aeroacoustic models ..................................................................... 13-3 Useful points on aeroacoustic analyses ......................................................... 13-4 Liquid Films ............................................................................................................ 13-5 Setting up liquid film models ........................................................................ 13-5 Film stripping ................................................................................................ 13-7 14 USER PROGRAMMING Introduction ............................................................................................................. 14-1 Subroutine Usage .................................................................................................... 14-1 Useful points .................................................................................................. 14-4 Description of UFILE Routines .............................................................................. 14-5 Boundary condition subroutines .................................................................... 14-5 Material property subroutines ........................................................................ 14-6 Turbulence modelling subroutines ................................................................ 14-9 Source subroutines ....................................................................................... 14-10 Radiation modelling subroutines ................................................................. 14-11 Free surface / cavitation subroutines ........................................................... 14-11 Lagrangian multi-phase subroutines ............................................................ 14-12 Version 4.02

vii

Liquid film subroutines ................................................................................14-14 Eulerian multi-phase subroutines .................................................................14-14 Chemical reaction subroutines .....................................................................14-15 Rotating reference frame subroutines ..........................................................14-16 Moving mesh subroutines ............................................................................14-16 Miscellaneous flow characterisation subroutines ........................................14-17 Solution control subroutines ........................................................................14-18 Sample Listing .......................................................................................................14-19 New Coding Practices ...........................................................................................14-20 User Coding in parallel runs ..................................................................................14-22 15 PROGRAM OUTPUT Introduction .............................................................................................................15-1 Permanent Output ....................................................................................................15-1 Input-data summary .......................................................................................15-1 Run-time output .............................................................................................15-3 Printout of Field Values ..........................................................................................15-3 Optional Output .......................................................................................................15-3 Example Output .......................................................................................................15-4 16 pro-STAR CUSTOMISATION Set-up Files ..............................................................................................................16-1 Panels .......................................................................................................................16-2 Panel creation .................................................................................................16-2 Panel definition files ......................................................................................16-5 Panel manipulation .........................................................................................16-6 Macros .....................................................................................................................16-6 Function Keys ..........................................................................................................16-9 17 OTHER STAR-CD FEATURES AND CONTROLS Introduction .............................................................................................................17-1 File Handling ...........................................................................................................17-1 Naming conventions ......................................................................................17-1 Commonly used files .....................................................................................17-1 File relationships ............................................................................................17-7 File manipulation ...........................................................................................17-9 Special pro-STAR Features ...................................................................................17-12 pro-STAR environment variables ................................................................17-12 Resizing pro-STAR ......................................................................................17-13 Special pro-STAR executables ....................................................................17-14 Use of temporary files by pro-STAR ...........................................................17-14 The StarWatch Utility ...........................................................................................17-15 viii

Version 4.02

Running StarWatch ..................................................................................... 17-15 Choosing the monitored values ................................................................... 17-17 Controlling STAR ....................................................................................... 17-17 Manipulating the StarWatch display ........................................................... 17-20 Monitoring another job ................................................................................ 17-21 Hard Copy Production .......................................................................................... 17-21 Neutral plot file production and use ............................................................ 17-21 Scene file production and use ...................................................................... 17-23

APPENDICES A pro-STAR CONVENTIONS Command Input Conventions .................................................................................. A-1 Help Text / Prompt Conventions ............................................................................. A-3 Control and Function Key Conventions .................................................................. A-4 File Name Conventions ............................................................................................ A-4 B FILE TYPES AND THEIR USAGE C PROGRAM UNITS D pro-STAR X-RESOURCES E USER INTERFACE TO MESSAGE PASSING ROUTINES F STAR RUN OPTIONS Usage .........................................................................................................................F-1 Options ......................................................................................................................F-1 Parallel Options .........................................................................................................F-3 Resource Allocation ..................................................................................................F-6 Default Options .........................................................................................................F-7 Cluster Computing ....................................................................................................F-8 Batch Runs Using STAR-NET .................................................................................F-8 Running under IBM Loadleveler using STAR-NET .......................................F-8 Running under LSF using STAR-NET ...........................................................F-9 Running under OpenPBS using STAR-NET ................................................F-10 Running under PBSPro using STAR-NET ....................................................F-11 Running under SGE using STAR-NET .........................................................F-11 Running under Torque using STAR-NET .....................................................F-12 G BIBLIOGRAPHY

INDEX INDEX OF COMMANDS

Version 4.02

ix

OVERVIEW Purpose The Methodology volume presents the mathematical modelling practices embodied in the STAR-CD system and the numerical solution procedures employed. In this volume, the focus is on the structure of the system itself and how to use it. This presentation assumes that the reader is familiar with the background information provided in the Methodology volume.

Contents Chapter 1 introduces some of the fundamental principles of computational continuum mechanics, including an outline of the basic steps involved in setting up and using a successful computer model. The important factors to consider at each step are mostly explained independently of the computer system used to perform the analysis. However, reference is also made to the particular capabilities of the STAR-CD system, where appropriate. Chapter 2 outlines the basic features of STAR-CD, including GUI facilities, session control and plotting utilities. Chapters 3 to 5 provide the reader with detailed instructions on how to use some of the basic code facilities, e.g. boundary condition specification, material property definition, etc., and an overview of the GUI panels appropriate to each of them. The description covers all facilities (other than mesh generation) that might be employed for modelling most common continuum mechanics problems. Mesh generation itself is covered in a separate volume, the Meshing User Guide. Chapters 2 to 5 should be read at least once to gain an understanding of the general housekeeping principles of pro-STAR and to help with any problems arising from routine operations. It is recommended that users refer to the appropriate chapter repeatedly when setting up a model for general guidance and an overview of the relevant GUI panels. Chapters 6 to 13 describe additional STAR-CD capabilities relevant to models of a more specialised nature, i.e. rotating systems, combustion processes, buoyancy-driven flows, etc. Users interested in a particular topic should consult the appropriate section for a summary of commands or options specially designed for that purpose, plus hints and tips on performing a successful simulation. Chapter 14 outlines the user programmability features available and provides an example FORTRAN subroutine listing implementing these features. All such subroutines are readily available for use and can be easily adapted to suit the model's requirements. Chapter 15 presents the printable output produced by the code which provides, among other things, a summary of the problem specification and monitoring information generated during the calculation. Chapter 16 explains how pro-STAR can be customised, in terms of user-defined panels, macros and keyboard function keys, to meet a user’s individual requirements. Finally, Chapter 17 covers some of the less commonly used features of STAR-CD, including the interaction between STAR and pro-STAR and how various system files are used. Version 4.02

1

Chapter 1

COMPUTATIONAL ANALYSIS PRINCIPLES Introduction

Chapter 1

COMPUTATIONAL ANALYSIS PRINCIPLES

Introduction The aim of this section is to introduce the most important issues involved in setting up and solving a continuum mechanics problem using a computational continuum mechanics code. Although the discussion applies in principle to any such code, reference is made where appropriate to the particular capabilities of the STAR-CD system. It is also assumed that the reader is familiar with the material presented in the Methodology volume. The process of computational mechanics simulation does not usually start with the direct use of such a code. It is indeed important to recognise that STAR-CD, or any other CFD, CAD or CAE system, should be treated as a tool to assist the engineer in understanding physical phenomena. The success or failure of a continuum mechanics simulation depends not only on the code capabilities, but also upon the input data, such as: • • • •

Geometry of the solution domain Continuum properties Boundary conditions Solution control parameters

For a simulation to have any chance of success, such information should be physically realistic and correctly presented to the analysis code. The essential steps to be taken prior to computational continuum mechanics (CCM) modelling are as follows: • • • •

Pose the problem in physical terms. Establish the amount of information available and its sufficiency and validity. Assess the capabilities and features of the STAR-CD code, to ensure that the problem is well posed and amenable to numerical solution by the code. Plan the simulation strategy carefully, adopting a step-by-step approach to the final solution.

Users should turn to STAR-CD and proceed with the actual modelling only after the above tasks have been completed.

The Basic Modelling Process The modelling process itself can be divided into four major phases, as follows: Phase 1 — Working out a modelling strategy This requires a precise definition of the physical system’s geometry, material properties and flow/deformation conditions based on the best available understanding of the relevant physics. The necessary tasks include: • • •

Version 4.02

Planning the computational mesh (e.g. number of cells, size and distribution of cell dimensions, etc.). Looking up numerical values for appropriate physical parameters (e.g. density, viscosity, specific heat, etc.). Choosing the most suitable modelling option from what is available (e.g. turbulence model, combustion option, etc.). 1-1


Chapter 1

Spatial description and volume discretisation

The user also has to balance the requirement of physical fidelity and numerical accuracy against the simulation cost and computational capabilities of his system. His modelling strategy will therefore incorporate some trade-off between these two factors. This initial phase of modelling is particularly important for the smooth and efficient progress of the computational simulation. Phase 2 — Setting up the model using pro-STAR The main tasks involved at this phase are: • • • • •

Creating a computational mesh to represent the solution domain (i.e. the model geometry). Specifying the physical properties of the fluids and/or solids present in the simulation and, where relevant, the turbulence model(s), body forces, etc. Setting the solution parameters (e.g. solution variable selection, relaxation coefficients, etc.) and output data formats. Specifying the location and definition of boundaries and, for unsteady problems, further definition of transient boundary conditions and time steps. Writing appropriate data files as input to the analytical run of the following phase.

Phase 3 — Performing the analysis using STAR This phase consists of: • •

Reading input data created by pro-STAR and, if dealing with a restart run, the results of a previous run. Judging the progress of the run by analysing various monitoring data and solution statistics provided by STAR.

Phase 4 — Post-processing the results using pro-STAR This involves the display and manipulation of output data created by STAR using the appropriate pro-STAR facilities. The remainder of this chapter discusses the elements of each modelling phase in greater detail.

Spatial description and volume discretisation One of the basic steps in preparing a STAR-CD model is to describe the geometry of the problem. The two key components of this description are: • •

1-2

The definition of the overall size and shape of the solution domain. The subdivision of the solution domain into a mesh of discrete, finite, contiguous volume elements or cells, as shown in Figure 1-1.

Version 4.02

Chapter 1

COMPUTATIONAL ANALYSIS PRINCIPLES Spatial description and volume discretisation

Figure 1-1

Example of solution domain subdivision into cells

This process is called volume discretisation and is an essential part of solving the above equations numerically. In STAR-CD both components of the spatial description are performed as part of the same operation, setting up the finite-volume mesh, but separate considerations apply to each of them. Solution domain definition Through its internal design and construction, STAR-CD permits a very general and flexible definition of what constitutes a solution domain. The latter can be: • • • •

A fluid and/or heat flow field fully occupying an open region of space Fluid and/or heat flowing through a porous medium Heat flowing through a solid A solid undergoing mechanical deformation

Arbitrary combinations of the above conditions can also be specified within the same model, as in problems involving fluid-solid heat transfer. The user’s first task is therefore to decide which parts of the physical system being modelled need to be included in the solution domain and whether each part is occupied by a fluid, solid or porous medium. Whatever its composition, the fundamental requirement is that the solution domain is bounded. This means that the user has to examine his system’s geometry carefully and decide exactly where the enclosing boundaries lie. The boundaries can be one of four kinds: 1. Physical boundaries — walls or solid obstacles of some description that serve to physically confine a fluid flow 2. Symmetry boundaries — axes or planes beyond which the problem solution becomes a mirror image of itself 3. Cyclic boundaries — surfaces beyond which the problem solution repeats itself, in a cyclic or anticyclic fashion The purpose of symmetry and cyclic boundaries is to limit the size of the domain, and hence the computer requirements, by excluding regions where the solution is essentially known. This in turn allows one to model the problem in greater detail than would have been the case otherwise. Version 4.02

1-3


Chapter 1


4. Notional boundaries — these are non-physical surfaces that serve to ‘close-off’ the solution domain in regions not covered by the other two types of boundary. Their location is entirely up to the user’s discretion but, in general, they should be placed only where one of the following apply: (a) Flow/deformation conditions are known (b) Flow/deformation conditions can be guessed reasonably well (c) The boundary is far enough away from the region of interest for boundary condition inaccuracies to have little effect Thus, locating this type of boundary may require some trial and error. The location and characterisation of boundaries is discussed further in “Boundary description” on page 1-10. Mesh definition Creation of a lattice of finite-volume cells to represent the solution domain is normally the most time-consuming task in setting up a STAR-CD model. This process is greatly facilitated by STAR-CD because of its ability to generate cells of an arbitrary, polyhedral shape. In creating a finite-volume mesh, the user should aim to represent accurately the following two entities: 1. The overall external geometry of the solution domain, by specifying an appropriate size and shape for near-boundary cells. The latter’s external faces, taken together, should make up a surface that adequately represents the shape of the solution domain boundaries. Small inaccuracies may occur because all boundary cell faces (including rectangular faces) are composed of triangular facets, as shown in Figure 1-2. These errors diminish as the mesh is refined.

triangular facet

Figure 1-2

Boundary representation by triangular facets

2. The internal characteristics of the flow/deformation regime. This is achieved by careful control of mesh spacing within the solution domain interior so that 1-4

Version 4.02

Chapter 1


the mesh is finest where the problem characteristics change most rapidly. Near-wall regions are important and a high mesh density is needed to resolve the flow in their vicinity. This point is discussed further in “Mesh distribution near walls” on page 1-7. Mesh spacing considerations The chief considerations governing the mesh spatial arrangement are: • • •

Accuracy — primarily determined by mesh density and, to a lesser extent, mesh distortion (discussed in “Mesh distortion” on page 1-5). Numerical stability — this is a strong function of the degree of distortion. Cost — a function of both the aforementioned factors, through their influence on the speed of convergence and c.p.u. time required per iteration or time step.

Thus, the user should aim at an optimum mesh arrangement which • • •

employs the minimum number of cells, exhibits the least amount of distortion, is consistent with the accuracy requirements.

Chapter 2 of the Meshing User Guide describes several methods available in STAR-CD, some of them semi-automatic, to help the user achieve this goal. However, even when suitable automatic mesh generation procedures are available, the user must still draw on knowledge and experience of computational fluid and solid mechanics to produce the right kind of mesh arrangement. Mesh distortion Mesh distortion is measured in terms of three factors — aspect ratio, internal angle and warp angle — illustrated in Figure 1-3.

φ

b

θ

a b/a = aspect ratio

Figure 1-3

θ = internal angle

φ = warp angle

Cell shape characteristics

When setting up the mesh, the user should try to observe the following guidelines: • Version 4.02

Aspect Ratio — values close to unity are preferable, but departures from this 1-5


Chapter 1


• •

are allowed. Internal Angle — departures from 90° intersections between cell faces should be kept to a minimum. Warp Angle — the optimum value of this angle is zero, which can occur only when the cell face vertices are co-planar.

Any adverse effects arising from departures from the preferred values of these factors manifest themselves through • •

the relative magnitudes of the coefficients in the finite-volume equations, especially those arising from non-orthogonality, and the signs of the coefficients (negative values are generally detrimental).

It is difficult to place rigid limits on the acceptable departures because they depend on local flow conditions. However, the following values serve as a useful guideline: Aspect Ratio Internal angle Warp angle

10 45° 45°

pro-STAR can calculate these quantities and identify cells having out-of-bounds values, as discussed in Chapter 3, “Mesh and Geometry Checking” of the Meshing User Guide. What is really important in this respect is the combined effect of the various kinds of mesh distortion. If all three are simultaneously present in a single cell, the limits given above might not be stringent enough. On the other hand, the effects of distortion also depend on the nature of the local flow. Thus, the above limits may be exceeded in the region of • •

simple flows such as, for example, uniform-velocity ‘free’ streams, wall boundary layers, where cells of high aspect ratio (in the flow direction) are commonly employed without difficulty.

Generally speaking, non-orthogonality at boundaries may cause problems and should be minimised whenever practicable. Mesh distribution and density Numerical discretisation errors are functions of the cell size; the smaller the cells (and therefore the higher the mesh density), the smaller the errors. However, a high mesh density implies a large number of mesh storage locations, with associated high computing cost. It is therefore advisable, wherever possible, to • •

ensure that the mesh density is high only where needed, i.e. in regions of steep gradients of the flow variables, and low elsewhere; avoid rapid changes in cell dimensions in the direction of steep gradients in the flow variables.

The flexibility afforded by STAR-CD’s unstructured polyhedral meshes facilitates such selective refinement. An illustration of some of the numerous cell shapes that may be employed is given in Figure 2-43 and Figure 2-44 of the Meshing User Guide. 1-6

Version 4.02

Chapter 1


Of course, it is not always possible to ascertain a priori what the flow structure will be. However, the need for higher mesh density can usually be anticipated in regions such as: • • • • • •

Wall boundary layers Jets issuing from apertures Shear layers formed by flow separation or neighbouring streams of different velocities Stagnation points produced by flow impingement Wakes behind bluff bodies Temperature or concentration fronts arising from mixing or chemical reaction

Mesh distribution near walls As discussed in Chapter 6, “Wall Boundary Conditions” of the Methodology volume, wall functions are an economic way of representing turbulent boundary layers (hydrodynamic and thermal) in turbulent flow calculations. These functions effectively allow the boundary layer to be bridged by a single cell, as shown in Figure 1-4(a).

Outer region

y

Inner region

(a) Wall function model

Figure 1-4

(b) Two-layer or Low Re models

Near-wall mesh distribution

The location y of the cell centroids in the near-wall layer, and hence the thickness of that layer, is usually determined by reference to the dimensionless normal distance y + from the wall. For the wall function to be effective, this distance must be • •

not too small, otherwise, the ‘bridge’ might span only the laminar sublayer; not too large, as the flow at that location might not behave in the way assumed in deriving the wall functions.

Ideally, y + should lie in the approximate range 30 to 150. Note that the above considerations apply to Reynolds Stress models as well as several classes of eddy viscosity model (see Chapter 3, “Turbulence Modelling”). Alternative treatments that do not require the use of wall functions are also available. These are: Version 4.02

1-7


Chapter 1

Problem characterisation and material property definition

1. Two-layer turbulence models, whereby wall functions are replaced by a one-equation k-l model or a zero-equation mixing-length model 2. Low Reynolds number models (including the V2F model), where viscous effects are incorporated in the k and ε transport equations For the above two types of model, the solution domain should be divided into two regions with the following characteristics: • •

An inner region containing a fine mesh An outer region containing normal mesh sizes

The two regions are illustrated in Figure 1-4(b). As explained in the Methodology volume (Chapter 6, “Two-layer models”), the inner region should contain at least 15 mesh layers and encompass that part of the boundary layer influenced by viscous effects. A more recent development, called the hybrid wall function is also available that extends the low-Reynolds number formulation of most turbulence models. This may be used to capture boundary layer properties more accurately in cases where the near-wall cell size is not adapted for the low-Reynolds number treatment and thus achieve y + independent solutions. Moving mesh features STAR-CD offers a range of moving mesh features, including: • • •

General mesh motion Internal sliding mesh Cell deletion and insertion

The first of these is straightforward to employ and the only caution required is the obvious one: avoid creating excessive distortion when redistributing the mesh. This caution also applies to the use of the other two features, but they have additional rules and guidelines attached to them. These are summarised in the Methodology volume, Chapter 15 (“Internal Sliding Mesh” on page 15-5 and “Cell Layer Removal and Addition” on page 15-7). Additional guidelines also appear in this volume, “Cell-layer Removal/Addition” on page 12-14 and “Sliding Meshes” on page 12-18; hence they are not repeated here.

Problem characterisation and material property definition Correct definition of the physical conditions and the properties of the materials involved is a prerequisite to obtaining the right solution to a problem, or indeed to obtaining any solution at all. It is also essential for determining whether the problem can be modelled with STAR-CD. The user must therefore ensure that the problem is well defined in respect of: • • • • 1-8

The nature of the fluid flow (e.g. steady/unsteady, laminar/turbulent, incompressible/compressible) Physical properties (e.g. density, viscosity, specific heat) External force fields (e.g. gravity, centrifugal forces) and energy sources, when present Initial conditions for transient flows Version 4.02

Chapter 1

COMPUTATIONAL ANALYSIS PRINCIPLES Problem characterisation and material property definition

Nature of the flow It is very important to understand the nature of the flow being analysed in order to select the appropriate mathematical models and numerical solution algorithms. Problems will arise if an incorrect choice is made, as in the following examples: • • •

Employing an iterative, steady-state algorithm for an inherently unsteady problem, such as vortex shedding from a bluff body Computing a turbulent flow without invoking a suitable turbulence model Modelling transitional flow with one of the turbulence models currently implemented in STAR-CD. None of them can represent transitional behaviour accurately.

Physical properties The specification of physical properties, such as density, molecular viscosity, thermal conductivity, etc. depends on the nature of the fluids or solids involved and the circumstances of use. For example, STAR-CD contains several built-in equations of state from which density can be calculated as a function of one or more of the following field variables: • • •

Pressure Temperature Fluid composition

In all cases where complex calculations are used to evaluate a material property, the following measures are recommended: • • •

•

The relevant field variables must be assigned plausible initial and boundary values. Where necessary, properties should be solved for together with the field variables as part of the overall solution. In the case of strong dependencies between properties and field variables, the user should consider under-relaxation of the property value calculations, in the manner described in the Methodology volume (Chapter 7, “Scalar transport equations”). When required, STAR-CD’s facility for alternative, user-programmable functions may be used.

Force fields and energy sources As already noted, STAR-CD has built-in provision for body forces arising from • •

buoyancy, rotation.

It is important to remember that as the strength of the body forces increases relative to the viscous (or turbulent) stresses, the flow may become physically unstable. In these circumstances it is advisable to switch to the transient solution mode. It is also possible to insert additional, external force fields and energy sources via the user programming facilities of STAR-CD. In such cases, it is important to understand the physical implications and avoid specifying conditions that lead to Version 4.02

1-9


Chapter 1

Boundary description

physical or numerical instability. Examples of such conditions are: • •

Thermal energy sources that increase linearly with temperature. These can give rise to physical instability called ‘thermal runaway’. Setting the coefficient β i in the permeability function K i = α i v + β i to a very small or zero value. If the local fluid velocity also becomes very small, the result may be numerical instability whereby small pressure perturbations produce a large change in velocities.

Initial conditions The term ‘initial conditions’ refers to values assigned to the dependent variables at all mesh points before the start of the calculations. Their implication depends on the type of problem being considered: •

•

In unsteady applications, this information has a clear physical significance and will affect the course of the solution. Due care must therefore be taken in providing it. It sometimes happens that the effects of initial conditions are confined to a start-up phase that is not of interest (as in, for example, flows that are temporally periodic). However, it is still advisable to take some precautions in specifying initial conditions for reasons explained below. In calculating steady state problems by iterative means, the initial conditions will usually have no influence on the final solution (apart from rare occasions when the solution is multi-valued), but may well determine the success and speed of achieving it.

Poor initial field specifications or, for transient problems, abrupt changes in boundary conditions put severe demands on the numerical algorithm when substituted into the finite-volume equations. As a consequence, the following special ‘start-up’ measures may be necessary to ensure numerical stability: • •

Use of unusually small time steps in transient calculations. Use of strong under-relaxation in iterative solutions.

Specific recommendations concerning these practices are given in “Numerical solution control” on page 1-13. In either case, increased computing times can be an undesirable side effect.

Boundary description As stated in “Spatial description and volume discretisation” on page 1-2, boundary identification and description are intimately connected with the generation of the finite-volume mesh, since the boundary topography is defined by the outermost cell faces. Furthermore, correct specification of the boundary conditions is often the main area of difficulty in setting up a model. Problems often arise in the following areas: • • •

Identifying the correct type of condition Specifying an acceptable mix of boundary types Ascribing appropriate boundary values

The above are in turn linked to the decisions on where to place the boundaries in the 1-10

Version 4.02

Chapter 1

COMPUTATIONAL ANALYSIS PRINCIPLES Boundary description

first instance. Boundary location Difficulties in specifying boundary location normally arise where the flow conditions are incompletely known, for example at outlets. The recommended solutions, in decreasing degree of accuracy, are to place boundaries • • •

in regions where the conditions are known, if this is possible; in a location where the ‘Outlet’ or ‘Prescribed Pressure’ option is applicable (see Chapter 5 in the Methodology volume); where the approximations in the boundary condition specification are unlikely to propagate upstream into the regions of interest.

Whenever possible, it is particularly important to avoid the following situations: 1. A boundary that passes through a major recirculation zone. 2. In transient transonic or supersonic compressible flows, an outlet boundary located where the flow is not supersonic. 3. A mix of boundary conditions that is inappropriate. Examples of this are: (a) Multiple ‘Outlet’ boundaries — unless further information is supplied on how the flow is partitioned between the outlets. (b) Prescribed flow split outlets coexisting with prescribed mass outflow boundaries in the same domain. (c) A combination of prescribed pressure and flow-split outlet conditions. Boundary conditions Another source of potential difficulty is in boundary value specification wherever known conditions need to be set, e.g. at a ‘Prescribed Inflow’ or ‘Inlet’ boundary. The basic points to bear in mind in this situation are: • •

All transport equations to be solved require specification of their boundary values, including the turbulence transport equations when they are invoked Inappropriate setting of boundary values leads to erroneous results and, in extreme cases, to numerical instability

The following recommendations can be given regarding each different type of boundary: 1. Prescribed flow — Here, care should be taken to: (a) Assign realistic values to all dependent variables, including the turbulence parameters, and also to auxiliary quantities, such as density. (b) Ensure that, if this is the only type of flow boundary imposed, overall continuity is satisfied (STAR-CD will accept inadvertent mass imbalances of up to 5%, correcting them by adjusting the outflows. An error message is issued if the imbalance exceeds this figure). 2. Outlet — The main points to note for this boundary type are: (a) The need to specify the boundary, where possible, at locations where the flow is everywhere outwardly directed; also to recognise that, if inflow Version 4.02

1-11


Chapter 1

Boundary description

occurs, it may introduce numerical instability and/or inaccuracies. (b) The necessity, if more than one boundary of this type is declared, of prescribing either the flow split between them or the mass outflow rate at each location. (c) The inapplicability of ‘prescribed split’ outlets to problems where the inflows are not fixed, e.g. i) in combination with pressure boundary conditions, or ii) in the case of transient compressible flows. 3. Prescribed pressure — The main precautions are: (a) To specify relative (to a prescribed datum) rather than absolute pressures. (b) If inflow is likely to occur, to assign realistic boundary values to temperature and species mass fractions. It is also advisable to specify the turbulence parameters indirectly, via the turbulence intensity and length scale or by extrapolating them from values in the interior of the solution domain. 4. Stagnation conditions — It is recommended to use this condition for boundaries lying within large reservoirs where properties are not significantly affected by flow conditions in the solution domain. 5. Non-reflecting pressure and stagnation conditions — A special formulation of the standard pressure and stagnation conditions, developed to facilitate analysis of steady-state turbomachinery applications 6. Cyclic boundaries — These always occur in pairs. The main points of advice are: (a) Impose this condition only in appropriate circumstances. Two-dimensional axisymmetric flows with swirl is a good example of an appropriate application. (b) For axisymmetric flows, make use of the CD/UD blending scheme to apply the maximum level of central differencing in the tangential direction (the default blending factor is 0.95; see also on-line Help topic “Miscellaneous Controls” in STAR GUIde). 7. Planes of symmetry — It is recommended to use this condition for two-dimensional axisymmetric flows without swirl 8. Free-stream transmissive boundaries — Used only for modelling supersonic free streams 9. Transient wave transmissive boundaries — Used only in problems involving transient compressible flows 10. Riemann boundaries — This condition is based on the theory of Riemann invariants and its application allows pressure waves to leave the solution domain without reflection

1-12

Version 4.02

Chapter 1

COMPUTATIONAL ANALYSIS PRINCIPLES Numerical solution control

Numerical solution control Proper control of the numerical solution process applied to the transport equations is highly important, both for acceptable computational efficiency and, sometimes, in order to achieve a solution at all. By necessity, the means of controlling the process depend heavily on the particular numerical techniques employed so no universal guidelines can be given. Thus, the recommended settings vary with the particular algorithm selected and the circumstances of application. Selection of solution procedure The basic selection should be based on a correct assessment of the nature of the problem and will be either • •

a transient calculation, starting from well-defined initial and boundary conditions and proceeding to a new state in a series of discrete time steps; or a steady-state calculation, where an unchanging flow/deformation pattern under a given set of boundary conditions is arrived at through a number of numerical iterations.

PISO and SIMPLE are the two alternative solution procedures available in STAR-CD. PISO is the default for unsteady calculations and is sometimes preferred for steady-state ones, in cases involving strong coupling between dependent variables such as buoyancy driven flows. SIMPLE is the default algorithm for steady-state solutions and works well in most cases. SIMPLE is also used for transient calculations in the case of free surface and cavitating flows, where it is the only option. In most other transient flow problems, PISO is likely to be more efficient due to the fact that PISO correctors are usually cheaper than outer iterations on all variables within a time step of the transient SIMPLE algorithm. However, there are situations in which PISO would require many correctors or even fail to converge unless the time step is reduced, whereas SIMPLE may allow larger time steps with a moderate number of outer iterations per time step. This is the case when the flow changes very little but certain slow transients are present in the behaviour of scalar variables (e.g. slow heating up of solid structures in the case of solid-fluid heat transfer problems, deposition of chemical species on walls in after-treatment of exhaust gases, etc.). In such cases, transient SIMPLE can be used to save on computing time. When doubts exist as to whether the problem considered actually possesses a steady-state solution or when iterative convergence is difficult to achieve, it is better to perform the calculations using the transient option. Transient flow calculations with PISO As stated in “The PISO algorithm” on page 7-2 of the Methodology volume, PISO performs at each time (or iteration) step, a predictor, followed by a number of correctors, during which linear equation sets are solved iteratively for each main dependent variable. The decisions on the number of correctors and inner iterations (hereafter referred to as ‘sweeps’, to avoid confusion with outer iterations performed as part of the steady-state solution mode) are made internally on the basis of the splitting error and inner residual levels, respectively, according to prescribed tolerances and upper limits. The default values for the solver tolerances and Version 4.02

1-13


Chapter 1

Numerical solution control

maximum correctors and sweeps are given in Table 1-1. Normally, these will only require adjustment by the user in exceptional circumstances, as discussed below. Table 1-1: Standard Control Parameter Settings for Transient PISO Calculations Variable Parameter Velocity

Pressure

Turbulence

Enthalpy

Mass fraction

Solver tolerance

0.01

0.001

0.01

0.01

0.01

Sweep limit

100

1000

100

100

100

Pressure under-relaxation factor = 1.0 Corrector limit = 20 Corrector step tolerance = 0.25 The remaining key parameter in transient calculations with PISO is the size of the time increment δt . This is normally determined by accuracy considerations and may be varied during the course of the calculation. The step should ideally be of the same order of magnitude as the smallest characteristic time δt c for convection and diffusion, i.e. 2

ρδL δt c = min ⎛⎜ δL ------, ------------⎞⎟ Γ U ⎝ ⎠

(1-1)

Here, U and Γ are a characteristic velocity and diffusivity, respectively, and δL is a mean mesh dimension. Typically, it is possible to operate with δt ≈ 50 δt c and still obtain reasonable temporal accuracy. Values significantly above this may lead to errors and numerical instability, whereas smaller values will lead to increased computing times. During the course of a calculation, the limits given in Table 1-1 may be reached, in which case messages to this effect will be produced. This is most likely to occur during the start-up phase but is nevertheless acceptable if, later on, the warnings either cease entirely or only appear occasionally, and the predictions look reasonable. If, however, the warnings persist, corrective actions should be taken. The possible actions are: • •

•

• 1-14

Reduction in time step by, say, an initial factor of 2 — if this improves matters, then the cause may simply be an excessively large δt . Increase in the sweep limits — if measure 1 fails, then this should be tried, only on the variable(s) whose limit(s) have been reached. Again, twofold changes are appropriate. Pressure under-relaxation — a value of 0.8 for pressure correction under-relaxation, using PISO, may be helpful for some difficult cases (e.g. for severe mesh distortion or flows with Mach numbers approaching 1). Corrector step tolerance — this may be set to a lower value but consult Version 4.02

Chapter 1


CD adapco first. Steady-state flow calculations with PISO When PISO operates in this mode, the inner residual tolerances are decreased and under-relaxation is introduced on all variables, apart from pressure, temperature and mass fraction. However, the last two variables may need to be under-relaxed for buoyancy driven problems. The standard, default values for these parameters and the sweep limits, which are unchanged from the transient mode, are given in Table 1-2. .

Table 1-2: Standard Control Parameter Settings for Steady PISO Calculations Variable Parameter Velocity

Pressure

Turbulence

Enthalpy

Mass fraction

Solver tolerance

0.1

0.05

0.1

0.1

0.1

Sweep limit

100

1000

100

100

100

Relaxation factor

0.7

1.0

0.7

0.95

1.0

Corrector limit = 20 Corrector step tolerance = 0.25 These settings should, all being well, result in near-monotonic decrease in the global residuals during the course of the calculations, depending on mesh density and other factors. If, thereafter, one or more of the global residuals R φ do not fall, then remedial measures will be necessary. In some instances, the offending variable(s) can be identified from the behaviour of the global residuals. The main remedies now available are: •

•

•

•

Version 4.02

Reduction in relaxation factor(s) — this should be done in decrements of between 0.05 and 0.10 and should be applied to the velocities if the momentum and/or mass residuals are at fault. Decrease in solver tolerances — as in the transient case, this may prove beneficial, especially in respect of the pressure tolerance and its importance to the flow solution. A twofold reduction should indicate whether this measure will work. Increase in sweep limits — if warning messages about the limits being reached appear and are not suppressed by measures 1 and 2, then it may be worthwhile increasing the limit(s) on the offending variables. Under-relaxation of density and effective viscosity — use of this method for density can be advantageous where significant variations occur, e.g. compressible flows, combustion, and mixing of dissimilar gases. Effective viscosity oscillations can arise in turbulent flow and non-Newtonian fluid flow and can be similarly damped by this device. 1-15


Chapter 1


Steady-state flow calculations with SIMPLE As noted previously, the control parameters available for SIMPLE are similar to those for PISO, except that, in the case of the former, a single corrector stage is always used and pressure is under-relaxed. The standard (default) settings are given in Table 1-3. .

Table 1-3: Standard Control Parameter Settings for Steady SIMPLE Calculations Variable Parameter Velocity

Pressure

Turbulence

Enthalpy

Mass fraction

Solver tolerance

0.1

0.05

0.1

0.1

0.1

Sweep limit

100

1000

100

100

100

Relaxation factor

0.7

0.3

0.7

0.95

1.0

In the event of failure to obtain solutions with the standard values, then the measures to be taken are essentially the same as those for iterative PISO, given in the previous section. However, here, reduction of the pressure relaxation factor is an additional device for overcoming convergence problems. The problems usually arise either from a highly distorted mesh, or from highly complex physics (many variables affecting each other). If the grid is distorted, one should reduce the relaxation factor for pressure from the beginning of the run (e.g. to 0.1). If convergence problems are still encountered, a substantial reduction of the under-relaxation factor for velocities and turbulence model variables should be tried (e.g. to 0.5). If this does not help, the problem may lie in severe mesh defects or errors in the set-up. Further reduction of under-relaxation factors may be tried if the grid is severely distorted and cannot be improved; otherwise, improving the mesh quality can be of much greater help. Note that the pressure under-relaxation factor needs to be adjusted within the limits of some range to make the iteration process converge, where the number of iterations required to reach such convergence is mainly dictated by the corresponding factors for velocities (and for scalar variables when strongly coupled to the flow). In the case of well-behaved flows and reasonable meshes, the relaxation factor for pressure can be selected as (1.0 - relaxation factor for velocities), e.g. 0.2 for pressure and 0.8 for velocities. Usually, for a given velocity relaxation factor, the one for pressure can be varied within some range without affecting the total number of iterations and computing time, but outside this range the iterative process would diverge. The lower the relaxation factor for velocities, the wider the range of pressure relaxation factors that can be used (e.g. between 0.05 and 0.8 if the velocity factor is low, say around 0.5). On the other hand, this range becomes narrower when the mesh is distorted. The limit to which the velocity relaxation factor can be increased is both problem- and mesh-dependent. When many similar problems need to be solved, it is worth trying to work near the optimum as this may save a lot of computing time. 1-16

Version 4.02

Chapter 1


On the other hand, for an one-off analysis, it may be more efficient to use a conservative setting. Note that under some conditions, such as those in Tutorial 13.1, a steady-state solution cannot be achieved due to the inherent unsteady character of the flow. This is often the case when the problem geometry possesses some form of symmetry but the Reynolds (or another equivalent) number is high and recirculation zones are present. In this case the residuals stop falling at some level and then continue to oscillate. The “solution” at that stage may be far from a valid solution of the governing equations and should not be interpreted as such unless the residual level is sufficiently small. An eddy-viscosity turbulence model (such as the standard k-e) combined with a first-order upwind scheme for convective fluxes may produce a steady-state solution, while a less diffusive turbulence model (such as Reynolds Stress and non-linear eddy-viscosity models) combined with a higher-order differencing scheme (such as central differencing) may not. In such cases, a transient simulation should be performed; the unsteady solution may oscillate around a mean steady state, in which case the quantities of interest (drag, lift, heat transfer coefficient, pressure drop, etc.) can be averaged over several oscillation periods. Transient flow calculations with SIMPLE The use of this algorithm in transient calculations essentially consists of repeating the steady-state SIMPLE calculations for each prescribed time step. The default control parameter settings are therefore as summarised in Table 1-4. .

Table 1-4: Standard Control Parameter Settings for Transient SIMPLE Calculations Variable Parameter Velocity

Pressure

Turbulence

Enthalpy

Mass fraction

Solver tolerance

0.1

0.05

0.1

0.1

0.1

Sweep limit

100

1000

100

100

100

Relaxation factor

0.9

0.3

0.7

1.0

1.0

Outer iteration limit = 5 The main difference compared to the PISO algorithm lies in the fact that all linearizations and deferred correctors are updated within the outer iterations, by recalculating the coefficient matrix and source term. For this reason, solver tolerances do not need to be as tight as for PISO; they are actually identical to those used for steady-state computations. However, since the discretization of the transient term enlarges the central coefficient of the matrix in the same way as under-relaxation does, the relaxation factors for velocities and scalar variables can be increased (the smaller the time step, the larger the values that can be used for relaxation factors — 0.95 or even more). The convergence criterion for outer iterations within each time step is by default Version 4.02

1-17


Chapter 1


the same as for steady-state flows. However, the number of outer iterations is also set to a default limit of 10; if substantially more iterations are needed to satisfy the convergence criterion, this is a sign that the time step is too large. In such a case, it is better to reduce the time step rather than allow more outer iterations for a larger time step, because this would lead to a more accurate solution at a comparable cost. On the other hand, if residuals drop below the limit after only a few iterations, one may increase the time step; experience shows that optimum efficiency and accuracy are achieved if 5 to 10 outer iterations per time step are performed. Note also that the reported mass residuals are computed before solving the pressure-correction equation; after this equation is solved and mass fluxes are corrected, the mass residuals are more than an order of magnitude lower. For this reason, one can accept mass residuals being somewhat higher than the convergence criterion when the limiting number of outer iterations is reached, provided that the residuals of all other equations have satisfied the criterion. In some cases, an increase in the under-relaxation factor for pressure (up to 0.8) can lead to a faster reduction of mass residuals. All these considerations are of course problemdependent and if several simulations over a longer period need to be performed, it may prove useful to invest some time in optimizing the relaxation parameters. Sometimes, it is necessary to select smaller time steps in the initial phase of a transient simulation than those at later stages. This is the case, for example, when starting with a fluid at rest and imposing a full-flow rate at the inlet, or full speed of rotation (in the absence of a better initial condition). This is equivalent to a sudden change of boundary conditions at a later time, which would also require that the time step be reduced. Another possibility of avoiding problems with abrupt starts from rest is to ramp the boundary conditions (e.g. a linear increase of velocity from zero to full speed over some period of time). The transient SIMPLE algorithm allows you to select either the default fully-implicit Euler scheme or the three-time-level scheme for temporal discretisation, described in Chapter 4, “Temporal Discretisation” of the Methodology volume. The latter scheme is second-order accurate but is currently applied only to the momentum and continuity equations. It should be chosen when temporal variation of the velocity field is essential, e.g. in the case of a DES/LES type of analysis. While PISO would normally be the preferred choice for the latter, under some circumstances (e.g. the existence of very small cells, poor mesh quality etc.), transient SIMPLE may allow the use of larger time steps than PISO without loss of accuracy. Effect of round-off errors Efforts have been made to minimise the susceptibility of STAR-CD to the effects of machine round-off errors, but problems can sometimes arise when operating in single precision on 32-bit machines. They usually manifest themselves as failure of the iterative solvers to converge or, in extreme cases, in divergence leading to machine overflow. If difficulties are encountered with problems of this kind, then it is clearly advisable to switch to double precision calculations. Instructions on how to do this are provided in the Installation Manual. As a general rule, however, you should try to avoid generating very small values for cell volumes and cell face areas by working with sensible length units. Alternatively, you could re-specify your 1-18

Version 4.02

Chapter 1

COMPUTATIONAL ANALYSIS PRINCIPLES Monitoring the calculations

problem geometry units while preserving relevant non-dimensional quantities such as Re and Gr. Choice of the linear equation solver STAR-CD offers two types of preconditioning of its conjugate gradient linear equations solvers: one which vectorises fully, and the other, which is numerically superior to the first one but vectorises only partially. Therefore, the first one (called ‘vector’ solver) is recommended when the code is run on vector machines (such as Fujitsu and Hitachi computers), and the second one (called ‘scalar’ solver) is recommended if the code is run on scalar machines (such as workstations).

Monitoring the calculations Chapter 5 and the section on “Permanent Output” on page 15-1 give details of the information extracted from the calculations at each iteration or time step and used for monitoring and control purposes. This consists of: •

•

Values of all dependent variables at a user-specified monitoring location. Care should be taken in the choice of location, especially for steady-state calculations. Ideally, it should be in a sensitive region of the flow where the approach to the steady state is likely to be slowest, e.g. a zone of recirculation. In transient flow calculations, the information has a different significance and other criteria for choice of location may apply. For example, a location may be chosen so as to confirm an expected periodic behaviour in the flow variables. The normalised global residuals R φ for all equations solved. Apart from turbulence dissipation rate residuals (see Chapter 7, “Completion tests” in the Methodology volume), these are used to judge the progress and completion of iterative calculations for steady and pseudo-transient solutions. In the early stages of a calculation, the non-linearities and interdependencies of the equations may result in non-monotonic decrease of the residuals. If these oscillations persist after, say, 50 iterations, this may be indicative of problems.

Remember that reduction of the normalised residuals to the prescribed tolerance (λ) is a necessary but not sufficient condition for convergence, for two reasons: 1. The normalisation practices used (see Chapter 7, “Completion tests” in the Methodology volume) may not be appropriate for the application. 2. It is also necessary that the features of interest in the solution should have stabilised to an acceptable degree. If doubts exist in either respect, it is advisable to reduce the tolerance and continue the calculations. It follows from the above discussion that strong reliance is placed on the global residuals to judge the progress and completion of iterative calculations of steady flows. These quantities provide a direct measure of the degree of convergence of the individual equation sets and are therefore useful both for termination tests and for identifying problem areas when convergence is not being achieved.

Version 4.02

1-19


Chapter 1

Model evaluation

Model evaluation Checking the model STAR-CD offers a variety of tools to help assess the accuracy and effectiveness of all aspects of the model building process. In performing the modelling stages discussed previously, the user should therefore take advantage of these facilities and check that: 1. The mesh geometry agrees with what it is supposed to represent. This is greatly facilitated by the built-in graphics capabilities that allow the mesh display to be (a) (b) (c) (d)

rotated, displaced, reduced, enlarged.

This enables the user to look at the mesh from any viewpoint, with the view showing the correct three-dimensional perspective. Frequent mesh displays during the mesh generation stage are very useful for verifying the accuracy of what is being created and are therefore strongly recommended, particularly for complex-geometry problems. It is best if such geometries are subdivided into convenient parts that can be individually meshed and then checked visually. 2. Materials of different physical properties occupy the correct location in the mesh. This can be checked visually by using the built-in colour differentiation scheme. Alternatively, each material’s mesh domain can be plotted individually. Precise values of specified properties can be checked via the screen printout. 3. Boundary conditions are correct, by producing special mesh views that show (a) boundary location, (b) boundary type, (c) a schematic of the conditions applied (e.g. inlet velocities). More complete information on specified boundary values can be obtained from the screen printout. 4. The initial conditions should also be checked, particularly for transient problems and initial fields specified through user subroutines, by running the STAR-CD solver for zero iterations/time steps and plotting the relevant field variables. Checking the calculations Having completed the model preparation, the next task is to run the STAR-CD solver and check the results of the numerical calculations. These results are presented in various ways, details of which are given in the Post-Processing User Guide. Briefly, printouts and/or plots can be produced of the following: • •

1-20

Field values of all primary variables at interior and boundary nodes. Interpolated values of the above quantities at arbitrary, user-specified points or surfaces within the solution domain. Version 4.02

Chapter 1

COMPUTATIONAL ANALYSIS PRINCIPLES Model evaluation

• •

Surface heat and mass transfer coefficients and forces; also values of the dimensionless coordinate y+ for near-wall mesh nodes. Global quantities such as total force components (e.g. drag, lift) on submerged bodies and their dimensionless counterparts, overall energy balances, etc.

It is important to examine this information carefully to verify that the calculations have been properly set up and are producing sensible results. In particular, the user should ensure that: • •

•

The interior fields are examined for plausibility and similar checks made on global quantities. For turbulent flow calculations, the near-wall node y+ values are within the recommended range (30-100) in regions where adherence to this constraint is important. In the case of calculations with a two-layer model, checks should be made that the mesh is sufficiently dense within the near-wall layer. The magnitude of numerical discretisation errors (spatial and, where relevant, temporal) is assessed and arrangements made for their reduction to acceptable levels, if necessary.

Of the above tasks, the last is currently the most difficult, for it is not possible to achieve it by a simple calculation. What is required are the following: •

•

A reliable means of evaluating the discretisation errors. At present, this is accomplished by repeating the calculations with finer meshes and smaller time steps (strictly, these should be done independently) and noting regions of appreciable change in the solution. Strategies for altering the mesh or time step to reduce errors. These adjustments are made manually.

Ideally, the error correction process should continue until the changes fall to acceptable levels. In practice, this approach may not be feasible, especially for three-dimensional problems involving complex geometries, due to the large preparation and computing overheads. An alternative way of gaining some insight into the presence of spatial truncation errors is to change the spatial discretisation scheme and note the effect on the solution. The second-order options, or blends thereof, available in STAR-CD will usually produce the lowest numerical errors.

Version 4.02

1-21

Chapter 2

BASIC STAR-CD FEATURES Introduction

Chapter 2

BASIC STAR-CD FEATURES

Introduction The main aim of this part of the manual is to provide users, whether experienced or not in the application of general-purpose computational continuum mechanics codes, with advice on effective ways of setting up and running a basic continuum mechanics model using STAR-CD. The reader is, however, expected to have gone through Chapter 1 and the material in the Methodology volume. All aspects of user interaction are handled by pro-STAR, the pre- and post-processing subsystem of the STAR-CD suite. As a pre-processor, pro-STAR is the means by which the user defines the • • • • • •

geometry, calculation mesh, boundary conditions, initial conditions, fluid and solid material properties, analysis controls,

which uniquely determine the problem to be solved. As a post-processor, pro-STAR can • • • • • • •

read and re-format the various data files produced by the analysis, manipulate the data read in, produce extensive and easily comprehensible printouts, summarise information on the calculated results, draw sophisticated 3-D graphical images, animate those images, draw graphs of various calculated quantities.

Both pre- and post-processing operations are served by an extensive set of plotting facilities, enabling rapid visualisation of even the largest models, plus on-line context sensitive help that provides detailed information on usage. pro-STAR is a combined command-, menu-, and process panel-driven program. The choice of working interface is entirely up to the user and depends on • • •

whether the available terminal can accept and display graphical input and output, whether the host computer’s operating system supports a windowed, graphical user interface (GUI) environment, user preference and level of experience with STAR-CD.

GUI facilities are available for UNIX, Linux or Windows implementations of STAR-CD using the OSF Motif graphics environment. They consist of two basic types: 1. Graphical tools such as drop-down menus, dialog boxes, push-buttons, sliders, etc. to assist users in specifying the desired pro-STAR actions. These facilities are arranged around the main pro-STAR window, or have their starting point located somewhere on that window. Their purpose and best way of using them are explained throughout this volume. Version 4.02

2-1


Chapter 2

Running a STAR-CD Analysis

2. A series of process-oriented panels contained within the STAR GUIde window. These represent an additional GUI facility, suitable for building STAR-CD models from scratch. An outline description is given in the section entitled “The STAR GUIde Environment” on page 2-38. Information on how to use this environment is provided by an on-line Help system accessed from within the STAR GUIde window. Note, however, that: •

•

•

In the present release, a number of pro-STAR facilities are not accessible via either of the GUI systems. Where this is the case, the discussion is in terms of commands rather than GUI operations. For the convenience of users who prefer to work with commands, the description of every GUI panel and dialog box also includes a list of commands that have equivalent functionality. A summary of all pro-STAR commands is given in Appendix B of the Commands volume. A summary of pro-STAR’s conventions regarding command syntax can be found in this volume, Appendix A. The same information is also available on line by choosing Help > pro-STAR Help from the menu bar in the main pro-STAR window and then selecting item PROGRAM (for command syntax) or COMLIST (for command summary) in the scroll list at the bottom of the Help dialog box. Details of all available commands and specific aspects of the command-driven mode of operation are discussed in the Commands volume.

Whichever operating mode is chosen, the same principles of use apply, namely: •

•

• •

A model is constructed or examined with the aid of numerous functions or ‘tools’, each of them represented by a menu-item choice, a special dialog box, a STAR GUIde panel or a command. Tools are selected as necessary, in a sequence that is sensible for modelling purposes. The recommended sequence is described in Chapter 1, “The Basic Modelling Process” and is further elaborated in the Tutorials volume. A tool always provides instant feedback so the user can tell immediately if it was used properly. Users can greatly influence the speed with which certain operations are performed by intelligent use of the available options.

Running a STAR-CD Analysis A STAR-CD analysis may be performed in one of the following two ways: •

•

2-2

By typing a series of script names in a shell or command prompt, each designed to help you build a CFD model, obtain a solution and then display the analysis results. This is the original method of working with STAR-CD and, for reasonably experienced users, may be the quickest way of getting results. By employing a new utility, STAR-Launch, as an aid to navigating through the various STAR-CD functions. This method should be particularly beneficial to novice users. Version 4.02

Chapter 2

BASIC STAR-CD FEATURES Running a STAR-CD Analysis

Using the script-based procedure To perform the analysis using scripts, the procedure described below should be followed in the order indicated: Step 1 Set up an appropriate environment for your STAR-CD system. The desired pro-STAR setup is defined by a number of environment variables such as: STARUSR —

path to the location of files PRODEFS (for command abbreviations) and PROINIT (for pro-STAR initialisation) — see Chapter 16, “Set-up Files” MACRO_LOCAL and MACRO_GLOBAL — paths to the local and global macro locations (see Chapter 16, “Macros”) PANEL_LOCAL and PANEL_GLOBAL — paths to the local and global user-defined panel locations (see Chapter 16, “Panel definition files”) TMPDIR — path to the location of pro-STAR’s temporary (scratch) files Further instructions on how to set the STAR-CD environment variables are given in the Installation and Systems Guide, supplied with the STAR-CD installation CD-ROM. Note that these settings can usually be made once and for all, at the time when STAR-CD is first installed on your computer. Step 2 Create a separate subdirectory for each case to be analysed and give it a descriptive name. This helps to organise the various files created during a run and makes it much easier to check or repeat previous work. Step 3 Move to the appropriate subdirectory and start a pre-processing (model building) session by typing: prostar The system will respond by prompting you to define the pro-STAR variant you wish to use Please enter the required graphics driver Available drivers are: x, xm, glm, mesa [xm] where the options refer to the various types of graphics libraries commonly used for graphical displays in workstations or X-terminals, i.e.

Version 4.02

x xm

— —

glm

—

mesa

—

X-windows X-windows using the Motif interface for pro-STAR’s GUI functions Motif interface plus the standard OpenGL libraries for pro-STAR’s GUI functions As above but using the Mesa OpenGL library (this option is not available in Windows ports) 2-3


Chapter 2


The precise list of options displayed by the prompt depends on how the pro-STAR environment was originally set up on your particular machine. Type in a response that is appropriate to the workstation or terminal you are using. Note that pro-STAR automatically searches for the highest depth pseudo colour, direct colour or true colour visual that exists for your screen and uses it. This may be overridden by specifying option -c when starting up pro-STAR, as shown below: prostar -c This is an 8-bit pseudo colour setting with shared colour map. The setting causes no screen flashing but requires sufficient available colours to work. Once the desired pro-STAR variant has been chosen, an introductory panel opens up leading you into STAR-CD’s model-building environment, as discussed in the section on “pro-STAR Initialisation”. From that point on, you may provide input for setting up your model according to the descriptions given in the remaining chapters of this manual. Step 4 When you have finished setting up your STAR-CD model, it is advisable to check the files created so far in your working directory. These should include: • •

• •

File .mdl, containing all user-supplied information about the model File .ccm, containing a full description of the model geometry. The STAR-CD solver operates only in SI units and all dimensions must therefore be defined in metres. However, it is possible to scale the mesh dimensions by a scaling factor if non-SI units were used during mesh generation. File .prob, containing problem data, such as material properties, boundary conditions, control parameters, etc. File .echo, containing a log (echo) of all instructions issued to pro-STAR during the session

Depending on the nature of your problem (e.g. whether it requires a special modelling facility such as Lagrangian multi-phase) additional files may be created. These are discussed fully in individual chapters of this volume dealing with such topics. A detailed description of all commonly used data files is given in Chapter 17, “Commonly used files”. Step 5 If user-defined subroutines are not required, go to Step 6. Otherwise, create a subdirectory called ufile and place your subroutine files in it. The most convenient way of doing this is to create both the subdirectory and the files from within the pro-STAR session (see Chapter 14, “Subroutine Usage”). Note that these files contain default (dummy) code to start with and you should edit them as necessary to insert your own code. Step 6 Based on the geometrical and physical data of the model just created, you are now in a position to run STAR. This may be done in one of the following ways: 1. Via STAR-GUIde’s “Run Analysis Interactively” panel. Examples of using this panel are provided in the Tutorials volume. This way, the STAR 2-4

Version 4.02

Chapter 2


executable will be run automatically and the analysis results (in terms of solution residuals) will be displayed on a separate window; see “The StarWatch Utility” on page 17-15 for more details. 2. By exiting from pro-STAR and then running STAR from your session’s shell or command prompt. For a large number of cases, it will be sufficient to type one command. For a single-precision run, type: star whereas for a double-precision run, type: star -dp Please note that it is not necessary to provide the case name of the model you are running. However, for better bookkeeping, it is still important to keep every case in its own directory. In most cases, and based on the model characteristics specified in pro-STAR, STAR automatically recognises the default run-time requirements and proceeds with the CCM analysis without further user input. Some cases, however, require the specification of additional options related to both run-time resources and/or behaviour. Briefly, the user can control the operational behaviour of STAR in one of the following areas: • • • • • •

Job precision (single or double precision) Job control (to abort, kill or restart a job) Environment (to export environment variables) User coding (to control the compilation and/or linking of user-supplied code) Parallel setup (pertaining to domain decomposition variations, data distribution and parallel communication libraries) Resource allocation (to choose which machines to use)

A full list of such options can be obtained by typing: star -h or star -help The listing will also contain a short description of each option’s purpose. A more complete description can be found in Appendix F of this manual. Please note that, in general, one needs to specify the machine (node) resources for running STAR and this input is automatically used to determine the type of run required. The following examples illustrate this point:

Version 4.02

star

Runs sequentially on the local node

star origin

Runs sequentially on a host called origin

star 4

Runs in parallel with 4 processes on the local node

2-5


Chapter 2


star origin,16

Runs in parallel with 16 processes on a host called origin

star cheese,2 pickle,2 curry,2 rice,2 Runs in parallel on a cluster of 4 machines with 2 processes each Please note that, for parallel cases, the computational domain decomposition is automatically handled by the star front-end script. The output files generated during the course of the run will be merged and placed in the case’s directory. There is then no visible difference between running in sequential and running in parallel. Extra options exist to cater for special situations which cannot be detected automatically. Please refer to Appendix F for a list of such options, their syntax and their intended purpose. Step 7 Once the run starts, iteration or time-marching continues until one of the following conditions is met: • • •

All the iterations or time steps specified for the current run have been completed. The normalised residual sum drops below a specified value (steady-state runs only). The solution starts to diverge. This occurs when a residual anywhere inside the solution domain reaches a very high value or a numeric overflow condition. Divergence is automatically detected by STAR, which then stops the calculations and writes a file with extension .div. This is identical in format and content to a normal solution data (.ccm) file and thus enables you to inspect the residuals and identify the mesh location(s) where numerical instability has occurred.

Check the condition under which your run has terminated. The parameters involved in controlling the STAR-CD simulation are set in pro-STAR using the facilities provided by the “Analysis Controls” folder in STAR GUIde. Additional information, such as printout of input data, boundary conditions, residual histories of the inner iterative loops, etc. can also be generated, as described in Chapter 15. Step 8 At the start of the analysis, STAR will read the following files: • •

case.ccm — geometry data (plus solution data for restart runs) case.prob — problem data

and, optionally, one or more problem-dependent files such as • • •

case.vfs — view factors for radiation problems case.evn — transient event data case.drp — droplet data

On completion of the run, file case.ccm will contain the current analysis results in a form suitable for post-processing or for starting another STAR run. A number of additional files will also be present in your working directory, including: 2-6

Version 4.02

Chapter 2


• • •

case.run — summary of input data plus numerical statistics and (optional) printout of solution variables case.info — STAR warning messages and (optional) additional numerical statistics case.rsi — Solution residuals, in a form that can be displayed graphically.

The above is the minimum number of output files created by a STAR run and you should confirm that they are all present. Additional files may appear depending on the nature of the problem. Such cases are discussed and explained individually in the relevant chapters of this volume. A description of all commonly used output files appears in Chapter 17, “Commonly used files”. Note that, at the beginning of every restart run, all current results files (such as the ones listed above) are automatically saved in a local sub-directory called RESULTS.xxx, where xxx stands for the run number. These sub-directories thus contain results obtained at the end of each successive run and are available for future inspection, or as a backup in case the restart run’s files are corrupted. If the case is subsequently run from initial conditions, the results of the last run performed are stored in sub-directory RESULTS.000 and all other RESULTS directories deleted. The process then repeats itself with the creation of a new RESULTS directory for each new restart. Step 9 You should now check the results of the analysis by looking at the run history (.run) file (see Chapter 15 for more information on its contents). The additional information (.info) file should also be examined for any signs of numerical problems. These are normally translated into warning messages. Both these files may be inspected via a suitable text editor or via panel “Run History of a Previous Analysis” in STAR-GUIde. Satisfactory completion of steady-state STAR runs can usually be judged by observing the following quantities: • •

The residual history printed during the run. The sum of the normalised absolute residuals should diminish steadily. The monitoring values of the dependent variables at a critical location within the solution domain. These should stabilise to the converged solution.

In transient calculations, completion is defined in terms of the elapsed (simulation) time or establishment of a steady state. In the latter case, information on the global change and monitoring values can be used in the same way as for a steady state analysis. It is important that checks are made regularly during the initial stages of the analysis to monitor the solution progress. If divergence occurs, the run should be terminated and appropriate adjustments made to the relevant control parameters such as under-relaxation factors. Neglecting this can result in costly and unproductive runs. Note, however, that increases in residuals and oscillations in the computed variables during the early stages of a run are not uncommon and should disappear after a few iterations. The run should therefore be given sufficient time to stabilise before any judgement is made on its progress. Step 10 Continue with an evaluation of the simulation results (post-processing) using the Version 4.02

2-7


Chapter 2


relevant facilities in STAR-GUIde. If you have previously exited from pro-STAR and run STAR separately (see Step 6 above), continue by typing prostar to re-enter pro-STAR. Reply as before to the initial prompt Please enter the required graphics driver Available drivers are: x, xm, glm, mesa [xm] and then supply the case name and other input, as described in Step 3. Using STAR-Launch STAR-Launch is a graphical interface that provides access to most of the CDadapco modelling tools, including pro-STAR, several es-tools and the STAR solver. Using STAR-Launch eliminates the need to enter multiple script names manually, as described in the previous section, and also ensures settings can be saved between sessions and between cases. STAR-Launch is intended to be used with only one case at a time. There is, however, no limit on the number of STAR-Launch windows that can be active simultaneously. Activating STAR-Launch On Unix/Linux Either double-click the appropriate icon on your desktop (for systems which support this), or else type starlaunch & in an appropriate X-terminal window. This will display the STAR-Launch main window shown below:

On Windows Double-click the appropriate icon on your desktop. Window layout The key parts of the STAR-Launch main window are highlighted below. The 2-8

Version 4.02

Chapter 2


Shortcut Buttons provide quick access to the three main functions of STAR-Launch, namely: • • •

Setting the working directory Launching a pre-/post-processing tool Running the STAR solver

These functions are also accessible through the Main Menubar running along the top of the window. The current working directory is displayed to the right of the Shortcut Buttons. This is the directory that will be used when launching a pre-/post-processing tool or running the STAR solver. Shortcut Buttons Main Menubar

Current Working Directory

Workspace for Process Output

Run STAR Interactively Launch Pre-Post Tool Set Working Directory

Setting the working directory Choose File > Set Working Directory or click the first shortcut button on the main window. This will display a directory browser as follows:

Version 4.02

2-9


Chapter 2


Navigate to the desired directory and click OK. Note that a path can be entered manually in the Look In entry box at the top of the browser window. The directory tree will be updated to reflect any valid path entered here. The path that will be set on clicking OK is shown along the base of the browser window. Starting a pre-/post-processing tool To start a pro-STAR session, or an equivalent pre-/post-processing tool, select the appropriate entry in the Pre-Post menu, or click the second shortcut button on the main window. The tool that will be started from this button is set using the Pre-Post tab of the Preferences dialog. Only tools available in the current installation will be listed in the Pre-Post menu. STAR-Launch will open a new Process Output window as shown below, which will contain any text generated by the Pre-/Post-processing tool as it starts up.

The STAR-Launch window can be resized as necessary to display more of the text appearing in the Process Output window. Only one pre-/post-processing tool can be running at any one time. If an attempt is made to start another one, a prompt will appear asking if the existing tool should 2-10

Version 4.02

Chapter 2


be closed. Choosing Yes will kill the existing process, which could result in loss of any unsaved data. When a process is active, the ball appearing in the Process Output window tab will be shaded red. This will change to black when the process is finished. Running STAR interactively Selecting Solver > Run Star Interactively, or clicking the third shortcut button, will display the Run Star Interactively dialog shown below. The dialog provides several options for running the STAR solver; detailed information on these options can be found in the STAR-Launch On-line Help, accessed from Help > Online Manual. When all settings have been made, the solver is started by clicking Run. STAR output will appear in a new Process Output window, similar to the one shown above for the Pre-/Post-processing tool. When the STAR solver finishes, the ball on the tab of the output window will turn black. Note that only one STAR solver can be run at any one time from a STAR-Launch session. If multiple solver processes are required, more STAR-Launch sessions must be opened.

STAR-Launch project files .starlaunch directory and launcherGlobal.xml When STAR-Launch is first used, it will attempt to create a hidden directory, .starlaunch in the users home directory (as given by $HOME). Within this directory, STAR-Launch will write file launcherGlobal.xml. The file is normally written on exit from STAR-Launch and contains details of the last working directory specified by the user. It also stores a flag indicating whether this stored path is to be used automatically in a new session. starProject.xml Another file, starProject.xml, can be written by STAR-Launch if requested by the user. This stores settings from the Preferences and Run Star Interactively dialogs. The various File menu options affecting this are explained below: •

Version 4.02

File > Open Project — This presents a file browser that should be used to find the required starProject.xml file. STAR-Launch will be updated to 2-11


Chapter 2

pro-STAR Initialisation

•

•

reflect settings from the new file. File > Save Project — This will write a starProject.xml file in the current working directory. Settings within the file will reflect the current state of STAR-Launch. File > Save As Default — This will write a starProject.xml file in the hidden .starlaunch directory within the user’s home directory.

STAR-Launch start-up procedure When STAR-Launch is first started, it will look for the launcherGlobal.xml file in the hidden .starlaunch directory. This will be read to determine the initial working directory. If a starProject.xml file is also contained in the hidden .starlaunch directory, STAR-Launch will read all settings within the file, and use these to configure the initial state of the GUI. If a starProject.xml file is also found within the initial working directory, STAR-Launch will read the settings within that file, and use these to update the initial state of the GUI. Settings contained in a local starProject.xml file (i.e. one within the initial working directory) will always take precedence over settings obtained from a starProject.xml file in the hidden .starlaunch directory. Preferences dialog Selecting File > Preferences... will display the Preferences dialog shown below:

The options contained here are explained fully in the STAR-Launch Online Help (Help > Online Manual). Their state will be saved in the starProject.xml file.

pro-STAR Initialisation Once the basic GUI mode of operation has been chosen (x, xm, glm or mesa, see “Running a STAR-CD Analysis”, Step 3 above, or via the Preferences dialog in STAR-Launch), the introductory panel shown below appears. The following three optional inputs may be provided: 1. The desired case name — star is the default name assigned to the current problem at the start of a pro-STAR session. Overtype this by the correct name in the Case Name text box. Note that: (a) If a model already exists in your present working directory, its name will be picked up automatically by pro-STAR. (b) If you have more than one model, you may choose the desired one by clicking on the file selection icon next to the Case Name box. This 2-12

Version 4.02

Chapter 2

BASIC STAR-CD FEATURES pro-STAR Initialisation

activates a File Selection browser (see page 2-33) that enables you to choose the desired model, stored in a file of form case.mdl 2. The Resume mode — This can be either a restart from an existing model definition, via its corresponding model (.mdl) file or a brand new case. Clear this option if the latter applies. 3. The Append mode — The session’s user input will be appended to an existing log or echo (.echo) file or a new echo file will be created. Clear this option if the latter applies. Refer to the description given in Chapter 17, “Commonly used files” for a definition of pro-STAR’s model and echo files. Click on Continue to display the basic pro-STAR GUI windows or Exit to abort the current session.

Two windows are displayed automatically immediately following the initialisation stage. These are described in the sections entitled “Input/output window” below and “Main window” on page 2-15. Input/output window This window, shown on the next page, consists of the following three sub-windows, in top-to-bottom sequence: 1. Command Output — displays the time and date of the run, plus summary data for the model in hand, if such data were read in from a Restart file at the initialisation stage. All subsequent output in that window are the echo of every instruction issued by the user plus pro-STAR’s response to it. The latter serves as feedback to help determine whether a facility was used properly. 2. Command Input — accepts pro-STAR instructions in the conventional ‘Command keyword plus parameters’ format described in the pro-STAR Commands volume. Thus, it is possible to work in ‘command’ mode at any stage of the model building process despite the fact that the GUI version of the code is active. This is useful when working with facilities that cannot be activated from a GUI panel or dialog box in the present pro-STAR version. This sub-window can be re-sized by dragging the control ‘sash’ (the small square at the top right-hand corner) up and down. 3. Command History — provides a numbered ‘command history’ list that keeps track of all pro-STAR instructions issued in the current session, either as Version 4.02

2-13


Chapter 2


choices from a menu in the main GUI window (see “Main window” on page 2-15) or as commands typed in the sub-window above. Menu choices are translated into their equivalent commands before being added to the list. The list can be used in the following two ways: (a) Single-click the command number to copy a command into the Input window and then edit it. (b) Double-click the command number for immediate re-execution. The Command History sub-window can be re-sized by dragging its control ‘sash’ up and down.

Note that: 1. The Command Input sub-window can accept multiple commands by cutting and pasting from the window of another application (e.g. a text editor). If any of the imported command text needs editing prior to execution, (a) (b) (c) (d)

click the Pause action button under the window (see the above panel) paste in the required group of commands make the necessary changes click the Pause action button again to allow pro-STAR to begin executing the commands one by one 2. The Command History sub-window will normally list all commands issued to pro-STAR, including those generated indirectly via an external command file (see Chapter 17, “Commonly used files”) or a user macro (see Chapter 16, 2-14

Version 4.02

Chapter 2


“Macros”). It will also list details (e.g. coordinate values) of items such as vertices, splines, cell faces, etc. that are directly picked from the main window display with the mouse. Such output may become extremely voluminous and may thus obscure the record of primary operations performed by the user. Clicking the Short Input History button will prevent this and will cause pro-STAR to list only the instructions directly issued by the user. Main window The main GUI window, shown below, is used for the following purposes:

• •

•

For graphical display of various aspects of the current model. As a launch pad for those pro-STAR utilities that are available in GUI form. The user should click one of the eleven drop-down menus appearing in the menu bar and select one of the displayed choices. Commonly used functions affecting the model display in the graphics area are also implemented, in the form of action buttons. These are distributed along the top and left-hand-side borders of the window and are described in Chapter 4 of the Meshing User Guide. Letting the mouse rest on top of any button causes a brief explanatory legend to appear in a special window provided for this purpose. To show messages for the user, such as prompts to supply data, in the space underneath the graphics area. The default display shows: (a) The current plot parameters (see “Plot Characteristics” on page 4-3 of the

Version 4.02

2-15


Chapter 2


Meshing User Guide). (b) A clock display showing the current time and date. This may be turned on or off by selecting Show Clock or Hide Clock from the Utility menu. (c) Three status indicators showing the result of processing the latest command, irrespective of whether it was typed in directly or issued via a GUI operation. The indicators are arranged as a set of ‘traffic’ lights whose significance when lit is as follows: i) Green — the command was executed successfully. The displayed message is Command: is Done ii) Amber — the indicator flashes to signal the presence of warning messages in the Output window. The displayed message is Command: has a Warning, check the output window iii) Red — the command has failed. The displayed message is Command: has an Error. Click on the red light to view the error Clicking on the red light displays an Error/Warning Summary pop-up window with more information on what has gone wrong, as discussed under “Error messages” below. Note that if a GUI operation generates a series of commands, a message is issued for each one in turn as soon as it is processed. If all goes well, the message finally seen on the screen is for the last command that was executed. The menu bar The menu bar items are listed below, along with a reference to chapters containing a detailed description of their functionality: 1. File Provides all basic housekeeping utilities, including those related to input/output operations — see Chapter 17, “File Handling”. 2. Tools Activates dialog boxes that allow definition and manipulation of basic pro-STAR entities (cells, vertices, splines, etc.). Most of these are covered in Chapter 2 of the Meshing User Guide. Another type of tool facilitates routinely-used, complex operations such as colour selection and mesh surface lighting effects (see Chapter 4, “Colour settings” in the Meshing User Guide). 3. Lists Displays lists of all available entities of a certain type (cells, vertices, boundaries, etc.) as well as those currently grouped into a user-defined set. 4. Modules Accesses special dialog boxes that set up various STAR-CD model 2-16

Version 4.02

Chapter 2


parameters in connection with (a) Animation control — see Chapter 12 of the Post-Processing User Guide (b) Transient condition definition — see Chapter 5, “Load-step based solution mode” 5. Plot Contains most of the facilities and options used for mesh plotting operations — see Chapter 4 of the Meshing User Guide. 6. Post Displays the results of a STAR run — see Chapter 1 of the Post-Processing User Guide. 7. Graph Produces various types of graph — see Chapter 14 of the Post-Processing User Guide. 8. Utility Provides miscellaneous utility functions designed to aid model control and development, such as calculation of cell volumes and distance between vertices — see Chapter 3, “Mesh and Geometry Checking”in the Meshing User Guide. It also supports special user-controlled operations, such as the assignment of user-defined functions to keyboard keys. 9. Panels Allows you to set up your own screen buttons or panel tools for performing common pro-STAR operations — see Chapter 16. 10. Favorites (optional) This menu appears only if you have chosen any ‘favourite’ (i.e. frequently used) panels in the STAR GUIde tree structure (see “Panel navigation system”). The relevant panels are listed under this menu, enabling you to jump to them directly. 11. Help Displays pro-STAR command help information in a scrolled-text fashion. Also contains on-line versions of the STAR-CD manuals and tutorials. A mouse click on any of the above menu names displays a drop-down list. In general, clicking an item on the list starts up the action indicated, unless the name is followed by • •

an ellipsis (…) which means the item displays a new dialog box, or an arrow (⇒) which means the item opens a secondary list with more items to choose from.

Throughout this manual, the “>” sign denotes successive mouse clicks on menu names, menu list items, dialog box buttons, etc. For example, Tools > Cell Tool > Edit Types means click Tools in the menu bar, then click the Cell Tool item in the drop-down list, then click the Edit Types button on the displayed Cell Tool dialog.

Version 4.02

2-17


Chapter 2

General Housekeeping and Session Control

General Housekeeping and Session Control When pro-STAR is initially installed on a computer system, default settings are provided for the program’s fundamental operating features. These settings, specified mostly via commands typed in the Command Input window, can be altered in special circumstances. The following aspects of the program’s operation are covered: Basic set-up These settings are helpful in establishing an appropriate environment for pro-STAR and for accessing facilities related to the operating system of the host machine. They are as follows: 1. Operating mode — command BATCH disables pro-STAR’s periodic prompts to stop or continue displaying long lists of data. 2. pro-STAR size — command SIZE lists the maximum number of cells, vertices, boundaries, etc. that the code can handle. If any of these values is inadequate for the model in hand, it may be increased by following the procedure described in Chapter 17, “Resizing pro-STAR”. 3. Reporting cpu time required to complete a pro-STAR function by typing command TPRINT. 4. Accessing special, user-written pro-STAR subroutines by typing command USER. It is advisable to use this facility only after consultation with CD-adapco. 5. Communicating with the operating system itself. This may be done by first choosing File > System Command from the menu bar to display the System Command dialog box shown below and then typing system commands in its text box.

Command:

SYSTEM

This is useful for issuing instructions to the host operating system without having to exit from the pro-STAR environment. Screen display control There are several facilities for controlling the screen display during a session, as follows: •

•

2-18

Defining the layout and look of the pro-STAR windows. Default settings are normally used for these but the user can override them at will, as explained in Chapter 16, “Set-up Files” and also in Appendix D. Switching from the terminal’s graphics screen to the text screen via command TEXT. This is applicable only when running a non-GUI version of pro-STAR Version 4.02

Chapter 2

BASIC STAR-CD FEATURES General Housekeeping and Session Control

• •

• •

•

and is used for controlling terminals that operate entirely either in text or in graphics mode. Setting the number of lines that appear on each ‘page’ of the Command Output window during lengthy listings using command PAGE. Displaying a history of the most recent commands issued during the session via command HISTORY. Again, this applies only when running non-GUI versions of pro-STAR since these do not provide a command history window. Echoing the user input stream to the same device as the output stream (e.g. the screen or a disk file) via command ECHOINPUT. Reading stored cursor picks from an input file, rather than displaying a crosshair cursor and reading the user-specified picks off the screen — command CURSORMODE. Providing a descriptive title for the current model that helps to identify each plot produced subsequently — choose File > Model Title from the menu bar to display the dialog box shown below. The desired title and up to two lines of subtitle text should be typed in the text boxes provided.

Command:

TITLE

Error messages pro-STAR issues error messages as a result of receiving incorrect commands or if it is unable to execute a valid command for whatever reason. Such messages appear in three places: • • •

Version 4.02

On the standard Output window At the bottom of the main pro-STAR window, after the red indicator light (see page 2-16) On the Error/Warning Summary pop-up panel, as in the example shown below:

2-19


Chapter 2

General Housekeeping and Session Control

The panel shows a list of all current errors, including their error id and command in which the error occurred. If you know the cause of the problem, click Clear to close the panel. Otherwise, select any item in the list to see the error description at the bottom of the panel. Command SUCCEED, typed on its own, tells you (in the output window) whether the previous command produced any errors or warnings. On the other hand, typed as SUCCEED, QUIT, it will immediately terminate the pro-STAR session. This can be useful when pro-STAR is run in batch mode, where it is not desirable for the job to continue after an error. Error recovery If mistakes are made during a session, the following operations are useful for error recovery: •

•

•

2-20

Re-executing a named range of previously issued commands by typing command RECALL. This can be most conveniently used in conjunction with the HISTORY command above. Retrieving the state of the model description as it was at the time of the previous SAVE or RESUME operation — command RECOVER. This is useful if a mistake is made but the user does not notice it until some time later. A list of commands issued since the last SAVE or RESUME operation is displayed, along with a prompt to choose the last command in the list to re-execute. The chosen command will normally precede the one where the mistake was made. Once all commands up to that point are re-executed, the user should type in a correct command and carry on from there. Note that the above safety features can be switched off using command SAFETY. This might speed up pro-STAR execution but at the potential cost of making any sort of recovery from mistakes nearly impossible. Thus, turning Version 4.02

Chapter 2

BASIC STAR-CD FEATURES Set Manipulation

off these features should be used with extreme caution. Session termination The current pro-STAR session is terminated by choosing File > Quit from the menu bar. This displays the Quit pro-STAR dialog box shown below, reminding you to save the results of the session to a .mdl file (in case this has not already been done explicitly). Alternatively, you may deliberately exit from pro-STAR without saving the present session’s work, by clicking Quit, Nosave.

Command:

QUIT

Set Manipulation pro-STAR has extensive facilities for collecting and modifying sets of objects. These are accessible by clicking one of the coloured buttons down the left-hand side of the main window. The pro-STAR entities serviced by the buttons are: • • • • • • •

C-> — cell sets V-> — vertex sets S-> — spline sets Bk-> — mesh block sets B-> — boundary sets Cp-> — couple sets D-> — droplet sets

Each button offers a wide range of possibilities to select, delete or re-select sets. For example, selection may be done by picking all objects falling within a given geometric range in a local coordinate system. Using other criteria, one can collect together all cells or boundaries connected to the current vertex set (and the reverse). Selection can also take place by simply using the screen cursor to point to items on the current plot. Each button gives direct access to the following set manipulation options: • • • • • • Version 4.02

All — select the entire set None — empty out the current set Invert — invert the current set, i.e. select all entities that are not currently selected and un-select the ones that are New — replace the current set with a new set, formed on the basis of a criterion given in a secondary drop-down list Add — add more members to the current set, selected using one of the criteria in the secondary drop-down list Unselect — remove some members from the current set, selected using one 2-21


Chapter 2

Set Manipulation

•

of the criteria in the secondary drop-down list Subset — select a smaller group of members from those in the current set, selected using one of the criteria in the secondary drop-down list

In addition, C-> offers one extra option, Surface, which selects all cells lying on the surface of the most recent mesh plot and makes them the current set. Further details on the above set selection options are given in Chapter 2 of the Meshing User Guide, for each of the mesh entities described there. Note that it is possible to save and restore useful cell, vertex, spline, block, boundary and couple sets without the need to rebuild them frequently. This is done by clicking the INFO button at the left-hand side of the main pro-STAR window. The following operations are possible: 1. To perform a ‘save set’ operation, select INFO > Store Set/Surface/View and then click the Sets tab to display the dialog shown below:

Commands: SETWRITE

SETDELETE

The input required is as follows: (a) Set File — The name of the set (.set) file that will store the set definition. If such a file already exists, pro-STAR’s built-in file browser may be used to help locate it. (b) Name — An identifier for the set being saved, up to 80 characters long Click Write to save the set definition. 2. To delete a set definition previously stored, use the same dialog as above and specify the following information: (a) Set File — The name of the set (.set) file containing the definition to be deleted. pro-STAR’s built-in file browser may be used to locate it. (b) Select Entry — The location of the set to be deleted, as select from the list. Click Delete to delete the set definition. 3. To perform a ‘restore set’ operation, select INFO > Recall Set/Surface/View 2-22

Version 4.02

Chapter 2

BASIC STAR-CD FEATURES Set Manipulation

and then click the Sets tab to display the dialog shown overleaf:

Command:

SETREAD

The input required is as follows: (a) Set File — The name of the set (.set) file containing the set definition. pro-STAR’s built-in file browser may be used to help locate it (b) Select Entry — Select the particular set required by name from the scroll list. The status of the selected entry is displayed in the box underneath (c) Choose Data — Specify the type of set to be read in (All, Cells, Vertices, etc.) by clicking one of the displayed option buttons (d) Read Option — Specify how the sets to be read in will modify any existing sets by selecting one of the menu options (Newset, Add, Unselect or Subset) Click Recall to recall the selected set. Note that it is possible to print a summary of all data sets stored so far by typing command FSTAT. Selecting sets of various entities has two major uses: 1. To display only items in the currently active set. For example, each time Cell plot is chosen from the Plot menu, pro-STAR plots only cells in the currently active cell set. Note that command SETADD causes all newly-defined cells to be automatically added to the current set. Thus, successive plots of the current state of the mesh can be made without needing to build a new set after each new cell definition. SETADD may also be used in the same way for other kinds of sets, i.e. boundaries, cell couples and splines. 2. To perform almost any modelling or post-processing operation on the Version 4.02

2-23


Chapter 2

Table Manipulation

currently active set, instead of on individual objects or a range of them. For example: (a) Choosing Lists > Cells from the menu bar and clicking the Show Cset Only option button will list only cells in the current set. (b) When working with commands, typing VMOD,VSET,2.5 will modify the X-coordinate of every vertex in the current vertex set. All set operations can also be performed by typing commands CSET, VSET, BSET, BLKSET, CPSET, SPLSET and DSET. These are described in detail in the pro-STAR Commands volume.

Table Manipulation pro-STAR tables are multi-variable entities akin to spreadsheets and can be used to store values for up to 100 dependent variables as functions of a combination of several independent variables. For most commonly used tables, the independent variables can be the three spatial coordinates, plus time for transient cases or iteration for steady-state cases. The dependent variables are normally flow field solution variables but, in principle, they could be anything of relevance to STAR-CD. Basic functionality At present, tables are used principally as a substitute for user subroutines in the following situations: •

•

Boundary Conditions — variable conditions along the surface of a boundary region; see Chapter 4, “Boundary Region Definition”, page 4-7. For most boundary types, the independent variables may be any combination of spatial coordinates and, for transient cases, time. The only exception is outlet boundaries where only time is allowed (i.e. there can be no spatial variation in outflow conditions along the outlet surface). The permissible dependent variables vary according to the boundary type considered; a full list is given under the various boundary type descriptions in Chapter 4, or the corresponding on-line Help topics for STAR GUIde’s “Define Boundary Regions” panel. Initial Conditions — non-uniform initial distributions of field variables; see Chapter 4, “Solution Domain Initialisation”. The independent variables may be any combination of spatial coordinates, for both steady and transient cases. The permissible dependent variables for fluid materials are listed under topic “Initialisation”. Scalar variables representing chemical species mass fractions may also be initialised, as described in a separate topic for scalar “Initialisation”. Note that: (a) The applicability of field variable and scalar initialisation tables can be restricted to a selected domain or a cell type (b) The only dependent variable allowed for solid materials is temperature

•

2-24

Source Terms — a description of mass, heat, momentum or scalar species sources; see Chapter 3, page 3-8. The independent variables may be any Version 4.02

Chapter 2

BASIC STAR-CD FEATURES Table Manipulation

combination of spatial coordinates and time for transient cases, or iteration number for steady-state cases. The permissible dependent variables vary according to the source type considered; a full list is given in the on-line Help topics for the various sources definable via panel “Source Terms”. Note that, as with initial conditions, the applicability of source tables can be restricted to a selected domain or a cell type. Rotational Speeds — variable angular velocity in rotating systems, specified in panel “Rotating Reference Frames”. The independent variable is time for transient cases, or iteration number for steady-state cases. The dependent variable is angular velocity, expressed in r.p.m. Run Time Controls — variable time step for transient cases, specified in panel “Set Run Time Controls”. The independent variable is time, the dependent variable the time step size. Note that STAR assumes a linear variation in step size between the size values entered at two consecutive time points. This is illustrated by the example below, showing the desired time step variation and the table structure needed to achieve it:

•

•

0.35 0.3

DT (sec)

0.25 0.2 0.15 0.1 0.05 0.0 0.00

5.00

10.00

15.00

20.00

Time (sec) Figure 2-1

Example of time step variation

Table 2-1: Time step size table

Version 4.02

TIME

DT

0.0

0.01

1.0

0.1

5.0

0.1

5.0

0.2

10.0

0.3

20.0

0.3 2-25


Chapter 2

Table Manipulation

In addition, a special table type is used to enter problem data for Lagrangian Multi-Phase cases. The following two options are available in this category: •

•

Mass Flow Rate — injection rate history, specified in panel “Spray Injection with Atomization” which activates STAR-CD’s built-in spray modelling facilities. The table is used in transient analyses only and contains injector mass flow rates vs. time (see also topic “Define Injectors”). The same table type may also be used in panel “Injection Definition” as part of an explicit specification of injection characteristics. Diameter Distribution Function — a definition of the droplet diameter distribution function, in terms of spray percentage mass vs. droplet diameter. This table may also be specified in panel “Injection Definition”.

The table editor Table data are stored in text files and may be created or modified either via a suitable text editor or via pro-STAR’s own GUI facilities. Both options are accessed by clicking the special table editor button

at the bottom left-hand side of the main window. The basic functionality of the editor is described below. New tables To create a new table, click New Table to display the table view shown below:

2-26

Version 4.02

Chapter 2


Panel Data Entry To use the dialog facilities directly, the following input is required (reading from left to right along the panel): 1. Table Title — enter a title up to 80 characters long, including spaces. Note, however, that only the first 30 characters found up to the first space in the string are usable by STAR. 2. Coordinate System — specify the coordinate system number to be used for spatial independent variables (see “Coordinate Systems” on page 2-8 of the Meshing User Guide). A search button is provided for choosing any of the currently defined systems from the Coordinate Systems dialog. Depending on your selection, the three space coordinates are interpreted as follows: Cartesian

Cylindrical

Spherical

Toroidal

x (X) y (Y) z (Z)

r (R) θ (ΤΗΕΤΑ) z (Z)

r (R) θ (ΤΗΕΤΑ) φ (PHI)

r (R) θ (ΤΗΕΤΑ) φ (PHI)

The coordinate names shown above inside parentheses should be used as table headers when creating a table outside this GUI environment. 3. Out of bound value options — prescribe the action to be taken if needing to calculate dependent variable values at points lying outside the table range. Obviously, this does not apply to mass flow rate tables. The available options are: (a) Error — issue an error message (b) Extrapolate — use the closest two data points to calculate an extrapolated value (c) Cutoff — use the closest data point as the variable value 4. Select Table Type — choose the basic table type from the list of options described under “Basic functionality”. The correct type is selected automatically if you enter the editor indirectly, i.e. by clicking button New in a STAR GUIde panel that requires the use of tables. 5. Select Dependent Variables — for boundary and source tables, select also the specific type of boundary or source required from a secondary menu. All valid variables for the chosen table type are displayed automatically in the adjacent scroll list. To select an item from this list: (a) For single items, click the desired variable (b) For two or more items in sequence, click the first variable, press and hold down the Shift key, then click the last variable in the group (c) For a random selection, hold down the Cntrl key and then click each variable in turn 6. Select Independent Variables — all valid variables for the chosen table type are displayed automatically as a series of option buttons. Choose those needed to define your table by clicking the corresponding button. 7. Click Setup to confirm your selections and enter the data input mode, as shown in the example below. Version 4.02

2-27


Chapter 2

Table Manipulation

Commands: TBDEFINE

TBCLEAR

TBWRITE

TBGRAPH

The following points should be kept in mind when specifying table data: •

•

•

•

Table values should be entered for each dependent variable selected in step 5 above. Your selection will be automatically reflected in the options shown on Dependent Variables scroll box. Fill in all required data for the currently selected variable before scrolling to the next one. The left-hand side of the panel will display a number of columns, one for each independent variable selected in step 6 above. Fill each column with all the values assumed by that variable in the table, in ascending order. Tables containing two or more independent variables are essentially multi-dimensional and need to be specified as a series of two-dimensional x-y tables, as in a spreadsheet. Accordingly, a pair of independent variable values are displayed as row and column headings and the user fills in appropriate values for the current dependent variable, as shown in the example above. To create such two-dimensional tables: (a) Select the required pair from the Independent Variables menu, noting that pro-STAR activates only those combinations that correspond to the choice made in step 6 above. The available pairs for the example shown (an X, Y, TIME selection) will be X - Y, X - TIME and Y - TIME and the pair chosen is X - Y. (b) Fix the other independent variable(s) to a desired value, by clicking the radio button next to that value in its column on the left-hand side. In the example, TIME is fixed to 0.

2-28

Version 4.02

Chapter 2


(c) Click the FILL button. This sets up the 2D table and displays the chosen pair’s values as row and column headings (d) Fill the table with the required dependent variable values and then click Save Data. (e) Fix the other independent variable(s) to a different value and repeat steps (b) to (d) above as many times as necessary (f) Select another pair from the Independent Variables menu and fill in another series of 2D tables. This might happen, for example, if instead of choosing to enter an X - Y set for a series of fixed TIME’s, you chose instead to enter X - TIME sets for fixed Y’s followed by Y - TIME sets for fixed X’s. •

Tables for rotational speeds, run-time controls and Lagrangian multi-phase specifications always have one independent variable and thus involve filling in a two-column table. The same also applies to the other tables if only a single independent variable is specified. A simplified display appears in the editor panel in these cases.

Once your data input is complete, you may: 1. Check the table contents graphically by plotting them as a pro-STAR graph (see Chapter 14 of the Post-Processing User Guide). To use this facility: (a) Select the variable to be checked from the Dependent Variables scroll box. This will be plotted along the graph’s y-axis. (b) Go to the graph setup section at the bottom of the panel (which now displays the chosen variable) and select an independent variable from the versus scroll box. This will be plotted along the graph’s x-axis. (c) The names of the remaining independent variable(s) will also be displayed in the const boxes. For the purposes of the graph, these will be fixed to the value indicated by the radio button in each variable’s column. These values will also appear inside the @ boxes. (d) Click Graph to see the result of your selection. 2. Save your data in a table file. The file name should have extension .tbl and should be entered in the File Name box at the bottom of the panel. pro-STAR’s built-in browser may also be used to locate an existing file. Click Write Table to save your data in this file. File Data Entry An alternative method of generating a new table is to import existing tabular data from an ASCII file created outside pro-STAR. To use this method: 1. After opening the Table Editor dialog, select the Import button situated under the New Table option. This will display the alternative panel view shown below.

Version 4.02

2-29


Chapter 2

Table Manipulation

2. Click Instruction to open a special text panel giving detailed information on how the user file shoud be structured and formatted. Check that your own file conforms to this standard and modify if necessary. 3. Enter the name of your file in the ASCII Data File Name box, or use pro-STAR’s built-in file browser to help locate it. 4. Select option space or comma from the Delimiter menu to indicate how the numerical values in your table are separated from each other. 5. Click Import to import your data into pro-STAR. 6. Enter the remaining table specification items on the right-hand side of the panel, as described on page 2-27. 7. Check the table contents graphically, if required, and then save them in a pro-STAR table file, as described on page 2-29. Existing table display/modification To read and display the contents of an existing table, click Read Table at the top left-hand side of the editor and then enter the file name (of form case.tbl) in the File Name box. pro-STAR’s built-in browser may be used to help locate the file.

Commands:

TBREAD

TBLIST

TBMODIFY

TBGRAPH

Once the table has been read, its contents can be checked visually using the graph function described in the previous section or modified as required. Note that: 1. You cannot add new dependent or independent variables to an existing table (or delete any that are currently defined) 2-30

Version 4.02

Chapter 2

BASIC STAR-CD FEATURES Plotting Functions

2. You may alter both individual values and the number of such values for any independent variable. Click Save Modified Data to confirm the changes. 3. Changes to existing dependent variable entries are made by over-typing and confirmed by clicking Save Data. 4. At the end of the editing session, you should always save your updated table in a named file by clicking Write Table Useful points 1. Only one table at a time may be loaded into the pro-STAR editor. If you need to access a second table, you must first save the current one to a named file (if you have made changes) before reading in the new one. 2. If you change your mind about the contents of your current table and wish to make drastic change, clicking New Table enables you to erase all entries and start afresh. 3. The scale factor applied when saving model geometry data (see Chapter 17, “Data repository file (.ccm)”) is also applied to table coordinate data when they are accessed by STAR. 4. Apart from the table file itself, table data needed for the next STAR-CD analysis are also stored in the STAR problem file (see Chapter 17, “Problem data file (.prob)”) so that they are available to STAR during the run. The user specifies which tables will be needed as part of the boundary, initial condition or other model specification requiring the use of tables. 5. You may use command TBSCAN to scan a named .tbl file. Information about its contents is displayed in the I/O window.

Plotting Functions Basic set-up The basic hardware-related plotting features are set by a single command, TERMINAL. This command sets: •

•

•

The display mode of X-based terminals (use option ALTERNATE only for improving the plotting speed of certain older types of workstation). This setting may also be accessed from the menu bar by switching between options Plot > Standard Plot Mode and Plot > Alternate Plot Mode. The plot destination — this specifies whether plots are to appear directly on the screen or written to the neutral plot file (see Appendix B in the Post-Processing User Guide). The operating mode of the plotting device — a choice between raster, vector or extended (for high-performance workstations). It is also possible to toggle between raster and extended plot mode by clicking the X / GL button at the bottom left-hand side of the main window. Note, however, that this option is available only if you are working with the glm version of pro-STAR (see “Running a STAR-CD Analysis”, Step 3).

The basic features of devices operating under one of the above modes are: 1. Vector devices, such as pen plotters, can draw lines in one or more colours, but are not generally capable of filling in closed polygons or erasing parts of the plot after drawing in them. When this mode is set: Version 4.02

2-31


Chapter 2

Plotting Functions

(a) All hidden-line plot calculations are done by software. (b) Large amounts of time may be required for large models. (c) All contour plots displayed as line contours rather than filled colours. 2. Raster devices, such as most workstation screens, Postscript laser printers, etc. are capable of filling in polygons quickly and overwriting previously coloured-in regions with new colours. When this mode is set: (a) Hidden-line plots are done by hardware. (b) Contour plots are rendered in filled colours. (c) VECTOR mode operation is still possible if, for example, the user wants fringe-style rather than filled-colour contour plots. 3. Extended mode devices offer additional functionality such as true (24-bit) colour, hardware Z-buffers, double-frame buffering, coordinate transformation pipelines, Gouraud shading, etc. Machines with these high-specification graphics attributes can provide: (a) Real-time rotation, translation and zooming of plots. (b) Contour plots rendered in smoothly varying colour bands. (c) Added lighting effects to enhance a user’s perception of the model geometry. This style of plot is limited to machines that support the OPENGL standard and cannot be stored in the neutral plot file at present. Appendix C in the Post-Processing User Guide lists all currently available combinations of plot mode and plot characteristics. The same information can also be listed on line by choosing Help > pro-STAR Help from the menu bar and then selecting the COMBINAT item from the list shown at the bottom of the pro-STAR Help dialog. Advanced screen control Advanced screen control functions are implemented as follows: •

•

• • •

2-32

Background/foreground colour reversal — from the menu bar, select Plot > Background > Standard (for white lines and text on a black background) or Plot > Background > Reverse (for black lines and text on a white background). Alternatively, use command CLRMODE. Maximising the graphics area — from the menu bar, select Plot > Maximum Plot Screen to hide the GUI buttons surrounding the graphics area so as to make the plot as large as possible. The window is also enlarged to take up almost the entire screen. This is helpful when making animations since the largest number of pixels are used, thereby obtaining the highest possible plotting resolution. Select Plot > Standard Plot Screen to return the window to its default size and appearance. Alternatively, use command WHOLE. Restoration of the original screen settings — command RESET. Temporary, on-line storage of complete screen images — command SCROUT. On-line retrieval of screen images previously stored with SCROUT — command SCRIN. This command also provides an elementary animation facility, by replaying a sequence of screen images in quick succession. Version 4.02

Chapter 2

BASIC STAR-CD FEATURES Plotting Functions

• • •

Deletion of screen images previously stored with SCROUT — command SCRDELETE. Customised scaling of text fonts used in pro-STAR — command TSCALE. Image display control — command PLTBACK. This enables images to be created and stored in memory and then popped onto the screen (as opposed to displaying them as they are being created).

For further details on using the above commands, refer to the pro-STAR Commands volume. Screen capture It is often very useful to be able to save the contents of the graphics screen as a picture file. The latter can then be pasted into a document created by another, say presentation or word-processing, application. pro-STAR provides this facility via the Utility > Capture Screen menu option (or by typing command SCDUMP). The result of this operation is the creation of a new window containing the picture currently displayed in pro-STAR’s main graphics area. The picture can be subsequently saved in a file by choosing Utility > Save Screen As and selecting one of the following options for the file format: • • • •

XWD (X Window Dump) — X-Motif version of pro-STAR only GIF (Graphics Interchange Format) PS (PostScript, either Level 1 or Level 2 format) EPSF (Encapsulated PostScript, either Level 1 or Level 2 format)

The user needs to make sure that the choice of format is appropriate to the end application. Selecting any of the above options opens the File Selection dialog shown below, enabling you to specify the name and destination directory of the picture file.

If you are working in OpenGL extended graphics mode (see page 2-32), you also have a choice of saving a high-resolution screen dump (HRSD) of the extended mode plotting window. This appears as an additional option, High Res. Screen Dump, in the Utility menu (alternatively, use command HRSDUMP). Selecting this Version 4.02

2-33


Chapter 2

Plotting Functions

option from the main menu opens the High Resolution Screen Dump dialog shown below:

The user input is as follows: 1. Select the required file format from the File Type menu as one of (a) (b) (c) (d)

png gif ps (PostScript) eps (Encapsulated PostScript)

2. Enter the file name in the box provided. Clicking the adjacent browser button opens the File Selection dialog shown above which helps locate the required file. 3. Clicking the Options button opens a secondary Image Options dialog that enables you to specify the required image resolution and/or page properties (for PostScript files). An example for GIF/PNG images is shown below.

It is also possible to use the HRSD facility in batch mode to produce high-quality plots using OpenGL style graphics (i.e. including translucency, special lighting effects, etc.). You do not require a special OpenGL graphics card on your machine to do this; the pictures can be made off-screen using the ‘mesa’ software emulation of OpenGL as follows: •

Run pro-STAR with mesa graphics in batch mode prostar mesa -b

•

Set extended mode graphics term,,exte

2-34

Version 4.02

Chapter 2

BASIC STAR-CD FEATURES The Users Tool

•

Set up the model as you wish, including any CPLOT/REPLOT operations needed to display the picture, and then use the HRSD command as follows: hrsd,png,output.png (write a .png file) hrsd,ps,test.ps (write a .ps file)

You can also use the various options to change image size, resolution, etc., just as for interactive mode above (see the HRSD command Help on options for setting image size, resolution, etc.)

The Users Tool The Users Tool enables you to create your own customised user interface, by running a Tcl/Tk script from within pro-STAR by means of a built-in interpreter. To make use of this tool, you need solid knowledge of Tcl/Tk programming. The basic idea is that the user builds a dialog box as he/she would for any other Tcl/Tk-based application, with widget callbacks designed to pass pro-STAR command strings back to pro-STAR (much as it happens now when you click a button in STAR GUIde). An introductory panel, shown below, is provided via the main menu, by choosing Tools > Users Tool. Clicking the left-hand button invokes the built-in interpreter which then runs your script.

To use this facility, it is important to • •

save your Tcl script in a file called STARTkGUI.tcl assign the path to this file to an environment variable called STAR_TCL_SCRIPT

Getting On-line Help The Help menu in the main pro-STAR window is divided into three parts. There are three options in the top part, About pro-STAR, Select Item and pro-STAR Help. Clicking About pro-STAR displays pro-STAR version information, as shown below:

Version 4.02

2-35


Chapter 2

Getting On-line Help

Clicking pro-STAR displays the pro-STAR Help dialog shown below:

This dialog contains on-line information on: • • • • • • • • • • • •

Conventions regarding command line syntax All valid combinations of plot mode and plot characteristics One-line summaries of every pro-STAR command, grouped by command module and listed in alphabetical order A list of all database files available under pro-STAR pro-STAR environment variable definitions All file extensions used A description of pro-STAR’s macro files A description of pro-STAR’s user-defined Motif panels A tabulation of radiation parameters required for walls and baffles Units for all physical quantities used in STAR-CD A list of user subroutine names and brief descriptions A list of all GUI tools and dialog boxes

Help on any of the above items is obtained simply by selecting the appropriate title in the scroll list underneath the main information display area. In addition, the default listing of any user subroutine may be displayed by selecting item UserSubs from the Module pop-up menu and then choosing the 2-36

Version 4.02

Chapter 2

BASIC STAR-CD FEATURES Getting On-line Help

required subroutine name from the second scroll list. Details on the functionality and syntax of every command may be displayed as follows: • •

• •

By typing the command name in the Find Command text box and pressing Return By selecting the appropriate command module from the Module pop-up menu (see the pro-STAR Commands volume for a description of modules) and then choosing the required command name in the scroll list By searching through the available help text for a keyword, as typed in the Keyword text box In a context-sensitive manner, by choosing option Select Item from the Help menu. This changes the mouse pointer from an arrow to a ‘hand’ (Help) pointer with which you can click any part of the main pro-STAR window. Such an action will automatically display the corresponding command description for that part of the window.

An example of command help is shown below:

The middle section of the Help menu gives on-line access to every volume in the STAR-CD documentation set, consisting of Release Notes for the current version, pro-STAR Commands, Methodology, Tutorials and also the CCM, Meshing and Post-Processing User Guides. To view these documents, users must make sure that Adobe’s Acrobat™ Reader is installed on their machine. Instructions on how to do this are given in the STAR-CD Installation and Systems Guide. There is also a Help section containing useful information on how to best use Acrobat for viewing on-line help text corresponding to each panel of the STAR GUIde system described below. The last section of the Help menu activates your machine’s web browser and directs it to useful web sites set up by CD-adapco.

Version 4.02

2-37


Chapter 2

The STAR GUIde Environment

The STAR GUIde Environment STAR GUIde represents the latest development in easy-to-use GUI tools for building STAR-CD models. It works by • •

•

dividing the CCM analysis task into groups of major modelling activities; displaying pre-defined groups of panels relating to each of the activities so that the user can specify model parameters and characteristics pertinent to the current activity; guiding the user through the modelling process in a logical sequence so that no steps of that process are overlooked.

At present, the STAR GUIde panels cover a subset of pro-STAR’s capabilities, i.e. those that relate to the most common tasks of the modelling process. Additional capabilities are being continually added and appear in each new version of STAR-CD. STAR GUIde may be accessed from pro-STAR’s main window using either of the following two methods: 1. Selecting Tools > STAR GUIde from the menu bar 2. Clicking the STAR GUIde button at the top left-hand side of the window.

This displays the introductory screen shown below. The screen consists of two parts: •

•

2-38

On the left is the Navigation Centre (NavCenter), a tool for guiding the user through the various stages of the model building process. These stages are represented by panels and are subdivided into logical groups. The panels and their groups are shown as a tree structure within the NavCenter sub-window. On the right is the initial Help screen explaining how STAR GUIde works and what its function buttons do. This is replaced by the contents of the current process panel as you go through each stage of model building.

Version 4.02

Chapter 2

BASIC STAR-CD FEATURES The STAR GUIde Environment

The following points should be borne in mind when using this tool: 1. The NavCenter tree contains a set of yellow folder icons representing each major modelling activity and acts as the starting point for defining your own model. A complete STAR-CD simulation can be set up and run by performing the activities in the folder tree and in the order shown. 2. Click on one of the yellow folder icons (or on the text next to them) to open and close the folder and to display its constituent process panels and sub-folders. 3. Click on a grey panel icon (or on the text next to it) to open the panel; its contents will be displayed on the right-hand side of the STAR GUIde window. Each process panel enables you to enter or generate data needed to complete that process. 4. Where appropriate, the input for a given process is distributed amongst colour-coded, ‘file tabs’. These are brought to the forefront by clicking on the appropriate tab. The colour coding depends on the entity (block, spline, cell, etc.) being processed and is consistent with the colour coding used in the main Version 4.02

2-39


Chapter 2

The STAR GUIde Environment

pro-STAR window. 5. In some instances, clicking a button on a panel activates a separate, free-floating dialog box. This happens whenever such a dialog provides the most convenient means of entering the data required. 6. To exit from the STAR GUIde, click the Close STAR GUIde button at the bottom of the NavCenter sub-window. Panel navigation system The set of five buttons at the top right-hand side of the STAR GUIde window are designed to help you navigate through the system and get more information about what to do. The function of each button is as follows: Go Back — returns to the previously selected panel.

Collapse/Expand Navcenter — Closes the left-hand (NavCenter) side of the STAR GUIde window to make more space on your screen. The window may be expanded back to its original size by clicking this button again.

Favorite — enables you to store the names of frequently used panels so that you may jump to them directly, i.e. without first opening the STAR GUIde window and then searching through the NavCenter tree. A ‘favourite’ panel is selected by first displaying it in STAR GUIde, clicking Favorite and then choosing the Add to favorites option. The reverse operation is performed by choosing Remove from favorites. The current favourites are listed under the Favorites menu in the main pro-STAR window. Help — provides concise information on the current panel, including descriptions of the data required, explanations of the choices available, suggestions on things to look out for, etc. Help screens use Adobe’s Acrobat™ Reader system; their contents therefore appear in a separate window opened by that system. Information on how to best use Acrobat for reading these screens is given under the Help menu in the main pro-STAR window (option Help).

2-40

Version 4.02

Chapter 2

BASIC STAR-CD FEATURES General Guidelines

Go Fwd — if the Go Back control has already been used, goes forward to the most recently displayed panel.

STAR GUIde usage The STAR GUIde panels should be used in conjunction with the facilities (pop-up menus and action buttons) offered by the main pro-STAR window. The input/output window should also be displayed to cater for operations that need command input (see also the “Introduction” section). For maximum ease of use, all three windows should be displayed side-by-side on your screen, as shown below:

General Guidelines The following general guidelines should be kept in mind when running STAR-CD models, including those described in the Tutorials volume: 1. Take advantage of the on-line Help facilities to check the code’s conventions and, if necessary, the structure and meaning of individual commands. These facilities are accessed either from the GUI Help menu (see “Getting On-line Help”) or by typing HELP, command_name in the pro-STAR I/O window. Version 4.02

2-41


Chapter 2

General Guidelines

2. Make frequent use of the File > Save Model option (or command SAVE) to store the current state of your model description on the pro-STAR model file (.mdl). This safeguards against unexpected mishaps (power failures, system crashes, etc.) by enabling you to restart your work from the point where the last SAVE operation was performed. You should, however, make sure that the model is in a satisfactory state before saving it. 3. If necessary, split lengthy model-building sessions into several parts, by using option File > Quit (or command QUIT) at any convenient point in your current session and then saving your work on the .mdl file. To continue working on the model, re-enter pro-STAR as discussed in Step 10 on page 2-7 and then perform the next operation. However, remember that for transient problems the transient data (.trns) file has to be explicitly re-connected to the pro-STAR session by using the Connect button in the Advanced Transients dialog (or command TRFILE). 4. Mistakes in pro-STAR can be rectified in two ways: (a) Use option File > Resume Model (or command RESUME) to go back to the state of the model saved with the last SAVE operation and start again from there (b) Use command RECOVER to play back all commands issued since the last SAVE operation, re-execute the code up to the one that went wrong, and continue from there 5. Note that command execution can be terminated half way through in the following circumstances: (a) By typing Abort instead of a parameter value while supplying parameter values to a command in ‘novice’ mode. (b) By typing Ctrl+C while waiting for a command to finish processing. Note that the effect of this operation is machine-dependent and therefore great caution should be exercised in its use; in some machines it will abort the entire pro-STAR session. 6. Display the relevant STAR GUIde panels frequently to check the settings of pro-STAR parameters; alternatively use command STATUS. In the latter case, the screen information relates to the active command module, so make sure you are in the right module by typing the appropriate keyword (MESH, PROPERTY, CONTROL, etc.) 7. Remember that all pro-STAR windows can be re-sized using the mouse. It is recommended that both the I/O and the main window are positioned and sized so that both are visible simultaneously. This is particularly helpful when you need to use commands for a particular operation, or if you want to check the commands that were generated automatically by a particular GUI operation.

2-42

Version 4.02

Chapter 3

MATERIAL PROPERTY AND PROBLEM CHARACTERISATION Introduction

Chapter 3

MATERIAL PROPERTY AND PROBLEM CHARACTERISATION

Introduction The physical properties of the fluid and/or solid materials within the model are typically defined immediately after setting up the mesh and performing a thorough visual and numerical check on it. STAR-CD can analyse problems containing arbitrary combinations of • • •

multi-domain fluids, where there is no mixing of fluid streams, porous materials, solids materials.

The Cell Table The process of setting up properties is usually quite simple and relies on the concept of cell identity and the consequent use of the cell table, as discussed under “Cell types” on page 2-37 of the Meshing User Guide. The cell table can be defined using pro-STAR’s Cell Table Editor, accessed by clicking the CTAB button on the left-hand side of the main pro-STAR window. All cells in the mesh can be indexed and differentiated in various ways with the aid of an entry in the cell table. This enables the user to specify a • • • • • • • • • • • • •

cell table index cell type material number colour table index porosity index spin index group number surface lighting material index processor number conduction thickness radiation switch initial free-surface identifier identifying name

for a set of cells, as shown in the dialog below. The meaning of the various parameters that may be set in this table is described in “Cell properties” on page 2-38 of the Meshing User Guide.

Version 4.02

3-1


Chapter 3

The Cell Table

Commands:

CTABLE CTDELETE

CTNAME CTCOMPRESS

CTMODIFY

CTLIST

The rules governing the use of the cell table are as follows: •

• • • • • •

•

•

3-2

All entries in the table are identified by an index, listed under the Table # heading in the editor’s scroll list. A new entry is set up by clicking on the next available number in the list and then specifying the relevant cell properties. Every cell in the model is associated with a cell table index. All cells linked via a common index belong to a common Cell Type (Fluid, Solid, Baffle, etc.), selected from the editor’s pop-up menu. Different materials are identified by separate material property numbers, typed in the Material Number text box. The default cell table index is number 1 and is associated with a fluid whose material number is 1. By default, material number 1 refers to air properties at standard conditions. Cell indexing normally differentiates the cells’ material type. However, it can also be used purely for visual and/or selection purposes. Thus, in the diffuser model shown in Figure 3-1 there is a single material number (no. 1), corresponding to the one and only domain in the model, but the cells can be indexed to different colours or different types of surface shading (see Chapter 4 of the Meshing User Guide). This is done by typing different values in the Color Table Index or Lighting Material text boxes, respectively. Colour selection is facilitated by clicking the multi-coloured button next to the Color Table Index box. This opens a Color Palette panel where the desired colour is selected by simply clicking the appropriate square. The corresponding colour number is then automatically entered into the box. Another possibility is to index cells on the basis of a common group number, typed in the Group Number text box. This groups together all cells belonging to a particular ‘object’, e.g. a distinct portion of the mesh. Such objects might typically be generated with the help of an external CAD package and Version 4.02

Chapter 3

MATERIAL PROPERTY AND PROBLEM CHARACTERISATION The Cell Table

•

imported into pro-STAR using IGES or VDA data files. Group numbers are normally generated automatically as part of the data import function (see “Importing Data from other Systems” on page 3-1 of the Meshing User Guide). Cell table entries can be further identified by a name, typed in the Name box.

A cell table definition is confirmed by clicking the Apply button.

Cell index 1

Cell index 2

Cell index 3

Colour 2 Colour 3 Colour 4

Figure 3-1

Cell indexing to implement differentiation by cell colour

Cell table entries may be displayed at any stage of the pro-STAR session by clicking CTAB on the main window. Any identifier, index, or reference number used in a cell table entry may be changed to a different value simply by selecting the entry in the Cell Table Editor’s scroll list and making the required changes. Cell table entries may also be deleted by clicking the Delete button. Note that all cells indexed to this entry must be deleted or changed to a different index before the table entry itself can be deleted. Tables that contain deleted (or undefined) entries such as this may be cleaned up by clicking the Compress button. This removes all redundant entries and re-numbers the remaining ones. Cell indexing Cells are assigned an identity (cell index) using the Cell Tool shown overleaf. This may be done in two ways: 1. Implicitly, by taking on the index that is active at the moment of their creation. The active cell type can be changed at any time by highlighting the type required in the Cell Table list displayed by the Cell Tool and then clicking the Set Active Type button. The selection is indicated in the list by a letter ‘A’ against the active type. 2. Explicitly, by collecting together a group of cells and then changing their identity to the currently-active type. This can be done by: (a) Pointing at the desired cells with the screen cursor — choose option Modify Type > Cursor Select. The action is terminated by clicking the Done button displayed on the plot. (b) Changing all cells contained within a polygon drawn on the screen with the screen cursor — choose option Modify Type > Zone. The action is terminated by clicking on i) the same point twice to complete the polygon; Version 4.02

3-3


Chapter 3

The Cell Table

ii) the Close button displayed, to let pro-STAR complete the polygon; iii) the Abort button displayed, to abort the selection operation. (c) Changing all surface cells encountered when searching from a starting position given by a ‘seed’ vertex (see the description on page 2-49 of the Meshing User Guide). This can be done by choosing option Modify Type > Surface (New Edge Vertex Set) (or Surface (Current Vertex Set)). The ‘seed’ vertex is selected with the screen cursor. (d) Changing all cells in the current cell set — choose option Modify Type > Cell Set.

Commands:

CTYPE

CCROSS

CFIND

CZONE

CTCOMPRESS

Another method of making changes is via the Cell List dialog, shown overleaf. This may be displayed by clicking the Cell List button on the Cell Tool or choosing Lists > Cells from the main menu bar. The cell or cell range to be changed must first be selected on the list. To change the associated cell type, click Change Type, choose a different cell table index on the displayed Change Cell Table box and then click Apply.

3-4

Version 4.02

Chapter 3

MATERIAL PROPERTY AND PROBLEM CHARACTERISATION Multi-Domain Property Setting

Command:

CMODIFY

The result of the above process can be checked using the Check Tool, option Double Cells (see “Microscopic checking” on page 3-28 of the Meshing User Guide). This will verify whether a cell table entry exists for every cell within the range specified.

Multi-Domain Property Setting The user is free to define as many material types (of the fluid or solid variety) as are necessary to represent the problem conditions. The most general case, involving multiple fluid domains in the presence of solids, is illustrated in the example below:

Domain 1 Metal plate Domain 2

Figure 3-2 Version 4.02

Multi-domain flow with solid material domains 3-5


Chapter 3

Multi-Domain Property Setting

Setting up models Step 1 Create an appropriate set of cell types and material indices for your model during mesh generation, using the procedure described in “The Cell Table” on page 3-1. The appropriate settings to be supplied via the Cell Table Editor for the example shown in Figure 3-2 are as follows: Domain 1

Metal plate

3-6

Version 4.02

Chapter 3

MATERIAL PROPERTY AND PROBLEM CHARACTERISATION Multi-Domain Property Setting

Domain 2

The Material Number indices 11, 12 and 13 above refer to the physical property sets associated with each fluid domain and with the solid plate domain. Note that different cell table indices 1, 2 and 4 are also assigned to each of these because each cell table index can only refer to one material number. In cases with multiple domains it is recommended that each domain is given a separate material number, even if domains have identical physical property sets. This is to allow each domain to have its own initialisation, reference values and residual normalisation. Step 2 Open the Thermophysical Models and Properties folder in STAR GUIde. For thermal problems, specify any special thermal transfer conditions (radiation, solar radiation or solid-fluid heat transfer) prevailing in your model by making the relevant selection(s) in the “Thermal Options” panel. Step 3 Set the physical properties of each fluid domain by opening sub-folder Liquids and Gases and then entering numerical values and/or selecting appropriate options in the “Molecular Properties” panel. Note that: •

•

The option chosen for density calculations determines whether the flow is treated as compressible or incompressible. Special considerations regarding the analysis of compressible flows are discussed in “Compressible Flow” on page 3-9 of this chapter. Non-Newtonian flow may be simulated by selecting the relevant molecular viscosity calculation option. The treatment of non-Newtonian fluids is discussed further in “Non-Newtonian Flow” on page 3-11 of this chapter.

Each domain must be selected in turn via the Material # control at the bottom of the panel (see also the “Liquids and Gases” Help topic).

Version 4.02

3-7


Chapter 3

Multi-Domain Property Setting

Step 4 If you have selected the solid-fluid heat transfer option, an additional sub-folder, Solids, will appear in the STAR GUIde tree structure. Specify the physical properties of the solid material by entering numerical values and/or choosing appropriate options in the “Material Properties” panel. If your model contains multiple solid domains possessing different properties, each domain may be selected in turn via the Material # control at the bottom of the panel (see also the “Solids” Help topic). Step 5 For turbulent fluid domains, choose an appropriate option from the “Turbulence Models” panel. Further details are given in “Turbulence Modelling” on page 3-12 of this chapter. Step 6 For thermal problems, turn on the enthalpy equation solver in all fluid domains using the “Thermal Models” panel. The enthalpy equation solver for solid materials is activated simply by selecting Solid-Fluid Heat Transfer in the “Thermal Options” panel. Special considerations regarding the use of this option are discussed in “Heat Transfer In Solid-Fluid Systems” on page 3-16 of this chapter. Step 7 Specify initial values for the flow variables in each fluid domain using the “Initialisation” panel (Liquids and Gases folder). The temperature distribution inside solid materials is specified via a separate “Initialisation” panel under the Solids folder. Step 8 Set the reference quantities (pressure and temperature) and monitoring cell location(s) for each domain using the “Monitoring and Reference Data” panel (Liquids and Gases folder). The reference temperature and monitoring cell location for solids is specified via a separate “Monitoring and Reference Data” panel under the Solids folder. Step 9 For buoyancy-driven or any other problems involving body forces, specify the necessary parameters using the “Buoyancy” panel. Special considerations regarding the use of this option are discussed in “Buoyancy-driven Flows and Natural Convection” on page 3-20 of this chapter. Step 10 If necessary, specify mass sources or additional source terms for the solution of the momentum, turbulence or enthalpy equation. The type of source is chosen by selecting the appropriate tab in STAR GUIde’s “Source Terms” panel (sub-folder Sources): •

•

3-8

Mass — specify mass sources or sinks, i.e. fluid injection or withdrawal, to be used in the solution of the mass conservation equation (tab “Mass”). Special considerations regarding the use of subroutine FLUINJ for this purpose are discussed in “Fluid Injection” on page 3-21 of this chapter. Momentum — specify momentum sources, e.g. a fan driving the flow at some location of your model, where the fan is not explicitly modelled (tab Version 4.02

Chapter 3

MATERIAL PROPERTY AND PROBLEM CHARACTERISATION Compressible Flow

•

•

“Momentum”). Turbulence — specify sources appropriate to the turbulence model used. These may be additional source terms or, in the case of the k-ε model, replacements for the existing terms (tab “Turbulence”). Enthalpy — specify heat sources or sinks, e.g. radioactive sources in a nuclear reactor cooling problem (tab “Enthalpy”).

All property and physical model settings in your problem may be inspected by selecting the relevant panels in the Thermophysical Models and Properties folder. In sub-folders Liquids and Gases and Solids, open each constituent panel in turn and scroll through the available materials. Alternatively, type command MLIST to display a brief or comprehensive listing of properties for any material in the Output window.

Compressible Flow The theory behind compressible flow problems and the manner of implementing it in STAR-CD is given in the Methodology volume (Chapter 16, “Compressible Flows”). This section contains an outline of the process to be followed when setting up such problems and important points to bear in mind. Also included are cross-references to appropriate parts of the STAR GUIde on-line Help system, containing details of the user input required. Setting up compressible flow models Step 1 Go to panel “Molecular Properties” in STAR-GUIde and select each compressible fluid domain via the slider at the bottom of the panel. Step 2 Declare the flow as (ideal gas) compressible by selecting option Ideal-f(T,P) from the “Density” pop-up menu. This effectively switches on the compressibility calculations by making the density a function of both pressure and temperature. Step 3 Set up boundary conditions that are appropriate to the type of flow being analysed. These are as follows: Subsonic flow (Ma < 1 throughout the solution domain) Inflow Stagnation conditions Inlet Inlet Inlet

Outflow Pressure Pressure Outlet (for steady flow, but see point no. 1 below) Wave transmissive (for transient flow)

Supersonic flow (Ma > 1 throughout the solution domain) Inflow Inlet Inlet Version 4.02

Outflow Outlet Pressure 3-9


Chapter 3

Compressible Flow

Transonic flow (Ma < 1 and Ma > 1 within the solution domain) Subsonic Inflow Stagnation conditions Inlet Supersonic Inflow Inlet Supersonic Inflow Inlet Subsonic Inflow Stagnation conditions

Subsonic Outflow Pressure Pressure Subsonic Outflow Pressure Supersonic Outflow Pressure Supersonic Outflow Pressure

The user should refer to the on-line Help text for panel “Define Boundary Regions” (especially that for “Inlet” boundaries) for a description of how to set up boundary conditions for this type of flow. Useful points on compressible flow 1. The combination of inlet and outlet boundary conditions for subsonic flows presented under Step 3 above does not constitute, strictly speaking, a ‘well posed’ problem. However, it is offered as an option for use in circumstances where the pressure is known at the inflow (or at some other point inside the solution domain) but not at the outflow. In such a case, users should designate the known pressure as the reference pressure and make sure the corresponding cell location lies as close as possible to the known location (e.g. the inlet boundary surface). The success of the simulation will depend on the magnitude of the Mach number. For the higher Mach numbers (e.g. Ma > 0.7) very low under-relaxation factors will have to be specified (e.g. 0.001 for pressure) in order to obtain a converged solution. 2. Special considerations apply to tetrahedral meshes or meshes containing trimmed (polyhedral) cells. If such meshes contain supersonic inlet boundaries then, to obtain a stable/convergent solution, it is necessary to create at least two cell layers immediately next to the boundary (see Figure 4-5 on page 4-23). If pro-STAR’s automatic meshing module is employed for this purpose, use its built-in mesh generation capabilities. If the mesh is imported from a package that lacks these facilities, you must extrude the mesh in a direction normal to the boundary and then shift the boundary location to the edge of the newly-created, layered structure. 3. In the case of a transonic problem with subsonic inflow, residual normalisation for momentum (and k, ε if appropriate) is based on the momentum (and k, ε) flux values at the inlet, as usual. However, because of the large difference in velocity magnitude between the inlet and the rest of the flow field, this may place an unnecessarily stringent condition on the built-in solution convergence criterion (as discussed in Chapter 1, “Monitoring the calculations”, this is based on the magnitude of the normalised residuals). In this situation, it could be more appropriate to inspect the convergence history of, say, mass and enthalpy and terminate the solution process after a sufficiently large number of iterations. 4. For inviscid flows, it is possible to calculate temperature from a constant 3-10

Version 4.02

Chapter 3

MATERIAL PROPERTY AND PROBLEM CHARACTERISATION Non-Newtonian Flow

stagnation enthalpy relationship rather than the standard enthalpy equation. To do this, go to panel “Thermal Models” in STAR-GUIde and select option Stagnation Enthalpy from the Conservation pop-up menu. The appropriate stagnation temperature should then be typed in the Stagnation Temp. text box. 5. It could be advantageous, even when a steady state is sought, to do a transient calculation using the “Pseudo-Transient Solution” method. To do this, select option Pseudo-Transient from the pop-up menu in the “Solution Method” STAR GUIde panel. 6. In the case of flow through ducts of non-uniform cross-section where supersonic conditions are expected over the whole or part of the solution domain, it is sometimes necessary to under-relax the initial velocities. This is done by activating special flux under-relaxation using panel “Miscellaneous Controls” in STAR GUIde. This operation affects only the velocity initialisation.

Non-Newtonian Flow The theory behind non-Newtonian flow and the manner of implementing it in STAR-CD is given in the Methodology volume (Chapter 16, “Non-Newtonian Flows”). This section contains an outline of the process to be followed when specifying non-Newtonian fluids and includes cross-references to appropriate parts of the STAR GUIde on-line Help system. The latter contains details of the user input required. Setting up non-Newtonian models Step 1 Decide whether the power law offers an adequate representation of the non-Newtonian fluid behaviour and what the value of the constants m and n in equation (1-6) of the Methodology should be. Alternatively, supply a suitable expression in subroutine VISMOL. Step 2 Go to panel “Molecular Properties” in STAR-GUIde and select the domain containing the non-Newtonian fluid via the slider at the bottom of the panel. Step 3 Use the “Molecular Viscosity” menu to either specify the model parameters m and n (option NonNewt, text boxes EM and EN) or call subroutine VISMOL (option User). Useful points on non-Newtonian flow 1. Bear in mind that constitutive relations for non-Newtonian flow are basically empirical curve-fitting formulae. It is therefore inadvisable to use them beyond the range of the available data. 2. The model parameters are functions of temperature, pressure and composition. They may also be functions of the rate of strain tensor’s range II s (see equation (1-5) in Chapter 1 of the Methodology volume), over which the equation is fitted. If any of these effects are significant, they should be allowed for in user subroutine VISMOL. Version 4.02

3-11


Chapter 3

Turbulence Modelling

Turbulence Modelling The theory behind the currently available models is given in Chapter 2 of the Methodology manual. A number of methods are also available for implementing the no-slip boundary conditions for turbulent flow, as follows: 1. Wall functions, applied to cells immediately adjacent to a wall. This method employs special algebraic formulae (described in Chapter 6, “High Reynolds number turbulence models and wall functions” of the Methodology volume) to represent velocity, temperature, turbulence parameters, etc. within the boundary layer that forms next to the wall; see Figure 3-3(a). The method is also appropriate for use with one-equation (k-l, Spalart-Allmaras), k-ω and Reynolds Stress models. An alternative, ‘non-equilibrium’ type of wall function is also provided for taking pressure gradient effects into account (see equation (6-17), (6-18) and (6-19) in the Methodology volume) but this is available only for k-ε models (linear and non-linear). 2. Two-layer models, employed as combinations of a high Reynolds number (k-ε) model with a low Reynolds number (one-equation or zero-equation) model. The latter is applied to the near-wall region where the mesh should be finely spaced, as shown in Figure 3-3(b); see also Chapter 6, “Two-layer models” in the Methodology volume. You are free to combine the wall function and two-layer approach within the same problem, provided that a linear k-ε type model is in use and the two treatments are applied to different boundary regions. However, care must be exercised at transition points between the two methods. 3. Low Reynolds number models, in which viscous effects are incorporated in the k and ε transport equations. No special near-wall treatment (other than an optional definition of wall surface roughness) is therefore required; see also Chapter 6, “Low Reynolds number turbulence models”. Both low Re and wall function treatments may be used in the same problem, but only if they apply to separate domains. 4. Hybrid wall boundary condition, which offers a special wall treatment for low Reynolds number models independent of the normalised parameter y + . For finely spaced meshes, this is identical to the standard low Reynolds number treatment. For coarser meshes, it provides special algebraic formulae to represent velocity, temperature, turbulence parameters, etc. similar to ordinary wall functions (see also Chapter 6, “Hybrid wall boundary condition”). The choice of wall treatment (where relevant) is made in the “Near-Wall Treatment” tab of the “Turbulence Models” panel. If a two-layer model is employed, you will need to indicate the wall or baffle region to which it applies via the “Define Boundary Regions” panel.

3-12

Version 4.02

Chapter 3

MATERIAL PROPERTY AND PROBLEM CHARACTERISATION Turbulence Modelling

k - ε model

match location NWL y Low Re model (a) Wall function model

Figure 3-3

(b) Two-layer models

Mesh spacing in the near-wall region

The following points should be borne in mind when considering the effectiveness or accuracy of a particular turbulence model or near-wall treatment: Wall functions 1. For reasons of accuracy, the normal distance y from the wall for near-wall cells (see Figure 3-3) should be such that the dimensionless parameter y + is kept within the limits 30 < y + < 100 , where: +

1⁄4 1⁄2

y ≡ ρ Cµ

k

y⁄µ

2. It is important to place y outside the viscous sublayer. This can be achieved by observing the lower limit on the value of y + . 3. The above considerations apply equally to both standard and non-equilibrium wall functions. The difference between the two is that the latter takes the pressure gradient into account. This provides more accurate results in terms of wall shear forces but has little effect on the character of the flow. 4. If the non-equilibrium option is chosen, the normal user inputs for wall roughness (specified via the Roughness pop-up menu for wall and baffle boundaries, see panel “Define Boundary Regions”) are not applicable. Two-layer models 1. These should be preferred for non-equilibrium flows, as they produce improved friction and heat transfer predictions. Their use, however, will result in larger meshes within the model and hence significantly higher calculation times. This is because the near-wall region requires a finer mesh than that needed by the wall function treatment. 2. In order to resolve properly the distributions of velocity and other variables within the near-wall region (i.e. at y + ≤ 40 ), it is necessary to ensure that it is spanned by about 15 mesh nodes. In general, this may require some trial and error adjustment of the mesh, since the near-wall region thickness is not known a priori. Once a suitable mesh density is chosen, the value of y + at the Version 4.02

3-13


Chapter 3

Turbulence Modelling

node next to the wall should be no larger than ~3 to resolve the velocity profile, but smaller to resolve the thermal profile. 3. If the prescribed NWL thickness is not sufficiently large to encompass the near-wall region throughout the domain in question (i.e. the switching location between high and low Re regions shown in Figure 3-3 lies outside the NWL in some places), the switching location there is assumed to be at the edge of the NWL and a warning message is issued on file case.info. In such cases, it is possible to increase the NWL thickness to a more suitable value and restart the calculations. 4. There is an additional option for fixing the above switching location to its current position. If this option is selected from the start of the analysis, its effect is to make the switching point distance equal to the NWL thickness. 5. During post-processing, the partitioning of the mesh into (a) near-wall region cells where the one-equation model applies (b) other cells in the NWL (c) ordinary cells in the flow field interior can be inspected by opening panel “Load Data” in STAR GUIde (“Data tab”), choosing “Cell Data” as the data type and then selecting option Two Layer from the Scalar Data scroll list. Option FMU allows inspection of the distribution of a quantity given by νt -----------------2 Cµk ⁄ ε Low Re models 1. These should be preferred for non-equilibrium flows, for the same reasons as two-layer models. However, their use may require meshes that are even larger than those for the two-layer approach. 2. The default treatment assumes a smooth wall but the wall surface roughness may also be specified, if required. 3. In order to resolve properly the distribution of velocity and other variables, approximately 20 mesh nodes are needed within the near-wall region ( y + ≤ 40 ). The value of y + at the node next to the wall should then be ~1. Note that this meshing strategy differs from that for two-layer models, where approximately 15 mesh nodes are needed over the near-wall region. This means that a mesh designed for two-layer models will not necessarily be suitable for low Re models. 4. As with two-layer models, computing times are substantially greater than when using a wall function approach. 5. It is recommended that such models are run in double precision. Hybrid wall boundary condition 1. The hybrid wall condition is an extension of low Reynolds number boundary conditions. It applies only to the following low Reynolds number turbulence models: (a) k-ε (linear, cubic and quadratic) (b) k-ω (standard and SST variants) 3-14

Version 4.02

Chapter 3

MATERIAL PROPERTY AND PROBLEM CHARACTERISATION Turbulence Modelling

(c) Spalart-Allmaras 2. The approach automatically selects a low Reynolds number wall treatment or a wall function, depending on the local flow field and near-wall mesh spacing. It should be preferred in situations where (a) the normalised parameter y + is unknown, or (b) large variations in y + create uncertainties as to whether a low Reynolds number boundary treatment or a wall function is appropriate. Reynolds Stress models 1. Both the Gibson-Launder and SSG models are high Reynolds number models so they need to be used in conjunction with wall functions. 2. Since Reynolds Stress models solve additional transport equations for Reynolds Stress components, they consume a substantially greater amount of computing time compared to k-ε models. 3. The ‘standard’ wall reflection term used in the Gibson - Launder model is not suitable for impingement flows. In such circumstances, it will return the wrong distribution of the stress component normal to the wall. It is therefore advisable to use the term calculated by the Craft model instead. DES models 1. A transient analysis setting is required, although the problem being modelled may in reality be a steady-state one. 2. The 3-time-level temporal discretisation scheme within the transient SIMPLE algorithm achieves second-order accuracy, but may be computationally expensive. Choosing the PISO algorithm results in an accuracy comparable to that of a formally second-order scheme, whilst being computationally cheaper. 3. Central differencing or automatic blending (see Chapter 2, “Blending Function” in the Methodology volume) is recommended for the discretisation of convective terms in the momentum equation; the MARS scheme is recommended for the turbulence equations. LES models 1. A transient analysis setting is required, although the problem being modelled may in reality be a steady-state one. 2. The 3-time-level temporal discretisation scheme within the transient SIMPLE algorithm achieves second-order accuracy, but may be computationally expensive. Choosing the PISO algorithm results in an accuracy comparable to that of a formally second-order scheme, whilst being computationally cheaper. 3. The time step size should be selected in such a way that the maximum Courant number does not exceed 0.5 4. Central differencing is recommended for the discretisation of convective terms in the momentum equation; the MARS scheme, with blending factor not less than 0.5, can be conveniently used for bounded scalars (e.g. mixture fractions or enthalpy).

Version 4.02

3-15


Chapter 3

Heat Transfer In Solid-Fluid Systems

Changing the turbulence model in use This facility allows you to run a turbulent flow case by restarting from a simulation done for the same case but with a different turbulence model. No special user input is required to run such a case. The table below illustrates the combinations allowed and the conversion formula adopted when STAR encounters a different turbulence model in the solution file to the one currently in force: FROM (Restart field) SpalartAllmaras

TO (New solution field)

SpalartAllmaras

k-ε type*

2

k ν t = C µ ----ε

k-ω (Wilcox and SST)

k ν t = ---ω

2

V2F

2

k k ν t = C µ ----- ν t = C µ ----ε ε

ε = C µ kω Not needed Not needed

k-ε type*

k-ω (Wilcox and SST)

ε ω = --------Cµk

Reynolds Stress (GL and SSG)

Not needed ε = C µ kω

V2F

Reynolds Stress (GL and SSG)

ε ω = --------Cµk

ε ω = --------Cµk

Not needed

Not needed ε = C µ kω Not needed

* k-ε, k-ε Quadratic, k-ε Cubic, k-ε RNG, k-ε CHEN, k-ε Speziale, k-ε Suga Quadratic and Cubic

Heat Transfer In Solid-Fluid Systems The theory behind this type of heat transfer models and the manner of implementing it in STAR-CD is given in Chapter 16, “Heat Transfer in Solid-Fluid Systems” of the Methodology volume. This section contains an outline of the process to be followed when setting up this type of model and includes cross-references to appropriate parts of the STAR GUIde on-line Help system. The latter contains details of the user input required and important points to bear in mind when setting up problems of this kind.

3-16

Version 4.02

Chapter 3

MATERIAL PROPERTY AND PROBLEM CHARACTERISATION Heat Transfer In Solid-Fluid Systems

Setting up solid-fluid heat transfer models Step 1 Specify the model regions occupied by the solids and fluids present and define their physical properties. Material 1 — steam

Heat flow Material 3 — steel

Material 2 — hot gas

Figure 3-4

Simple heat exchanger

In terms of the heat exchanger example shown in Figure 3-4, this requires the following actions (see also “Multi-Domain Property Setting” on page 3-5): • •

Set up cell table entries for fluid materials 1,2 and solid material 3 Assign all cells in the mesh to the appropriate cell type (1, 2, 3) as described in the section on “Cell indexing” on page 3-3. • Specify the physical properties of each material Step 2 Turn on Solid-Fluid Heat Transfer in the “Thermal Options” STAR-GUIde panel. Note that this also has the effect of switching on the temperature solver in solid materials. Step 3 Switch on the temperature solver in each fluid material using the “Thermal Models” panel. Step 4 Normally, STAR-CD treats the solid-fluid interface as part of the default wall region (region 0). However, unlike other parts of this region whose default thermal condition is adiabatic, the solid-fluid interface is treated as a conducting wall. Therefore: •

If an additional thermal resistance exists at the interface, define the latter as a separate region and use the “Define Boundary Regions” panel to specify it as a conducting wall having the required thermal resistance value (see the STAR GUIde “Wall” Help topic for more information). • STAR uses default expressions to calculate heat transfer (film) coefficients at all solid/fluid interfaces, including those at external walls and baffles. You can supply alternative expressions for these quantities via subroutine MODSWF Step 5 If a printout of temperature distribution in the model is required, use command Version 4.02

3-17


Chapter 3

Heat Transfer In Solid-Fluid Systems

PRTEMP to specify whether the printed values are absolute or relative to the datum temperature previously defined (see topic “Reference Data” in the STAR GUIde on-line Help system). Heat transfer in baffles Thermal conduction along the plane of a baffle’s surface is currently neglected (see the STAR GUIde “Baffle” Help topic for more information). However, this effect may still be modelled by expanding a baffle into a single layer of solid cells using command CBEXTRUDE (see also Chapter 2, “Extrusion” in the Meshing User Guide). The surrounding mesh is automatically adjusted to make room for the solid cells, as shown in Figure 3-5. Before

Ordinary baffle

Figure 3-5

After

Fully-conducting baffle

‘Fully-conducting’ baffle creation

Note that: •

Special cell shapes (such as prisms) are created at the edges of the solid cell layer, as shown in the exploded view of the baffle in Figure 3-5. This brings the baffle thickness down to zero and avoids the need to create coupled cells in those parts of the mesh. • The modelling of heat conduction will be slightly in error as a result of the introduction of the above artificial cell shapes. • A baffle of the kind described here may be attached directly to an external boundary or to internal boundaries such as solid-fluid interfaces to model a conducting fin. In the latter case, you need to make sure that the cell type assigned to baffle cells is different from that assigned to solid cells at the base of the baffle. Alternative treatment for baffle heat transfer It can be seen that the expansion process described above will create a disturbance in the fluid cells around the baffle and may result in a highly irregular mesh. In order to avoid this problem, a facility is provided for specifying a finite baffle thickness (to be used internally for heat conduction calculations) but without actually expanding the baffle to that thickness. Thus, the fluid flow calculations are based on an undisturbed mesh structure. 3-18

Version 4.02

Chapter 3

MATERIAL PROPERTY AND PROBLEM CHARACTERISATION Heat Transfer In Solid-Fluid Systems

To use this facility, the following steps are needed: Step 1 Using the Cell Table Editor, create a separate baffle cell type and a separate solid cell type. The latter will be used to represent the ‘conducting baffles’. Step 2 Create the baffle cells in the appropriate mesh location using the baffle cell type defined in Step 1. Step 3 Apply command CBEXTRUDE to the baffle cells created in Step 2 and extrude them into solid cells using the solid cell type created in Step 1. Note that: •

Upon extrusion, the baffle cells will be removed from the mesh and replaced by the solid cells that they have been extruded into. • If no solid cell type identification, ICTID, is supplied in the CBEXTRUDE command, the solid cell identification will be set as cell type 1. • If no solid cell thickness, DT, is supplied in the CBEXTRUDE command (this is the normal practice), the default thickness will be applied, currently set at 0.2 × 10-3 m. Step 4 Go back to the Cell Table Editor and select the solid cell type defined in Step 1. Enter the actual conduction thickness in the box labelled Conduction Thickness Step 5 Turn on Solid-Fluid Heat Transfer in the “Thermal Options” STAR-GUIde panel. Step 6 Apply the appropriate wall boundary condition to the solid cells created in Step 3. If none is specified, the default wall boundary condition for region number 0 will be used. This results in a conducting, no-slip wall. Note that: •

• •

Conducting baffles of the same thickness DT specified in command CBEXTRUDE and of the same Conduction Thickness specified in the Cell Table Editor can share the same cell type. Conducting baffles that have a different DT or different Conduction Thickness must also have a different cell type. A conducting baffle that is attached to a solid base must have a different cell type to that of the solid to which it is attached.

Useful points on solid-fluid heat transfer 1. The On button in the Solid-Fluid Heat Transfer section of the “Thermal Options” STAR-GUIde panel must always be used to turn on the solution of the energy equation in solids, even if the entire model is made up of solid cells. 2. It is usually advisable to run solid-fluid heat transfer simulations in double precision. This helps to overcome potential convergence problems arising as a result of a large disparity in thermal conductivity between fluid and solid. The Version 4.02

3-19


Chapter 3

Buoyancy-driven Flows and Natural Convection

3.

4.

5.

6.

choice of single or double precision mode can be made when running STAR (see Chapter 2, “Running a STAR-CD Analysis”, Step 6). A convenient way of modelling thermal contact resistance between two adjacent solid domains is to define a baffle of suitable properties at the faces of the appropriate solid cells in one of the domains. In some situations the energy under-relaxation factor in fluid domains has to be reduced below its default value of 1.0 to aid convergence. In such cases, we recommend that the corresponding factor for solids is left at 1.0. If your model contains an arbitrary or embedded mesh interface between the fluid and solid cells, you will need to match cells on either side of the interface, as described in Chapter 3, “Couple creation” in the Meshing User Guide. If your model contains scalar variables, the only valid scalar boundary condition for walls located at the solid-fluid interface is Adiabatic.

Buoyancy-driven Flows and Natural Convection The theory behind flow problems of this kind and the manner of implementing it in STAR-CD is given in the Methodology volume (Chapter 16, “Buoyancy-driven Flows and Natural Convection”). The present chapter contains an outline of the process to be followed when setting up buoyancy-driven flows and includes cross-references to appropriate parts of the STAR GUIde on-line Help system. The latter contains details of the user input required and important points to bear in mind when setting up problems of this kind. Setting up buoyancy-driven models Step 1 Switch on the temperature solver using the “Thermal Models” STAR-GUIde panel Step 2 Switch on the density solver by selecting one of the following options from the “Density” pop-up menu in the “Molecular Properties” panel: • Isobaric — isobaric density variation (normally used for liquids) • Ideal-f(T) — density variation based on the Ideal Gas Law • User-f(T) — density variation based on user-defined relationships Step 3 Set up the problem’s initial conditions using the “Initialisation” panel controls Step 4 Define the reference pressure and temperature plus the reference pressure cell location using the “Monitoring and Reference Data” panel Step 5 Use the “Buoyancy” panel to specify suitable buoyancy parameters for your problem. Useful points on buoyancy-driven flow 1. Check the settings in STAR GUIde’s “Gravity” panel (which determine the 3-20

Version 4.02

Chapter 3

MATERIAL PROPERTY AND PROBLEM CHARACTERISATION Fluid Injection

gravitational body force effects) before starting a buoyancy calculation. Also note that, if droplets and/or liquid wall films are present in your model, gravitational effects for these features must be switched on separately. 2. It is usually advisable to run buoyancy-driven flow simulations in double precision. This is because the body force terms in the momentum equation are often so small compared to the other terms that they can be masked by the round-off error of the calculation. The consequences of working in single precision mode are oscillation in the residual values and non-convergence of the solution. The choice of single or double precision mode can be made when running STAR (see Chapter 2, “Running a STAR-CD Analysis”, Step 6). 3. In multi-domain problems, the reference density and datum location should be defined domain-wise. 4. If you use the option for direct specification of the reference density, the latter should be assigned a realistic value based on the expected density variation in the fluid. For simulations without pressure boundaries: (a) In steady-state calculations, unrealistic values can give rise to a body force that is out of balance with the piezometric pressure gradient. This can cause delay in the solution convergence. (b) In transient calculations, these initial disturbances could also produce unrealistic initial fields and therefore invalidate the results of the analysis. 5. If convergence problems are encountered, it is advisable to begin the calculations with a small amount of under-relaxation on both temperature and density, e.g. 0.9. The desired values may be entered in the corresponding Relaxation Factor boxes inside panel “Solver Parameters” in STAR GUIde. This measure often helps to stabilise the solution and promote convergence. 6. In problems of this type, there is very strong coupling between the temperature, scalar mass fraction and flow fields. It is therefore advisable to use the PISO algorithm which is more suitable for this type of coupling. 7. If convergence problems are encountered, it may be necessary to run the model in transient mode. This involves approaching the steady-state solution, if one exists, by means of time steps. The most convenient way of doing this is to use the single-transient solution mode (see Chapter 5, “Default (single-transient) solution mode”), since this way one does not need to set up load steps. 8. Buoyancy-driven flows with high Grashof number (i.e. Gr > 109) are sometimes naturally unstable (i.e. time-dependent without a single unique solution). In such cases, a converged steady-state solution cannot be obtained and you should opt for the transient approach. A method of calculating the time step size is given in the Methodology volume (Chapter 16, “Buoyancy-driven Flows and Natural Convection”).

Fluid Injection The theory behind flow problems of this kind and the manner of implementing it in STAR-CD is given in the Methodology volume (Chapter 16, “Local Fluid Injection/Extraction”). This section contains an outline of the process to be Version 4.02

3-21


Chapter 3

Fluid Injection

followed when setting up fluid injection problems. Also included are crossreferences to appropriate parts of the on-line Help system, containing details of the user input required. Setting up fluid injection models Step 1 Create a set of all cells where fluid injection or removal is to be take place. A separate cell table index number should be assigned to this set (see “The Cell Table” on page 3-1). Step 2 Activate the injection facility using the “Mass” tab in STAR-GUIde’s “Source Terms” panel. Step 3 Copy subroutine FLUINJ into the ufile sub-directory of your working directory, as described in Chapter 14, “Subroutine Usage”. Step 4 Insert appropriate code in subroutine FLUINJ using a suitable editor. Usually, the code specifies the mass flux injected or removed (on a per unit volume basis) for cells of the required type, so that a single value can be used for the entire cell set selected. An example of this is given in the sample coding supplied in subroutine FLUINJ. If only the total amount of mass injected is known, the required value may be obtained by dividing by the total volume of the cell set. Thus, you may need to calculate this volume first, either by choosing Utility > Calculate Volume > Cell Set from pro-STAR’s main menu bar or by using command VOLUME. If mass is being injected, specify all relevant properties of the incoming fluid (i.e. it is assumed that the fluid is bringing all its properties into the computational domain). The properties in question may be velocity components (U, V, W), turbulence parameters (k, ε), temperature and chemical species mass fractions. If mass is being removed, only the mass flux needs to be specified as the withdrawn fluid is assumed to possess the (known) properties in its vicinity.

3-22

Version 4.02

Chapter 4

BOUNDARY AND INITIAL CONDITIONS Introduction

Chapter 4

BOUNDARY AND INITIAL CONDITIONS

Introduction The process of defining boundaries in a model can be divided into two major steps: 1. Identify the location of individual, distinct boundaries (i.e. where the boundaries are). 2. Specify the conditions at the boundaries (i.e. what the conditions are). It is of the utmost importance that boundaries are chosen and implemented correctly, since the outcome of the simulation depends on them. Users should have a good understanding of the physical significance and numerical implications of different boundary conditions and should apply them correctly to their model. It is therefore advisable to refer to the relevant sections of the Methodology volume for guidance.

Boundary Location The two important geometrical features of boundaries are: 1. They are created on the outer surfaces of the mesh, except for: (a) so-called baffle boundaries, which are normally positioned at the interface of two cells; (b) solid/fluid interface boundaries in heat transfer problems. 2. They are grouped into boundary regions. A boundary region consists of a group of cell faces that cover the desired boundary surface. Figure 4-1 shows a boundary region made up of nine cell faces.

Figure 4-1

Boundary region definition

The rules governing the use of boundary regions are as follows: • •

Version 4.02

Regions are numbered in an arbitrary manner by the user, in order to identify them. The indexing of boundary cell faces (or boundaries, for short) comprising a region is done automatically by pro-STAR, in a similar manner to the automatic cell numbering discussed in “Cells” on page 2-37 of the Meshing User Guide. In the example shown in Figure 4-2, boundary nos. 1 to 9 are assigned to region 1. 4-1


Chapter 4

Boundary Location

7

8

9

4

5

6

1

2

3

Figure 4-2

Boundary cell face indexing

Thus, each boundary in the model is identified by a region number (user-defined) and composed of boundary cell faces that are automatically numbered by pro-STAR. pro-STAR offers two methods for setting up boundary regions: 1. Typing commands from the keyboard, as described below 2. Using the facilities of panel “Create Boundaries” in STAR GUIde (“Regions” tab) Command-driven facilities The available functions are as follows: •

•

• • • •

Assignment of boundaries to a region using the keyboard — command BDEFINE. This requires input of the region number, cell number and cell face number on which the boundary will be created. pro-STAR generates the boundary number automatically. Further boundaries can be created individually or generated from an existing set, using command BGENERATE. This creates additional boundaries by applying an offset to the cell numbers of a previously-defined set. Modification of the region number assigned to a boundary face — command BMODIFY. Re-assignment of a boundary to a different region graphically — command BCROSS. Conversion of a set of shells into a set of boundaries — command BSHELL. The starting shells are not deleted by this process. Counting the currently defined boundaries — command COUNT. The same operation can also be executed by choosing Utility > Count > Boundaries from the menu bar.

For further details on the function and application of boundary commands, refer to the pro-STAR Commands volume.

4-2

Version 4.02

Chapter 4

BOUNDARY AND INITIAL CONDITIONS Boundary Location

Boundary set selection facilities Boundaries may need to be grouped together for the purposes of mass manipulation or plotting, thus defining a boundary set. This is done by selecting one of the list options provided by the B-> button in the main pro-STAR window. The available options are: 1. All — puts all existing boundaries in the current set 2. None — clears the current set 3. Invert — replaces the current set with one consisting of all currently unselected boundaries 4. New — replaces the current set with a new set of boundaries 5. Add — adds new boundaries to the current set 6. Unselect — removes boundaries from the current set 7. Subset — selects a smaller group of boundaries from those in the current set For the last four options, the required boundaries are collected by choosing an item from a secondary drop-down list, as follows: • •

• •

•

• • •

Cursor Select — click on the desired boundaries with the cursor, complete the selection by clicking the Done button on the plot Zone — use the cursor to draw a polygon around the desired boundaries. Complete the polygon by clicking the right mouse button (or the Done button outside the display area to let pro-STAR do it for you). Abort the selection by clicking the Abort button. Region (Current) — select all boundaries whose region number is currently highlighted in the boundary region table Region (Cursor Select) — select all boundaries belonging to a given region. The required region is selected by clicking on a representative boundary with the cursor. Patch (Cursor Select) — select all boundaries containing radiation patches (see Chapter 7, Step 6). The patches in question are selected by clicking with the cursor. Vertex Set (All) — all constituent vertices of the selected boundaries must be in the current vertex set Vertex Set (Any) — the selected boundaries must have at least one constituent vertex in the current vertex set Attach, Baffle, Cyclic, Degas (Phase Escape boundary condition used in Eulerian multi-phase problems), Freestream, Inlet, Monitoring, NonReflective_Pressure, NonReflective_Stagnation, Outlet, Pressure, Radiation, Riemann, Stagnation, Symplane, Transient, Wall — all boundaries must be of the type selected, regardless of region number More boundary set operations are available in the Boundary List dialog (see “Boundary listing” below) or by typing command BSET (see the pro-STAR Commands volume for a description of additional selection options).

Boundary listing Boundary information is displayed in the Boundary List dialog shown below, obtained by selecting Lists > Boundaries from the main menu bar. Boundary definitions are displayed in a scroll list in numerically ascending order, in terms of: Version 4.02

4-3


Chapter 4

Boundary Location

• • • • • •

Boundary serial number Parent cell serial number Face number of this cell that has been defined as a boundary Patch serial number of any radiation patch that has also been created on that face Region number Boundary type

There is also a choice of listing all boundaries or just the current set (marked by asterisks in the Bset column). The choice is made by simply selecting the Show All Boundaries or Show Bset Only option, respectively. To select boundaries from the list: • •

For single items, click the number of the required boundary. For two or more items in sequence, click the first boundary you want to select, and then press and hold down the Shift key while you click the last boundary in the group.

Commands:

BLIST

BDELETE

BMODIFY

BSET

Once the desired boundaries are selected, the following additional operations are possible: 1. Addition to (or removal from) the current set — click the Add to Set/Remove from Set button. 2. Deletion — click the Delete Boundary button. 3. Change of boundary region — click the Change Region button. This activates an additional dialog, shown below. To change the region type associated with the selected boundaries, choose a different region number on the displayed Change Region box and then click the Apply button. 4-4

Version 4.02

Chapter 4

BOUNDARY AND INITIAL CONDITIONS Boundary Region Definition

Note that all the above operations have an immediate effect on the boundary definitions, reflected by immediate changes to what is displayed in the list. However, any subsequent boundary changes made outside this dialog, e.g. by issuing commands via the pro-STAR I/O window, will not be listed. To display these changes, click Update List at the top of the dialog.

Boundary Region Definition Having specified the location of all boundaries in the model, the next step is to • •

define their individual type (i.e. set the boundary condition); supply information relevant to that type.

The boundary types available at present are: 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17.

Inlet Outlet Pressure Non-reflective pressure Stagnation Non-reflective stagnation Wall Baffle Symmetry plane Cyclic Free-stream transmissive Transient-wave transmissive Riemann Invariant Attachment Radiation Monitoring Phase-escape (Degassing)

The extent of the information required to define each boundary properly depends in many cases on the variables being solved. For example, in problems using the k-ε model, an inlet boundary needs information concerning the turbulence quantities k and ε. In most cases, the appropriate variables are activated automatically as a result of choosing a given modelling option, e.g. Version 4.02

4-5


Chapter 4

Boundary Region Definition

• •

In the “Molecular Properties” STAR GUIde panel, the ideal gas option for density will switch the density solver on In the “Turbulence Models” panel, any of the K-Epsilon options will switch on the k, ε and viscosity solver

Note that: 1. In the case of a variable such as temperature, you need to switch on the temperature solver explicitly (in panel “Thermal Models”) before proceeding with region definitions. 2. Specification of alternative sets of variables needed to completely define boundaries of type ‘Inlet’ or ‘Pressure’ is possible, as discussed in the sections dealing with such boundaries. 3. It is possible to check for common mistakes in prescribing boundary conditions (e.g. boundary velocities specified in an undefined local coordinate system) by using the facilities available within the “Check Everything” STAR-GUIde panel. 4. Boundary regions may be given an optional alphanumeric name to help distinguish one region from another more easily. The easiest way of applying a desired boundary condition to a given region is via the STAR GUIde system; go to the Define Boundary Conditions folder and open the “Define Boundary Regions” panel, as in the example shown below:

The number and purpose of the text boxes appearing in the panel and whether they are active or not depends on • • 4-6

the type of condition selected; which variables are being solved for. Version 4.02

Chapter 4

BOUNDARY AND INITIAL CONDITIONS Boundary Region Definition

On the other hand, all forms of the panel possess a number of common features, listed below: 1. New regions are defined by: (a) Selecting an unused region in the boundary regions scroll list (b) Choosing the desired boundary condition via the Region Type menu options. The effect of this is to immediately display input boxes for supplying boundary values for all flow variables required. (c) Typing an optional name in the Region Name text box 2. Modification of existing regions is performed in a similar way. The changes are made permanent by clicking the Apply button. 3. Additional boundary regions with identical properties to a pre-defined base region set may also be generated by typing command RGENERATE in the pro-STAR I/O window. 4. Selected region definitions can be deleted by clicking Delete Region. 5. The Compress button eliminates all deleted or undefined regions from the boundary regions scroll list and renumbers the remaining ones contiguously. 6. All free surfaces in your model that are neither defined as boundaries nor explicitly assigned to a region will become part of region no. 0 (shown in the example above). The latter’s properties may be specified in the same way as for any other region. By default, this region is assumed to be a smooth, stationary, impermeable, adiabatic wall. 7. Non-uniform or time-varying conditions may be specified for some boundary types. This is done by choosing one of the following from the Options menu (the default setting, Standard, means constant and uniform conditions): (a) User — specify the required conditions in one of the user subroutines listed below (see also Chapter 14): i) ii) iii) iv) v) vi) vii) viii) ix) x)

BCDEFI — Inlet BCDEFO — Outlet BCDEFP — Pressure BCDNRP — Non-reflective pressure BCDEFS — Stagnation BCDNRS — Non-reflective stagnation BCDEFW — Wall or Baffle BCDEFF — Free-stream transmissive BCDEFT — Transient-wave transmissive BCDEFR — Riemann invariant

The panel also displays a Define user coding button. Click it to store the default source code in sub-directory ufile, ready for further editing. (b) Table — use values stored in a table file as boundary conditions. The file name is of form case.tbl (see Chapter 2, “Table Manipulation”) and may be entered in the Table Name text box. Alternatively, the file may be selected using pro-STAR’s built-in browser. Note that whilst one table can be applied to multiple boundary regions, multiple tables cannot be applied to the same boundary region. A list of Version 4.02

4-7


Chapter 4

Boundary Region Definition

valid dependent variable names that may be used in tables is given for each boundary type in the sections that follow. In addition, the coordinate system used in a table must be the same as the coordinate system specified for its associated boundary regions. Table values are actually assigned to a boundary by the STAR-CD solver during the analysis. This is done as follows: i) Table data are mapped onto the appropriate boundary region in the mesh ii) Boundary face-centre coordinates are compared with the table coordinates iii) Variable values at face centres are calculated from the table data using inverse distance-weighted interpolation iv) The resulting values are assigned to the boundary for the whole duration of the analysis Figure 4-3 shows an example of using a table to assign boundary conditions to a computational boundary. The coordinates and user-supplied values are stored at the nodes of the table data grid and the STAR flow variables are stored at the boundary face centres. In the example, boundary values at face centre 1 are calculated as a weighted average of the table data located at ABCD. Similarly, values at face centre 2 are a weighted average of the table data located at EFGH. Table data map

Table data node

Boundary mesh

B

Boundary face centre

A 1

C

F

E

D 2 G

Figure 4-3

H

Mapping and interpolation of table data onto a boundary

Please also note the following: i) It is possible to produce contour or vector plots of the boundary conditions specified by the table, as a means of checking that the table values have been entered correctly. To do this, click Plot Boundary after you have read in the table and then specify which flow variables you wish to plot. 4-8

Version 4.02

Chapter 4

BOUNDARY AND INITIAL CONDITIONS Inlet Boundaries

ii) The use of boundary condition tables is not supported for cases using the load-step method to define transient conditions (see Chapter 5, “Load-step based solution mode”) (c) GT-POWER — set up a link with the GT-POWER engine system simulation tool (see Chapter 11 of the Supplementary Notes). This provides automatic updating of boundary conditions at inlet and/or pressure boundaries during engine simulation runs. Note that this facility becomes active only after the relevant option has been selected in the “Miscellaneous Controls” panel. (d) Rad. Eq. Tip — impose a radial equilibrium condition by specifying the static pressure at the tip of a turbomachinery case. (e) Rad. Eq. Hub — impose a radial equilibrium condition by specifying the static pressure at the rotor hub of a turbomachinery case.

Inlet Boundaries Introduction This condition describes an inflow boundary and thus requires specification of inlet fluxes for • • • • •

mass momentum turbulence quantities energy chemical species mass fraction

as appropriate. The same boundary type may also be used to specify an outflow condition (i.e. ‘negative inlet’). Note that boundary values are needed only for variables pertinent to the problem being analysed (see “Boundary Region Definition” on page 4-5). In specifying turbulence quantities, it is possible to select in advance the form in which the required boundary values will be input. It is also possible to specify how mass influx is treated under subsonic compressible flow conditions. The choices are made in the “Define Boundary Regions” panel for inlets, as shown in the example below, and are fully described in the “Inlet” on-line Help topic.

Version 4.02

4-9


Chapter 4

Inlet Boundaries

Useful points 1. If the Flow Switch and Turb. Switch settings are changed after velocity components and turbulence boundary conditions have been input, the existing values are not converted in any way, but are interpreted differently. You should therefore use “Define Boundary Regions” to correct these values. 2. Boundary values for turbulence in domains using a Reynolds Stress model may be specified solely in terms of k and ε instead of Reynolds Stress components. If this option is chosen, turbulence conditions at the boundary are assumed to be isotropic. 3. At negative inlets, i.e. inlet boundaries with velocity components pointing out of the solution domain, values for temperature, turbulence quantities and chemical species mass fractions are ignored. 4. Special considerations apply to tetrahedral meshes or meshes containing trimmed (polyhedral) cells. If such meshes contain supersonic inlet boundaries then, to obtain a stable/convergent solution, it is necessary to create at least two cell layers immediately next to the boundary (see Figure 4-5 on page 4-23). If pro-STAR’s automatic meshing module is employed for this purpose, use its built-in mesh generation capabilities. If the mesh is imported from a package that lacks these facilities, you must extrude the mesh in a direction normal to the boundary and then shift the boundary location to the edge of the newly-created, layered structure. 5. If boundary conditions are set using a table (see page 4-7), the permissible variable names that may appear in the table and their meaning is as follows: (a) U — U-component of velocity 4-10

Version 4.02

Chapter 4

BOUNDARY AND INITIAL CONDITIONS Outlet Boundaries

(b) V — V-component of velocity (c) W — W-component of velocity (d) TE — Turbulence kinetic energy or intensity, depending on the Turb. Switch setting (e) ED — Turbulence kinetic energy dissipation rate or length scale, depending on the above setting (f) UU - Reynolds stress component (g) VV - Reynolds stress component (h) WW - Reynolds stress component (i) UV - Reynolds stress component (j) VW - Reynolds stress component (k) UW - Reynolds stress component (l) T — Temperature (absolute) (m) DEN — Density (n) Scalar_name — Mass fraction; use the scalar species name, e.g. H2O, N2 etc. as the scalar variable name(s) You must also ensure that the coordinate system used is the same as the coordinate system specified in the “Define Boundary Regions” panel. 6. When running a transient, compressible flow case in which the mass flux is to be maintained at a constant value (the Flow Switch menu setting is Mass Flux), you must specify both velocity components and density as inlet conditions. This applies to all methods of boundary condition input, i.e., Standard STAR GUIde panel entry, User coding or Table. 7. Inlet boundaries should only be placed on the external surfaces of a fluid domain.

Outlet Boundaries Introduction This condition should be applied at locations where the flow is outwardly directed but the conditions are otherwise unknown. There are two types of outlet boundary: 1. Prescribed flow split boundary. The conditions that must be observed are: (a) The specified split factor f s must be positive. (b) Flow splits for all outlet regions belonging to a given fluid domain should sum to unity, i.e.

∑

fs = 1

(4-1)

(c) This type of boundary must not be used in combination with a pressure or a stagnation pressure boundary within the same fluid domain. (d) This type of boundary must not be used for transient compressible flow cases 2. Prescribed mass outflow rate boundary. The conditions that must be observed in this case are: (a) The specified outflow rate m˙ out must be positive. (b) This type of boundary must be used in combination with at least one Version 4.02

4-11


Chapter 4

Pressure Boundaries

pressure boundary. The desired boundary type is imposed via the “Define Boundary Regions” panel for outlets, as shown in the example below, and is fully described in the “Outlet” on-line Help topic.

Useful points 1. Outlet boundaries of the two basic types described above must not coexist in the same domain. 2. For solution stability and accuracy, outlet boundaries should be used only far downstream of strong recirculation areas, where it is reasonable to expect true outflow everywhere on the boundary. 3. Prescribed mass outflow boundaries are recommended for obtaining fully developed flow in pipes, channels, etc. 4. The difference between outflow conditions described using negative inlet as opposed to prescribed mass outflow boundaries is that the former prescribes both the velocity distribution as well as the mass rate, whereas the latter prescribes only the mass rate. 5. If boundary conditions are set using a table (see page 4-7), only one variable name FSORMF, is allowed. The meaning of this variable is either flow split or mass outflow rate, depending on the Condition pop-up menu setting described above. Note that the variable must be a function of time only. 6. Outlet boundaries are incompatible with: (a) Transonic flows (b) Cavitating flows 7. Outlet boundaries should only be placed on the external surfaces of a fluid domain.

Pressure Boundaries Introduction This condition specifies a constant static pressure or piezometric pressure on a given boundary. For turbomachinery cases, it is also possible to specify the static pressure at the tip or hub and impose a pressure distribution that satisfies radial 4-12

Version 4.02

Chapter 4

BOUNDARY AND INITIAL CONDITIONS Pressure Boundaries

equilibrium. The direction and magnitude of the flow are determined as part of the solution. Thus, • •

if the flow is directed outwards, the values of the other variables are extrapolated from the upstream direction; if the flow is directed inwards, the values are obtained from the supplied boundary conditions.

In specifying turbulence quantities, temperature, mass fraction or (optional) tangential velocity components, it is possible to select in advance the way in which these quantities will be determined. The choices are made in the “Define Boundary Regions” panel for pressure boundaries, as shown in the example below, and are fully described in the “Pressure Boundary” on-line Help topic. However, any such conditions are only applied if the flow direction is towards the solution domain interior.

Useful points 1. For a given fluid domain, pressure boundaries must not coexist with outlet boundaries of the ‘Flow Split’ type. 2. Analyses with multiple pressure boundaries inherently converge more slowly than those where the inlet flow rates and flow splits have been specified. 3. Numerical instability may occur when large or curved surfaces are used as pressure boundaries. 4. It is advisable to choose a reference pressure that is of the same order as the pressure values on the boundaries. For example, if the model contains two boundaries at 10 and 11 bars a reasonable reference pressure would be 10 Version 4.02

4-13


Chapter 4

Stagnation Boundaries

5. 6.

7.

8.

9.

10.

bars. This practice will help to avoid start-up difficulties and to minimise problems due to machine round-off errors. If the Turb. Switch setting is changed to Zero Grad after turbulence boundary conditions have been input, the values already supplied are ignored. If the piezometric setting is chosen for problems involving buoyancy driven flow, you must ensure that the datum level location and density (as specified in the “Buoyancy” panel) are for a point lying on the pressure boundary itself. In cases where a pressure boundary coexists with another pressure or stagnation boundary, it is recommended that the user supplies an estimate for the maximum velocity within the solution domain in the relevant text box of the “Initialisation” panel (see also “Solution Domain Initialisation” on page 4-42). To obtain a stable/convergent solution for tetrahedral meshes or meshes containing trimmed (polyhedral) cells, use of the UVW On option (i.e. explicit velocity specification, see the “Pressure Boundary” STAR GUIde panel) is recommend. Any type of mesh may be used for problems containing radial equilibrium boundaries but only one such region must be employed in the model. Note also that in cases of high circumferential velocity gradients in the radial direction, the user may change the number of averaging intervals to capture the problem details more accurately. The default interval value (50) is however adequate for most cases. If boundary conditions are set using a table (see page 4-7), the permissible variable names that may appear in the table and their meaning is as follows: (a) (b) (c) (d) (e)

PR — Pressure (relative) TE — Turbulence intensity ED — Turbulence length scale T — Temperature (absolute) Scalar_name — Mass fraction; use the scalar species name, e.g. H2O, N2 etc. as the scalar variable name(s)

11. When option Mean On is used (see the “Pressure Boundary” STAR GUIde panel), the scope for tabular input of pressure is limited. Temporal variations in pressure may be prescribed through tabular input, but not spatial variations. 12. Pressure boundaries should only be placed on the external surfaces of a fluid domain.

Stagnation Boundaries Introduction This condition is typically used on a boundary lying in a large reservoir where fluid properties are not significantly affected by flow conditions in the solution domain. It normally appears in compressible flow calculations, but you may also employ it for incompressible flows. Information relevant to such a region is supplied in the “Define Boundary Regions” panel for stagnation boundaries, as shown in the example below, and is described in the “Stagnation Boundary” on-line Help topic.

4-14

Version 4.02

Chapter 4

BOUNDARY AND INITIAL CONDITIONS Stagnation Boundaries

Useful points 1. If a fluid domain contains a stagnation boundary, it must also contain a pressure boundary. 2. Boundary values for turbulence in domains using a Reynolds Stress model may be specified solely in terms of k and ε instead of Reynolds Stress components. If this option is chosen, turbulence conditions at the boundary are assumed to be isotropic. 3. It is recommended that the user supplies an estimate for the maximum velocity within the solution domain via the relevant text box of the “Initialisation” panel (see also “Solution Domain Initialisation” on page 4-42). This will ensure that the calculations start with a reasonable initial velocity field. 4. For a given fluid domain, stagnation boundaries must not co-exist with outlet boundaries of the ‘Flow Split’ type. 5. To obtain a stable/convergent solution for tetrahedral meshes or meshes containing trimmed (polyhedral) cells, it is necessary to create at least two cell layers immediately next to the boundary (see Figure 4-5 on page 4-23). If pro-STAR’s automatic meshing module is employed for this purpose, use its built-in mesh generation capabilities. If the mesh is imported from a package that lacks these facilities, you must extrude the mesh in a direction normal to the boundary and then shift the boundary location to the edge of the newly-created, layered structure. 6. Stagnation boundaries are incompatible with: (a) Eulerian multi-phase flows Version 4.02

4-15


Chapter 4

Non-reflective Pressure and Stagnation Boundaries

(b) Free surface flows (c) Cavitating flows 7. If boundary conditions are set using a table (see page 4-7), the permissible variable names that may appear in the table and their meaning is as follows: (a) (b) (c) (d) (e) (f)

DCX — Direction cosine for U-component of velocity DCY — Direction cosine for V-component of velocity DCZ — Direction cosine for W-component of velocity PSTAGB — Stagnation pressure (relative) TSTAG — Stagnation temperature (absolute) TINTB — Turbulence kinetic energy or intensity, depending on the Turb. Switch setting (g) TLSCB — Turbulence kinetic energy dissipation rate or length scale, depending on the Turb. Switch setting (h) Scalar_name — Mass fraction; use the scalar species name, e.g. H2O, N2 etc. as the scalar variable name(s)

You must also ensure that the coordinate system used is the same as the coordinate system specified in the “Define Boundary Regions” panel. 8. Stagnation boundaries should only be placed on the external surfaces of a fluid domain.

Non-reflective Pressure and Stagnation Boundaries Introduction This type of boundary condition was specially developed for turbomachinery applications. It may only be used in situations where the working fluid is an ideal gas and the flow is compressible. Furthermore, it requires the presence of periodic (cyclic) boundaries in a transverse direction relative to the dominant flow direction, as illustrated in Figure 4-4 below.

4-16

Version 4.02

Chapter 4

BOUNDARY AND INITIAL CONDITIONS Non-reflective Pressure and Stagnation Boundaries

Circumferential direction Wall

Cyclic boundary

Flow (axial) direction Cyclic boundary Wall Figure 4-4

Example of non-reflecting boundary mesh structure

Boundaries of this kind are frequently used as non-reflective pressure/stagnation pairs. The information required for each type represents the average value of the dependent variables that need to be satisfied by the simulation and is supplied in the “Define Boundary Regions” panel. The relevant form of this panel for non-reflecting stagnation boundaries is shown in the example below and is fully described in the “Non-reflective Stagnation Boundary” on-line Help topic.

Version 4.02

4-17


Chapter 4

Non-reflective Pressure and Stagnation Boundaries

The panel for a non-reflecting pressure boundary is shown below and is fully described in the “Non-reflective Pressure Boundary” on-line Help topic.

Useful points 1. Non-reflective pressure and stagnation conditions impose a number of restrictions on the type of mesh employed at the boundary surface: (a) The boundary must contain only quadrilateral faces, aligned along the circumferential direction as shown in Figure 4-4. (b) The cell layer adjacent to the boundary must contain only hexahedral cells 2. Such conditions cannot be assigned to boundary region no. 0. 3. Each strip of boundary faces along the circumferential direction must be assigned to a different non-reflective region number. However, these regions can have the same boundary conditions. 4. The boundary surface must be delimited by cyclic boundaries along the transverse direction, as shown in Figure 4-4. 5. If N is the number of cells along the circumferential direction, the maximum number of harmonics to be used by the Discrete Fourier Transform algorithm is N/2 -1. The minimum number is 0. 6. To ensure that the analysis runs smoothly, it may be necessary to start the simulation by using standard pressure and stagnation boundary conditions over a number of iterations. This can then be followed by a restart run where the non-reflecting boundaries have been applied. 7. For a given fluid domain, non-reflective boundaries must not coexist with outlet boundaries of the ‘Flow Split’ type. 8. At present, certain physical features must not be present in cases containing non-reflective boundaries. The excluded features are: (a) (b) (c) (d) (e) 4-18

Transient calculations Chemical reactions and scalar variables Radiation Reynolds Stress and V2F turbulence models Two-phase flow Version 4.02

Chapter 4

BOUNDARY AND INITIAL CONDITIONS Wall Boundaries

(f) Moving meshes (g) Liquid films (h) Free surface and cavitation 9. If boundary conditions are set using a table (see page 4-7), the permissible variable names that may appear in the table and their meaning is as follows: (a) Non-reflective pressure boundaries i) PR — Pressure (relative static) ii) TE — Turbulence kinetic energy or intensity, depending on the Turb. Switch setting iii) ED — Turbulence kinetic energy dissipation rate or length scale, depending on the above setting (b) Non-reflective stagnation boundaries i) ii) iii) iv) v) vi)

DCX — Direction cosine for U-component of velocity DCY — Direction cosine for V-component of velocity DCZ — Direction cosine for W-component of velocity PSTAGB — Stagnation pressure (relative) TSTAG — Stagnation temperature (absolute) TINTB — Turbulence kinetic energy or intensity, depending on the Turb. Switch setting vii) TLSCB — Turbulence kinetic energy dissipation rate or length scale, depending on the above setting The user must also ensure that the coordinate system used is the same as the coordinate system specified in the “Define Boundary Regions” panel. 10. Non-reflective boundaries should only be placed on the external surfaces of a fluid domain.

Wall Boundaries Introduction STAR-CD’s implementation of wall boundaries involves a generalisation and extension of the no-slip and impermeability conditions commonly used at such surfaces. Thus, a wall boundary may be defined as: •

• •

• • • Version 4.02

Of the no-slip or slip type. The latter is applicable to inviscid flows (in practice µ is set to 10–30 Pa s). The no-slip boundary conditions for turbulent flow are implemented using one of the methods discussed in Chapter 3, “Turbulence Modelling”. Smooth or rough. Moving or stationary. A wall may move within the surface it defines. If motion normal to that surface is desired, use the moving mesh features discussed in Chapter 12, “Moving Meshes”. Permeable or impermeable to heat and/or mass flow. Resistant or not to heat flux due to a thermal boundary layer or intervening solid material. Radiating or non-radiating (see also Chapter 7). 4-19


Chapter 4

Wall Boundaries

As with other boundaries, wall boundary values are needed only for variables pertinent to your problem (see also “Boundary Region Definition” on page 4-5). These are specified via the “Define Boundary Regions” panel for walls, shown in the example below, and are fully described in the “Wall” on-line Help topic.

Thermal radiation properties In thermal radiation problems: 1. Values for thermal emissivity, reflectivity and transmissivity [dimensionless] are required (see also Chapter 7). These should be typed in the text boxes provided. 2. The thermal absorptivity is calculated as (1- reflectivity - transmissivity). 3. The defaults are those for a black body (emissivity equal to 1.0, reflectivity and transmissivity equal to 0.0). Kirchoff’s law (emissivity = absorptivity) is not enforced by the solver. For your wall boundary condition to obey Kirchoff’s law, you must enter the condition: emissivity = absorptivity = 1 - transmissivity - reflectivity Solar radiation properties In solar radiation problems: 1. External walls must be declared as Exposed or Unexposed to incident radiation, by selecting the appropriate option from the Solar Heating pop-up 4-20

Version 4.02

Chapter 4

BOUNDARY AND INITIAL CONDITIONS Wall Boundaries

menu. Note that: (a) This option does not apply to internal ‘walls’ (i.e. baffles and solid/fluid interfaces) (b) The external (FASTRAC) DTRM method distinguishes between direct and diffuse solar radiation — see topic Solar Radiation. However, the above option affects both of them equally. 2. The thermal resistance of an exposed wall to incident solar radiation is neglected. 3. Walls can be made transparent to incident radiation, in which case a value of transmissivity [dimensionless] should be supplied in the text box provided. Thus, the direct solar radiation received by walls can be (a) absorbed, (b) reflected as diffuse radiation, or (c) transmitted. 4. Direct radiation transmitted through transparent walls (e.g. windows), is tracked along the angle of solar inclination (specified via the Solar Radiation option in the “Thermal Options” panel) until it falls on an obstructing surface. 5. The remaining user input depends on the problem conditions: (a) If only solar radiation is present: i) The reflected diffuse radiation is neglected ii) The absorptivity is calculated as (1 – transmissivity) (b) If both thermal and solar radiation are present: i) The code treats the two radiation components separately ii) Values of reflectivity and transmissivity for each component are supplied in separate text boxes and the corresponding absorptivity calculated as (1 – reflectivity – transmissivity) iii) Choosing the internal DTRM method for radiation calculations has the effect of making the solar transmissivity equal to the thermal transmissivity. Other radiation modelling considerations • • •

Version 4.02

The FASTRAC method must be used for thermal/solar radiation problems with transmissive external walls. If the FASTRAC method is used, it is necessary to specify the transmissivity value prior to the view factor calculation. The user input required under the various combinations of thermal and/or solar radiation conditions may be conveniently summarised in the table below:

4-21


Chapter 4

Wall Boundaries .

Table 4-1: Summary of radiative surface property requirements Property Condition Emissivity

Reflectivity

Absorptivity

Transmissivity

Exposure

Thermal

Y

Y

N (=1-R-T)

Y

N

Thermal & Solar

Y N (=0)

Y Y

N (=1-R-T) N (=1-R-T)

Y Y

Y

Solar

N (=0)

N (=0)

N (=1-T)

Y

Y

Useful points 1. For stationary mesh cases, only velocities in directions parallel to the wall surface may be specified, e.g. a planar wall can move only within its own plane. For moving mesh cases, all velocity components should in general be specified. 2. Wall function and two-layer models can be used with any kind of mesh. However, for tetrahedral meshes or meshes containing trimmed (polyhedral) cells, it is advisable to create at least one cell layer immediately next to the wall boundary (see Figure 4-5 below). If pro-STAR’s automatic meshing module is employed for this purpose, use its built-in mesh generation capabilities. If the mesh is imported from a package that lacks these facilities, you must extrude the mesh in a direction normal to the boundary so that the wall is located at the edge of the newly-created, layered structure. 3. The practice recommended above is particularly important for wall boundaries that strongly influence the character of the flow. 4. If boundary conditions are set using a table (see page 4-7), the permissible variable names that may appear in the table and their meaning is as follows: (a) (b) (c) (d) (e) (f)

U — U-component of wall velocity V — V-component of wall velocity W — W-component of wall velocity TORHF — Wall temperature (absolute) or heat flux RESWT — Wall thermal resistance Scalar_name — Mass fraction at the wall; use the scalar species name, e.g. H2O, N2 etc. as the scalar variable name(s) (g) Scalar_name-RSTSC — Wall resistance for a given species, e.g. H2O-RSTSC, N2-RSTSC, etc.

You must also ensure that the coordinate system used is the same as the coordinate system specified in the “Define Boundary Regions” panel. 5. Wall boundaries should only be placed on the surfaces of a fluid or solid domain. They are the only valid boundary type for the interfaces between solid and fluid domains. 6. Scalar boundary conditions at solid-fluid interfaces must be set to zero flux.

4-22

Version 4.02

Chapter 4

BOUNDARY AND INITIAL CONDITIONS Baffle Boundaries

Figure 4-5

Example of tetrahedral plus layered mesh structure

Baffle Boundaries Introduction Baffles are zero-thickness cells within the flow field. They represent solid or porous domains whose physical dimensions are much smaller than the local mesh dimensions, as shown in the example of Figure 4-6.

Figure 4-6

Example model with baffles: duct bend with turning vanes

Baffle ‘cells’ are normally defined via the Cell Tool, as described in Chapter 2, page 2-48 of the Meshing User Guide and should be placed on cell faces inside a fluid domain. If no boundary conditions are specified for the baffle surfaces, they are assumed to be smooth, stationary, impermeable, adiabatic walls. If one needs to specify any other conditions, it is necessary to define special boundaries (called baffle boundaries) explicitly on the baffle surfaces. These boundaries can then be grouped into regions and the “Define Boundary Regions” panel can be used to apply the desired conditions, as shown in the example below. Version 4.02

4-23


Chapter 4

Baffle Boundaries

The discussion of porous media in Chapter 6 also applies, in a modified form, to porous baffles. Thus, it is possible to calculate such a flow by formulating the porous media equation in terms of a pressure drop, ∆p , across the baffle. The definition of the baffle resistance coefficients is also adjusted to account for this change. Obviously, it is now necessary to provide only one pair of such coefficients. Setting up models Inputs for baffle regions are very similar to inputs for walls, including a choice between wall functions and the two-layer model (see the “Baffle” on-line Help topic). There are a few exceptions which are noted below: 1. It is usually possible to impose different boundary conditions on either side of the baffle. As shown in the example dialog above, conditions for Side 1 are supplied first. It is then necessary to click the Apply button, which displays the Side 2 dialog and a message to enter appropriate parameters for that side. Once this is done, the process should be completed by clicking Apply a second time. 2. The numbering of the sides is based on the manner in which the baffle was defined. Side 1 is the ‘outward normal’ side as defined by the cross product of two vectors pointing from the first node to the second node and from the first node to the fourth node, as shown in Figure 4-7.

4-24

Version 4.02

Chapter 4

BOUNDARY AND INITIAL CONDITIONS Baffle Boundaries 4

3,3

3

1

1 Side 2

Side 2

2 Side 1

Figure 4-7

2 Side 1

Numbering convention for various baffle shapes

Another way of determining side numbers is to view the baffle cell and consult the cell definition. If the ordering of the cell vertices is counter clockwise, you are viewing Side 1. 3. The fact that the boundary conditions can be designated separately for each side enables the user to have one side moving and the other stationary or one side isothermal and the other side adiabatic. The conditions can be mixed in any combination with two exceptions: (a) If the thermal boundary condition for Side 1 of the baffle is Conduction, STAR-CD calculates the one-dimensional heat transfer across the baffle based on the local temperature and flow conditions on either side. This choice of boundary condition naturally excludes a different choice for Side 2 and therefore the Wall Heat pop-up menu is deactivated for that side. An exception to this rule occurs when thermal radiation is switched on, in which case radiation properties for both sides of the baffle need to be supplied. (b) In a similar way, if the baffle is porous, only one set of resistance coefficients is needed. The required values are supplied as input for Side 1. Since these naturally apply to the entire baffle, no input is necessary for Side 2. Specific input required for baffles is fully described in the STAR GUIde “Baffle” Help topic. The user should supply values first for Side 1 and then for Side 2 (with the exceptions noted above). Thermal radiation properties In thermal radiation problems: 1. Values for thermal emissivity, reflectivity and transmissivity [dimensionless] are required (see also Chapter 7). These should be typed in the text boxes provided. 2. The absorptivity is calculated as (1 – reflectivity – transmissivity). 3. The defaults are those for a black body (emissivity equal to 1.0, reflectivity Version 4.02

4-25


Chapter 4

Baffle Boundaries

and transmissivity equal to 0.0). The effect of baffle transmissivity is taken into account during the view factor calculations. Therefore, any changes in transmissivity during the run (for example, as part of a transient calculation) will activate beam tracking and a re-calculation of view factors. Note that use of transparent baffles is restricted to surface-to-surface radiation only and thus excludes participating media radiation. Note also that Kirchoff’s law (emissivity = absorptivity) is not enforced by the solver. For your baffle boundary condition to obey Kirchoff’s law, you must enter the condition: emissivity = absorptivity = 1 - transmissivity - reflectivity Solar radiation properties In solar radiation problems, user input depends on the problem conditions: 1. If solar radiation only is present: (a) It is assumed to be completely absorbed by the baffle (i.e. absorptivity = 1) (b) The reflected diffuse radiation is neglected (c) As a result, no user input is required 2. If both thermal and solar radiation are present: (a) Values for emissivity, reflectivity and transmissivity [dimensionless] are supplied separately for each radiation component (b) Choosing the internal DTRM method for radiation calculations has the effect of making the solar transmissivity equal to the thermal transmissivity. Therefore, only a reflectivity value needs to be supplied for the solar component. Other radiation modelling considerations • •

If the FASTRAC method is used, it is necessary to specify the transmissivity value prior to the view factor calculation. The user input required under the various combinations of thermal and/or solar radiation conditions may be conveniently summarised in the table below: .

Table 4-2: Summary of radiative surface property requirements Property Condition

4-26

Emissivity

Reflectivity

Absorptivity

Transmissivity

Exposure

Thermal

Y

Y

N (=1-R-T)

Y

N

Thermal & Solar

Y N (=0)

Y Y

N (=1-R-T) N (=1-R-T)

Y Y

N

Solar

N (=0)

N (=0)

N (=1)

N (=0)

N

Version 4.02

Chapter 4

BOUNDARY AND INITIAL CONDITIONS Symmetry Plane Boundaries

Useful points 1. For stationary mesh cases, only velocities in directions parallel to the baffle surface may be specified, e.g. a planar baffle can move only within its own plane. For moving mesh cases, all velocity components should in general be specified. 2. If boundary conditions are set using a table (see page 4-7), the permissible variable names that may appear in the table and their meaning is as follows: (a) (b) (c) (d) (e) (f)

U — U-component of baffle velocity V — V-component of baffle velocity W — W-component of baffle velocity TORHF — Baffle temperature (absolute) or heat flux RESWT — Baffle thermal resistance Scalar_name — Mass fraction; use the scalar species name, e.g. H2O, N2 etc. as the scalar variable name(s) (g) Scalar_name-RSTSC — Baffle resistance for a given species, e.g. H2O-RSTSC, N2-RSTSC, etc. You must also ensure that the coordinate system used is the same as the coordinate system specified in the “Define Boundary Regions” panel.

Symmetry Plane Boundaries Symmetry boundaries are used for two purposes: 1. To reduce the size of the computational mesh by placing the boundary along a plane of geometrical and flow symmetry. 2. To approximate a free-stream boundary. No user input is required beyond definition of the boundary location. The quantities set to zero at the boundary are: • •

The normal component of velocity The normal gradient of all other variables

Symmetry boundaries should only be placed on the external surfaces of a fluid or solid domain. They cannot be used in FASTRAC radiation calculations.

Cyclic Boundaries Introduction Cyclic boundaries impose a repeating or periodic flow condition on a pair of geometrically identical boundary regions, numbers 1 and 2 in the example of Figure 4-8. Selected scalar variables are forced to be equal at corresponding faces on the two regions. As shown in Figure 4-8, velocity components are also equalised in a common local coordinate system specified by the user. Such boundaries thus serve to reduce the size of the computational mesh. This is illustrated by the example of Figure 4-9, showing a cascade of repeating baffles.

Version 4.02

4-27


Chapter 4

Cyclic Boundaries

U2 V1

U1

V2

Cyclic boundary 1

Cyclic boundary 2 YL U1 = U2 V1 = V2 W1 = W2

RL ΘL

Figure 4-8

XL

Local cylindrical system

Cyclic conditions defined using a local coordinate system Cyclic boundary 1

Inlet

Cyclic boundary 2

Figure 4-9

Regular cyclic boundaries with integral match

Setting up models Cyclic boundaries are defined using STAR GUIde panels in the following multistage process: 1. In panel “Create Boundaries”, use tab “Regions” to set up a pair of regions, of identical size and shape, and designate them as cyclic 2. In tab “Cyclics”, specify a number of parameters that enable them to be matched to each other geometrically and which take into account the mesh characteristics at either end. This involves the following considerations: (a) Specification of suitable coordinate increments (offsets) that allow one 4-28

Version 4.02

Chapter 4

BOUNDARY AND INITIAL CONDITIONS Cyclic Boundaries

member of the pair to be located if one starts at the other member. A local coordinate system in which the regions are matched is also specified. (b) Whether the regions form a regular cyclic (as in Figure 4-9) or an anticyclic pair (as in Figure 4-10). The latter appears in problems where all flow variable profiles have to be reversed in a specified direction of the matching coordinate system. This operation also reverses the coordinate value of each boundary face in that direction before adding the corresponding offset. Thus, placing the coordinate system origin on an axis of symmetry and choosing its location carefully can eliminate the need for offsets, as in the anticyclic system shown in Figure 4-10. Cyclic boundary 1

Local Cartesian coordinate system

Cyclic boundary 2

Figure 4-10

Partial anticyclic boundaries with integral match

(c) Whether there is a one-to-one correspondence between boundary faces on either side of the cyclic pair, as in the examples shown in Figure 4-9 and Figure 4-10. This requires a so-called integral matching operation. If no such correspondence exists, typically because one side is more finely meshed than the other (as in Figure 4-11), the system requires an arbitrary matching operation. The latter is similar to matching cell faces on either side of an interface between mesh blocks (see Chapter 3, “Arbitrary connectivity” in the Meshing User Guide). It thus involves matching of so-called master boundary faces on one side of the cyclic pair with slave faces on the other side. 3. In tab “Cyclics”, finish up by performing the geometric matching operation between boundary faces on either side of the pair to form so-called cyclic sets. Note that the same operation may also be performed manually, whereby each cyclic set and the boundaries contributed by each cyclic pair member are named explicitly using command CYCLIC. The list of cyclic pairs can also be extended with the CYGENERATE command, beginning from a pre-existing starting set.

Version 4.02

4-29


Chapter 4

Cyclic Boundaries

Figure 4-11

Cyclic boundaries with arbitrary match

4. In panel “Define Boundary Regions”, specify the physical cyclic boundary conditions that exist between the members of the pair, as shown below:

These can be of two types: (a) Ordinary cyclic conditions, whereby all flow variable values on one member are matched with the corresponding values on the other member. (b) Partial cyclic conditions, whereby the matching process is subject to an additional constraint of either a prescribed pressure drop or a fixed mass flow rate across the cyclic pair. An example of a fixed mass flow rate system, representing one half of a continuous loop flow system, is shown in Figure 4-10. For thermal problems, the bulk mean temperature on the inflow side of the cyclic pair is also required. Useful points 1. One member of the cyclic pair must be designated as an Inflow boundary and the other as an Outflow boundary 4-30

Version 4.02

Chapter 4

BOUNDARY AND INITIAL CONDITIONS Cyclic Boundaries

2. If the partial cyclic condition is specified via the STAR GUIde interface, the Inflow and Outflow sides are indicated via the Flow Direction menu and the pressure drop or flow rate must be a positive number. Note that this number must be the same for both members of the pair. 3. If the partial cyclic condition is specified via the RDEFINE command, the inflow and outflow sides are distinguished by assigning a pressure drop or flow rate to one member that is equal in magnitude but of opposite sign to that for the other member. The sign convention is as follows: (a) Pressure Drop + Inflow – Outflow (b) Flow rate + Outflow – Inflow 4. Partial cyclic conditions can only be applied to boundaries matched in Cartesian coordinates 5. Such conditions are not available for chemical species mass fractions and cannot be used in variable-density flows 6. Arbitrary cyclic matching (see page 4-29 above) is not allowed for partial cyclic conditions 7. Cyclic boundaries cannot be used in FASTRAC radiation calculations. 8. Cyclic boundaries should only be placed on the external surfaces of a fluid domain. Cyclic set manipulation All currently defined cyclic sets are shown in the Cyclic Set List below:

Commands:

CYLIST

CYDELETE

CYCOMPRESS

The list may be displayed by choosing Lists > Cyclic Sets from the main menu bar. The sets are numbered and listed in numerically ascending order, together with their constituent master and slave boundary numbers for arbitrarily matched regions (see page 4-29 above). There is a choice of showing all cyclic sets (click button Show All Cyclic Sets) or just those with at least one member (master or slave boundary) in the current boundary set (click button Show Cyclic Sets with Boundaries in Version 4.02

4-31


Chapter 4

Free-stream Transmissive Boundaries

Bset Only). Items in the second category are marked by asterisks in the Bset column. To select cyclic sets from the list: • •

For single items, click the required set number. For two or more items in sequence, click the first set you want to select, press and hold down the Shift key and then click the last set in the group.

Once the desired sets are selected, the following operations are possible: • •

Deletion — click on the Delete button. Compression — click on the Compress button. This involves the elimination of all deleted cyclic sets and renumbering of the remaining ones.

A third operation, for validating arbitrarily matched cyclic boundaries, is implemented in the “Check Everything” panel. The operation checks that • • •

all sets in a given range exist and reference arbitrarily-matched cyclic regions; there is overlap between boundaries on the two sides of the cyclic set; the overlapping areas from either side match up.

All checks are performed to within a specified tolerance.

Free-stream Transmissive Boundaries Introduction This type of boundary may be used only in models involving supersonic free streams where the working fluid is an ideal gas. The facility enables shock waves generated in the interior of the solution domain to be transmitted, without reflection, through the boundary to the wider (free stream) space surrounding the domain. Flow can be out of the solution domain (compression waves) or into the solution domain (expansion waves). In either case, boundary values of scalar variables are extrapolated from the solution domain interior. In the case of turbulent inflow (expansion waves), the turbulence quantities have to be specified as part of the user input. To set up boundaries of this kind, you need to: 1. Decide on an appropriate location for the boundary, preferably parallel to the main (supersonic) stream. 2. Supply values in the “Define Boundary Regions” panel for all free-stream properties, as shown in the example below. The required input is fully described in the “Free-stream Transmissive Boundary” on-line Help topic.

4-32

Version 4.02

Chapter 4

BOUNDARY AND INITIAL CONDITIONS Free-stream Transmissive Boundaries

The STAR-CD solver calculates the magnitude and direction of the flow at the boundary as part of the analysis, based on the simple wave theory given in [3] and [4]. Useful points 1. A value of temperature at the boundary is obligatory. The user must therefore ensure that temperature calculations are activated, via the “Thermal Models” panel, before defining the boundary conditions. 2. Boundary values for turbulence in domains using a Reynolds Stress model may be specified solely in terms of k and ε instead of Reynolds Stress components. If this option is chosen, turbulence conditions at the boundary are assumed to be isotropic. 3. To obtain a stable/convergent solution for tetrahedral meshes or meshes containing trimmed (polyhedral) cells, it is necessary to create at least two cell layers immediately next to the boundary (see Figure 4-5 on page 4-23). If pro-STAR’s automatic meshing module is employed for this purpose, use its built-in mesh generation capabilities. If the mesh is imported from a package that lacks these facilities, you must extrude the mesh in a direction normal to the boundary and then shift the boundary location to the edge of the newly-created, layered structure. 4. Boundary conditions specified in a table will be applied only if fluid is entering the solution domain from the outside. If this is not the case, i.e.the flow is parallel to the boundary or crossing it from inside the domain, boundary values will be extrapolated from interior values and the table data will not be used. 5. If boundary conditions are set using a table (see page 4-7), the permissible Version 4.02

4-33


Chapter 4

Transient-wave Transmissive Boundaries

variable names that may appear in the table and their meaning is as follows: (a) (b) (c) (d) (e) (f)

UINF — U-component of velocity VINF — V-component of velocity WINF — W-component of velocity PINF — Pressure (relative) TINF — Temperature (absolute) TEINF — Turbulent kinetic energy or intensity, depending on the Turb. Switch setting (g) EDINF — Turbulent kinetic energy dissipation rate or length scale, depending on the Turb. Switch setting

You must also ensure that the coordinate system used is the same as the coordinate system specified in the “Define Boundary Regions” panel. 6. Free-stream transmissive boundaries should only be placed on the external surfaces of a fluid domain.

Transient-wave Transmissive Boundaries Introduction This type of boundary may be used only in transient, compressible flows where the working fluid is an ideal gas. It enables transient waves to leave the solution domain without reflection. STAR-CD uses the simple wave theory to calculate conditions behind the wave and to specify such conditions at the boundaries. To set up boundaries of this kind, you need to: 1. Decide on an appropriate location for the boundary 2. Supply values in the “Define Boundary Regions” panel for all dependent variables, representing conditions outside the boundary (at ‘infinity’). The required input is fully described in the “Transient-wave Transmissive Boundary” on-line Help topic.

4-34

Version 4.02

Chapter 4

BOUNDARY AND INITIAL CONDITIONS Transient-wave Transmissive Boundaries

STAR-CD calculates the magnitude and direction of the flow at the boundary as part of the analysis, based on the transient wave theory given in [3] and [4]. Useful points 1. A value of temperature at the boundary is obligatory. The user must therefore ensure that temperature calculations are activated, via the “Thermal Models” panel, before defining the boundary conditions. 2. Boundary values for turbulence in domains using a Reynolds Stress model may be specified solely in terms of k and ε instead of Reynolds Stress components. If this option is chosen, turbulence conditions at the boundary are assumed to be isotropic. 3. To obtain a stable/convergent solution for tetrahedral meshes or meshes containing trimmed (polyhedral) cells, it is necessary to create at least two cell layers immediately next to the boundary (see Figure 4-5 on page 4-23). If pro-STAR’s automatic meshing module is employed for this purpose, use its built-in mesh generation capabilities. If the mesh is imported from a package that lacks these facilities, you must extrude the mesh in a direction normal to the boundary and then shift the boundary location to the edge of the newly-created, layered structure. 4. Boundary conditions specified in a table will be applied only if fluid is entering the solution domain from the outside. If this is not the case, i.e.the flow is parallel to the boundary or crossing it from inside the domain, boundary values will be extrapolated from interior values and the table data will not be used. 5. If boundary conditions are set using a table (see page 4-7), the permissible Version 4.02

4-35


Chapter 4

Riemann Boundaries

variable names that may appear in the table and their meaning is as follows: (a) (b) (c) (d) (e) (f)

UINF — U-component of velocity VINF — V-component of velocity WINF — W-component of velocity PINF — Pressure (relative) TINF — Temperature (absolute) TEINF — Turbulent kinetic energy or intensity, depending on the Turb. Switch setting (g) EDINF — Turbulent kinetic energy dissipation rate or length scale, depending on the above setting

You must also ensure that the coordinate system used is the same as the coordinate system specified in the “Define Boundary Regions” panel. 6. Transient-wave transmissive boundaries should only be placed on the external surfaces of a fluid domain.

Riemann Boundaries Introduction This type of boundary is typically employed in external aerodynamics simulations and may be used only if the working fluid is an ideal gas. It enables weak pressure waves to leave the solution domain without reflection and is valid for both steady-state and transient problems. To set up boundaries of this kind, you need to: 1. Decide on an appropriate location for the boundary 2. Supply values in the “Define Boundary Regions” panel for all dependent variables, representing conditions outside the boundary (at ‘infinity’). The required input is fully described in the “Riemann Boundary” Help topic.

4-36

Version 4.02

Chapter 4

BOUNDARY AND INITIAL CONDITIONS Riemann Boundaries

STAR-CD calculates the magnitude and direction of the flow at the boundary as part of the analysis, based on the Riemann invariant theory given in [7]. Useful points 1. Check that the density of the domain to which such a boundary belongs is set to Ideal-f(T,P) 2. Boundary values for turbulence in domains using a Reynolds Stress model may be specified solely in terms of k and ε instead of Reynolds Stress components. If this option is chosen, turbulence conditions at the boundary are assumed to be isotropic. 3. A value of temperature at the boundary is obligatory. The user must therefore ensure that temperature calculations are activated, via the “Thermal Models” panel, before defining the boundary conditions. 4. To obtain a stable/convergent solution for tetrahedral meshes or meshes containing trimmed (polyhedral) cells, it is necessary to create at least two cell layers immediately next to the boundary (see Figure 4-5 on page 4-23). If pro-STAR’s automatic meshing module is employed for this purpose, use its built-in mesh generation capabilities. If the mesh is imported from a package that lacks these facilities, you must extrude the mesh in a direction normal to the boundary and then shift the boundary location to the edge of the newly-created, layered structure. 5. Boundary conditions specified in a table will be applied only if fluid is entering the solution domain from the outside. If this is not the case, i.e.the flow is parallel to the boundary or crossing it from inside the domain, boundary values will be extrapolated from interior values and the table data Version 4.02

4-37


Chapter 4

Attachment Boundaries

will not be used. 6. If boundary conditions are set using a table (see page 4-7), the permissible variable names that may appear in the table and their meaning is as follows: (a) (b) (c) (d) (e) (f)

UINF — U-component of velocity VINF — V-component of velocity WINF — W-component of velocity PINF — Pressure (relative) TINF — Temperature (absolute) TEINF — Turbulent kinetic energy or intensity, depending on the Turb. Switch setting (g) EDINF — Turbulent kinetic energy dissipation rate or length scale, depending on the above setting (h) Scalar_name — Mass fraction; use the scalar species name, e.g. H2O, N2 etc. as the scalar variable name(s)

You must also ensure that the coordinate system used is the same as the coordinate system specified in the “Define Boundary Regions” panel. 7. Riemann boundaries should only be placed on the external surfaces of a fluid domain.

Attachment Boundaries Attachment boundaries are used for the following two purposes: 1. To define the interface between cells that may be connected or disconnected from each other (see Chapter 12, “Cell Attachment and Change of Fluid Type”). 2. To define the interface between mesh blocks that slide past each other, either in an ‘integral’ or ‘arbitrary’ manner — see “Regular sliding interfaces” on page 12-18.

Two input parameters are needed: • •

A local coordinate system in which the boundaries are to be matched An alternate boundary region number

The second parameter is required for cell layer attachment cases and serves to maintain appropriate boundary conditions in the solution domain if the cells on either side of the interface become disconnected. The alternate boundary region 4-38

Version 4.02

Chapter 4

BOUNDARY AND INITIAL CONDITIONS Radiation Boundaries

must be of ‘wall’ or ‘inlet’ type. Examples of cases requiring attachment boundaries are given in Chapter 12. Useful points 1. To obtain a stable/convergent solution for meshes containing trimmed (polyhedral) cells, it is necessary to create at least two cell layers immediately next to the boundary (see Figure 4-5 on page 4-23). If the pro-STAR Auto Mesh module is employed, use its built-in mesh generation capabilities for this purpose. If the mesh is imported from a package that lacks such facilities, you must extrude the mesh in a direction normal to the boundary and then shift the boundary location to the edge of the newly-created, layered structure. 2. Attachment boundary regions must be created in pairs, one on each of the mesh blocks that are attached to, detached from or sliding past each other. 3. Attachment boundaries should be placed on the surfaces of domains/ subdomains.

Radiation Boundaries Radiation boundaries are used for the purpose of separating that part of your model where radiation effects are important from other parts where such effects are negligible. This type of boundary only influences radiation calculations and is completely transparent to the fluid flow and non-radiative heat transfer in your model.

Two input parameters are needed (see also Chapter 7): 1. The boundary radiation temperature [K], normally set to a value close to the expected temperature in the surrounding area 2. The boundary surface emissivity [dimensionless], normally set to 1.0 The location and properties of such a boundary should be chosen so that: •

• • Version 4.02

Radiant energy passing through it escapes to the outside world with minimal back-radiation into the sub-domain where it emanated. The escaped radiation should be low enough not to influence conditions in the outside world. Its presence does not adversely affect the accuracy of the calculations inside the radiative sub-domain(s) If a coupled-cell interface exists between the radiative and non-radiative sub-domains, the boundary must be placed on the cells that are inside the 4-39


Chapter 4

Phase-Escape (Degassing) Boundaries

radiative sub-domain. Useful points 1. The use of radiation boundaries is not required in radiation problems employing the Discrete Ordinates method. Turning radiation off for a particular cell type is sufficient to exclude radiation calculations in that mesh region. 2. Radiation boundaries are currently incompatible with FASTRAC radiation calculations. 3. Radiation boundaries should be placed on cell faces inside a fluid domain.

Phase-Escape (Degassing) Boundaries This type of boundary appears exclusively in Eulerian multi-phase problems (see Chapter 10 of this volume) and represents a degassing free-surface bounding a two-phase system of gas bubbles in a liquid, corresponding to the dispersed and continuous phases, respectively. The boundary conditions applied to each phase are as follows: 1. For the continuous phase, the boundary acts like a slip wall, allowing the liquid to flow parallel to the boundary surface without friction 2. For the dispersed phase, the boundary acts like an opening allowing bubbles to escape into the surrounding medium, unless retained within the solution domain by the drag forces acting on them. No further user input is required on the “Define Boundary Regions” panel. Note that only one boundary of this type should be present in your model. Degassing boundaries should only be placed on the external surfaces of a fluid domain.

Monitoring Regions These are arbitrary surfaces, defined in the same way as ordinary boundaries but placed on any cell faces within a fluid or solid domain so as to form internal surfaces. They are used purely for monitoring engineering data such as mass flux (see panel “Monitor Boundary Behaviour”) so no further user input is required on the “Define Boundary Regions” panel. Monitoring regions do not affect the flow field in any way; STAR simply calculates the monitored data values at the specified region’s surface and stores them for subsequent display as a function of time or number of iterations (see panel “Engineering Data”). The same data values are available at monitoring regions as at open boundary regions, except that: • • •

Item Heat Flux is not available Field values are taken from the neighbouring cell centres and are not interpolated to the boundary Item Enthalpy In/Out is based on convection only, so it will be zero in solid materials

Each face of a monitoring region “belongs” to a neighbouring cell, such that the mass flux is defined as being positive when it leaves this cell through the face. This in turn determines the face’s orientation and, for consistent calculation of the total 4-40

Version 4.02

Chapter 4

BOUNDARY AND INITIAL CONDITIONS Boundary Visualisation

mass flux through the region, it is important that all its faces are oriented the same way. The choice of which cell a monitoring region face belongs to is made when that face is defined. Commands BFIND, BCROSS, and BZONE do this by picking a cell face, as do their associated GUI operations, and are therefore suitable for this purpose. On the other hand, commands BDEF, BGEN and BDX should not be used to define monitoring region faces as they do not involve the explicit selection of a cell face and hence the orientation of the monitoring region face is indeterminate. Caution should also be exercised when generating such regions automatically, for example by cell refinement. When visualising monitoring regions, their orientation is indicated using an arrow normal to the boundary and whose direction indicates the direction of positive flux, as shown in Figure 4-12 below.

Figure 4-12

Monitoring region display

Boundary Visualisation As described in “Boundary set selection facilities” on page 4-3, boundaries can be collected into sets. The currently defined set can then be displayed on top of the calculation mesh by choosing Cell Plot Display Option Bound from the main window and re-plotting. The cell faces representing the boundaries will be marked by distinctive fill patterns and colours, characteristic of the boundary type represented. Boundary faces will be superimposed on any kind of plot already displayed on the screen other than a section plot. Note that the boundary display option may also be selected by choosing Plot > Cell Display > Boundaries from the menu bar. Alternatively, you may type commands BDISPLAY, ON or CDISPLAY, BREGION in the I/O window.

Version 4.02

4-41


Chapter 4

Solution Domain Initialisation

Solution Domain Initialisation Steady-state problems User action depends on whether the solution is to start from the initial state of the model (initial run) or to continue from a previously computed solution (restart run). Initial runs Initial conditions for flow field variables are assigned in STAR-GUIde’s Thermophysical Models and Properties folder: •

For fluid field variables, use panel “Initialisation” in the Liquids and Gases sub-folder. Note that there is a separate “Turbulence tab” for initializing turbulence parameters. If this is done by specifying the initial turbulence intensity I and length scale l, the turbulence kinetic energy k and dissipation rate ε are computed as follows: 2 2

k = 1.5U I

(4-2)

1.5

k ε = -------l

(4-3)

where U is the initial velocity magnitude. For turbulence models other than the k-ε type, the turbulence scales (ω for k-ω models or ν t for the Spalart-Allmaras model) are computed automatically. • For chemically reacting flows, you may also need to use panel “Initialisation” in the Additional Scalars sub-folder to specify initial mass fractions for chemical species. • In conjugate heat transfer problems, another panel also called “Initialisation” in the Solids sub-folder can be used to specify initial temperatures in solid materials. Restart runs Various options for this operation are available in panel “Analysis (Re)Start” within the Analysis Preparation/Running folder. If option Standard Restart is chosen, the solution from a previous run serves as the starting point for the current run. If Initial Field Restart is chosen in this panel, the STAR-CD solver only corrects the mass fluxes to satisfy continuity. The Initial Field Restart option should be chosen if any change has been made to the boundary conditions or reference quantities (pressure and/or temperature). This option must also be chosen if new scalars have been defined by selecting additional modelling options such as Lagrangian multi-phase or chemical reaction. Special considerations apply to cases where the restart also involves a change in the mesh configuration, typically a refinement of a coarser starting mesh. These are covered in Chapter 5, “Solution Control with Mesh Changes”. Transient problems In transient problems, all flow field variables should be given the correct values for the problem at hand. Depending on the physical conditions being modelled, this can be done in one of the following ways: 1. Specify uniform values — select option Constant in the “Initialisation” panel 4-42

Version 4.02

Chapter 4

BOUNDARY AND INITIAL CONDITIONS Solution Domain Initialisation

and then type values for each variable in the text boxes provided. Turbulence parameters, scalar mass fractions and solid temperatures may be initialised as described in section “Initial runs” above for steady-state cases. 2. Set values through a user-supplied subroutine — select option User in the Initialization panel and then specify the required distributions in subroutine INITFI. 3. Read in a previously computed distribution that corresponds to the desired setting — select an option from the “Analysis (Re)Start” panel (usually Initial Field Restart plus one of the options in the Initial Field Restart pop-up menu depending on the problem at hand). Option Standard Restart must be used for all moving mesh cases and should also be chosen to start a transient analysis from a previously computed steady-state solution. Note that such restarts should not be performed for Lagrangian multi-phase cases, as the meaning of the droplet treatment is different between steady-state and transient analyses.

Version 4.02

4-43

Chapter 5

CONTROL FUNCTIONS Introduction

Chapter 5

CONTROL FUNCTIONS

Introduction At this stage of modelling, the following tasks should have been completed: 1. Mesh set up 2. Material property and continuum mechanics model specification 3. Definition of boundary type and location The penultimate task before a STAR analysis run is to set the parameters that control that run. This consists of • •

setting various parameters that affect the progress of the numerical solution algorithm used by STAR; specifying the type and amount of run-time output and post-processing data.

The user should also decide whether the problem is steady-state or transient so as to perform the appropriate operations for the above tasks.

Analysis Controls for Steady-State Problems Solution controls Solution control parameters have a strong influence on the progress of the analysis, so it is important to have a basic understanding of their significance and effect during a run. You are therefore advised to refer to Chapter 7 in the Methodology volume for a detailed discussion of under-relaxation and other solution control topics. STAR-CD offers two alternatives for solving steady-state problems: 1. An iterative method employing under- relaxation factors 2. A pseudo-transient time marching to the steady-state solution with a fixed-length time step. Note, however, that an under-relaxation factor (default value 0.2) is still used on the pressure correction equation. Incompressible, non-reacting and low Mach number flows usually converge smoothly and fast in a combination with the inertial under-relaxation shown in equation (7-14) of the Methodology volume. If fluid flows which are characterized by travelling waves (e.g. pressure waves in compressible fluids, or gravity waves in free surface flows) can reach a steady state, the convergence process is typically much more robust and stable if one can resolve to some extent the waves travelling during the iteration process. For this, it is important that waves travel in all cells with the same pseudo time step, and the pseudo-transient mode is usually better suited to this class of problems. In other problems (e.g. inviscid flows and where the initial velocity field is zero), we can have very small or zero values of the central coefficients A P at an early stage of the iteration process. In such cases, local pseudo time steps become very large or infinite (see equation (7-19)), which again has an impact on the convergence and stability of the solution. In flows which exhibit this kind of problem, use of the pseudo-transient mode is recommended. The main advantage of the iterative method employing under-relaxation factors compared with the pseudo-transient mode is that, in the former, the under-relaxation Version 4.02

5-1

CONTROL FUNCTIONS

Chapter 5

Analysis Controls for Steady-State Problems

factors vary between 0 and 1 and a large number of cases run nearly optimally with default values (e.g. 0.7 for momentum and 0.2 for pressure correction) that are based on considerable past experience. However, in the pseudo transient approach, the time step varies between zero and infinity and a suitable value is not always easy to find. The optimum value can be determined only by numerical experiments. As a guideline, one should choose a time step such that the Courant number based on the characteristic velocity and the characteristic mesh size is between 1 and 8. Note that, as with under-relaxation, lower values of time step are more likely to promote convergence, while larger ones lead to a faster solution. The task of setting up solution controls for either of these methods can be divided into the following steps: Step 1 Start up the STAR GUIde system and then define the type of problem you are solving by selecting Steady State from the Time Domain pop-up menu in the “Select Analysis Features” panel Step 2 Go to the Solution Controls folder and open the “Solution Method” panel. From the pop-up menu at the top of the panel select: •

Steady State for conventional steady-state runs. Also choose the numerical algorithm to be used (see topic “Steady-State Solution”). In every case, specify the maximum residual error tolerance (i.e. maximum acceptable level of remaining error in the solution), plus any additional parameters required by the algorithm you have chosen. • Pseudo-Transient for pseudo-transient runs (see topic “Pseudo-Transient Solution”). The maximum residual error tolerance (i.e. maximum acceptable level of remaining error in the solution) should be specified; the normalised residuals are displayed on the screen and also saved on file case.run, as in ordinary steady-state runs. Step 3 In the “Primary Variables” panel, inspect the solution status for flow variables and material properties (see topic “Equation Status”) to confirm that the right variables will be solved for. Step 4 Check the “Solver Parameters” (under-relaxation factors, number of calculation sweeps and residual error tolerances for each solution variable). Step 5 Choose one of the available “Differencing Schemes”. It is suggested that higher-order differencing schemes such as LUD or MARS should be used if high spatial discretisation accuracy is required. Output controls Having set the solution control parameters, the next task is to choose the type and volume of output from the forthcoming STAR run. The bulk of this output consists of solution variable values at cell centroids. Output controls can be applied by going to the Output Controls folder in the STAR GUIde system and following the steps 5-2

Version 4.02

Chapter 5

CONTROL FUNCTIONS Analysis Controls for Steady-State Problems

below: Step 6 Consider whether detailed printout on the solution progress is required and if necessary specify the appropriate settings in the “Monitor Numeric Behaviour” panel. Step 7 Decide whether you want to follow the progress of the analysis by generating various types of monitoring data at every iteration. If so, go to the Monitor Engineering Behaviour sub-folder and use one or both of the following panels: • •

“Monitor Boundary Behaviour” — select one or more boundary regions and the type of monitoring information to be generated for them “Monitor Cell Behaviour” — select one or more sub-domains, defined in terms of cell sets, and the type of monitoring information to be generated for them

The requested data are stored in special files (case.erd and case.ecd for boundary and cell data, respectively), from where they may be displayed as pro-STAR graphs at the end of the analysis (see panel “Engineering Data” in the Post-Processing folder) or read by an external post-processing package. Step 8 Specify the manner of saving mesh data for use in post-processing and/or restart runs via the “Analysis Output” panel (“Steady state problems”). If desired, go to the “Additional Output Data” section to select any wall data to be included in the solution (.ccm) file. This is important, as these settings will affect the availability of data for post-processing. You can also select what wall data are to be ‘printed’ (i.e. displayed on your screen) and stored in the .run file at the end of the run. For both post and print control parameters, it is up to you to check the default settings and change them, if necessary, according to the type of problem being analysed. Other controls Step 9 Go to the Sources sub-folder and inspect the “Source Terms” panel to see if any additional information (such as extra source terms for flow variables) is needed to completely describe your problem. Note that STAR-CD provides special switches and constants for activating various beta-level features in the code, or for turning on calculation procedures designed for debugging purposes. These are found in the “Switches and Real Constants” panel and are normally used only after consultation with CD-adapco. An alternative way of performing this function is to enter special debugging instructions into the Extended Data panel, accessible from the Utilities menu in the main window (or issue command EDATA). Step 10 Go to the Analysis Preparation/Running folder and open the “Set Run Time Controls” panel: • Version 4.02

For conventional steady-state runs, enter the maximum number of iterations 5-3

CONTROL FUNCTIONS

Chapter 5

Analysis Controls for Transient Problems

(or calculation loops, see “Steady state problems”). • For pseudo-transient runs, specify the time step size and the maximum number of time steps. A variable step magnitude may also be specified via user subroutine DTSTEP, by selecting option User in the Time Step Option pop-up menu (see “Steady state problems (Pseudo-transient)”). Step 11 To complete the controls specification, you need to decide whether the analysis is to start from initial conditions or restart from a previous run. Set the appropriate solution controls in the “Analysis (Re)Start” panel.

Analysis Controls for Transient Problems Transient problems can be divided into three groups: 1. Systems whose flow, thermal and chemical fields are originally in thermodynamic equilibrium and which are subjected to a set of non-equilibrium boundary conditions at the start of the calculation. The system’s response is to gradually approach a new steady state. Such problems can be analysed in STAR either in the steady-state or transient mode; some buoyancy driven flows are best run in transient mode (see also Chapter 3, “Buoyancy-driven Flows and Natural Convection”). 2. Systems whose boundary conditions change in a prescribed fashion, e.g. due to opening and shutting of flow valves. 3. Inherently unstable systems that never reach a steady state and exhibit either (a) a cyclic (or periodic) behaviour, as in some vortex shedding problems, or (b) chaotic behaviour, as in some buoyancy driven flows. Procedures for solving all of these problem types are described below. Default (single-transient) solution mode This procedure, referred to as the ‘single-transient’ solution mode in earlier versions of STAR-CD, is the quickest and easiest way of setting up transient problems. It is also suitable for steady-state compressible or buoyancy driven flows that require close coupling between the momentum, enthalpy, chemical species and density equations. Other important characteristics are: • • •

It is fully supported by pro-STAR’s STAR GUIde interface It can accommodate problems with time-varying boundary conditions through the use of tables (see Chapter 2, “Table Manipulation”) Changes in boundary region type (e.g. a pressure boundary changing to a wall boundary) are also possible but require stopping and restarting the analysis at those times when such changes occur

The single-transient mode provides an alternative to the “Load-step based solution mode” discussed below, by eliminating the need for a transient history file and explicit load step definitions. It is in fact equivalent to performing a single load step, hence the name ‘single transient’. To use this approach, follow the procedure below.

5-4

Version 4.02

Chapter 5

CONTROL FUNCTIONS Analysis Controls for Transient Problems

Solution controls Step 1 Start up the STAR GUIde system and then define the type of problem you are solving by selecting Transient from the Time Domain pop-up menu in the “Select Analysis Features” panel. Step 2 Go to the Solution Controls folder and open the “Solution Method” panel. Select the numerical algorithm to be used (see topic “Transient problems”). Also select the time differencing scheme required. Step 3 Display the “Primary Variables” panel and check that the solution parameter settings are appropriate for your case. If there is any need for alterations, consult Chapter 1, “Transient flow calculations with PISO” or “Transient flow calculations with SIMPLE” in this volume for information and advice. Output controls The output to be produced by a transient run is chosen in a similar manner to that for steady-state problems. However, since the volume of data that can be generated is potentially very large, additional controls are provided to limit the amount to what is absolutely essential. Step 4 Open the “Analysis Output” panel (“Transient problems”). 1. In the “Post tab”, specify control parameters for the wall data that will be written to the solution (.ccm) file and/or printed and saved in the .run file at the end of the run, in the same manner as for steady-state problems. 2. In the “Transient tab”, specify control parameters for data destined for: (a) The transient post data (.pstt) file. The difference between this and the usual solution (.ccm) file is as follows: i) File case.ccm contains analysis results only for the last time step. These form a complete set of all cell data relevant to the current problem and the file can therefore be used to restart the analysis. ii) File case.pstt, on the other hand, contains user-selected data, such as cell pressures, wall heat fluxes, etc. written at predetermined points in time. These are defined by the parameters entered in the “Transient tab”. The file is therefore suitable for post-processing runs but cannot be used to restart the analysis. (b) The data display appearing on your screen at predetermined points in time (not necessarily the same as the ones specified for the post data). This information is also saved in the run history (.run) file. The “Transient tab” control parameters must be used with care since they could cause excessively large data files to be written. On the other hand, they must not be used too sparingly as they may fail to record important data. If the analysis is split into several stages, as is usually the case with large models and/or lengthy Version 4.02

5-5

CONTROL FUNCTIONS

Chapter 5


transients, it is advisable to give the .pstt file produced at the end of each stage a unique file name. This helps to spread the output produced amongst several files and thus eases the data management and manipulation processes. Step 5 Specify any other output controls required, e.g. whether you want to generate monitoring data at every time step, in the same manner as for steady-state problems (see “Analysis Controls for Steady-State Problems”, Step 7). Other controls Step 6 Specify any other necessary controls in the Sources and Other Controls sub-folders, in the same manner as for steady-state problems. Step 7 Go to the “Set Run Time Controls” panel (Analysis Preparation folder) and specify: 1. The analysis run time 2. The method of calculating the time step size and the total number of time steps, see “Transient problems” Step 8 To complete the controls specification, you need to decide whether the analysis is to start from initial conditions or restart from a previous run. Set the appropriate solution controls in the “Analysis (Re)Start” panel. Load-step based solution mode This older procedure allows for all intricacies in the transient problem specification, including variable boundary conditions. However, it is more complex to set up and maintain as it requires definition of so-called ‘load steps’ (see “Load step characteristics” below) and their storage in special transient history files. Other important characteristics are: • • • •

It is driven by its own special user interface, the Advanced Transients dialog, accessed by selecting Modules > Transient in pro-STAR’s main menu bar Time variations may be specified only in terms of load steps, as described in the sections to follow; the use of tables is not permissible It is part of the recommended procedure for setting up moving-mesh cases defined via pro-STAR ‘events’ (see Chapter 12, “Moving Meshes”) It does not support models containing features introduced in STAR-CD V3.20 or later, such as Eulerian multi-phase and liquid films

Load step characteristics For problems involving changing boundary conditions, the main considerations are: •

• 5-6

To define the variation in boundary conditions as a series of events which occur over a period of time. These events, called load steps in pro-STAR terminology, represent a transition from one state of the boundary conditions to another with increasing time. To divide each load step into several time increments, or time steps. Version 4.02

Chapter 5


The allowable variations in the boundary conditions are as follows: 1. Step — where the boundary values change discontinuously from one state to the next, see Figure 5-1(a). 2. Ramp — where the values change linearly between the state at the beginning of the load step to that at the end, see Figure 5-1(b). 3. Function of time — where the variation is arbitrary and is prescribed via a user subroutine. Any combination of load step types can be specified, as shown in Figure 5-1(c)–(d). Boundary condition value

Boundary condition value

S

S

1

2

S

S

3

4

Time

R

R

R

R

R

1

2

3

4 (b)

5

(a)

Time



S

S

R

R

R

1

2

3 (c)

4

5

Figure 5-1

Time

R

R

R

R

S

S

1

2

3

4

5

6

Time

(d)

Representation of boundary value changes by load steps

The difference between the available alternatives is illustrated in Figure 5-2 for load step number n and a time increment of DT. The following information is specified every time a load step is defined: 1. The number of time steps to be performed. 2. The boundary values prevailing at the end of the load step. 3. The manner in which the boundary values should vary between the start and end of the load step. The action of the program is then as follows: (a) For step settings, the value at the start and at all intermediate times is kept equal to value at the end time, as specified in stage 2. above. (b) For ramp settings, the value at the start is made equal to that specified at the previous load step. All intermediate values vary linearly between the start and end values, as shown in Figure 5-2. Version 4.02

5-7

CONTROL FUNCTIONS

Chapter 5


(c) For user settings, values between the start and end times vary in an arbitrary manner, according to what is prescribed in the user-supplied subroutine (see Figure 5-2). Boundary condition value

Load step n–1

A Load step n User coding Ramp B Load step n+1

Step

DT Time

Figure 5-2

Types of change in boundary conditions

Some examples of different load step sequences are shown in Figure 5-1 where the letters S and R denote a step or ramp setting respectively. Load step definition The user should bear in mind the following points when defining load steps: 1. Special considerations apply if the very first load step has a ramp setting. This is because there is no previous load step to fix the value of its starting point. The problem is resolved by defining an extra, dummy load step which merely serves to supply the required boundary value. Examples of this situation are shown in Figure 5-1, cases (b) and (d). 2. At each new load step, the user is free to modify any existing boundary region definition. For example, boundaries that were previously outlets can now become walls and vice versa. However, new boundary regions cannot be added or existing ones deleted, nor can the physical extent of the boundaries be modified in any way. The user must therefore plan the model’s boundary region definitions adequately before starting a transient analysis. A step setting is always imposed at every boundary type change. 3. When the boundary values at the start and end of a load step are identical, the sole purpose of defining the load step would be to permit subdivision of time into discrete time steps so as to track the transient behaviour of the flow field. 4. The time step size can vary from one load step to the next to suit the problem conditions. The size should be small enough to meet the following two 5-8

Version 4.02

Chapter 5


targets: (a) Stability of the numerical solution algorithm, by minimising the cumulative error in the numerical solution. (b) Capture of the transient details of the flow. A good way of testing the sufficiency of the time step size is by calculating the Courant number Co, a dimensionless quantity given by v ∆t Co = -----------l

(5-1)

where v and l are a characteristic velocity and dimension, respectively. Note that in compressible flows v should be replaced by v + c , where c is the velocity of sound. For optimum results, the user should calculate the Courant number in two ways: 1. Cell-wise, by setting v to an estimated local velocity and l to the corresponding local mesh dimension (e.g. cell diagonal). The time step should be chosen such that the maximum Courant number does not exceed 100. 2. Globally, by setting v to the estimated average velocity in the flow field and l to a characteristic overall dimension of the model (e.g. pipe length in pipe flow). The time step should be chosen so that it is commensurate with the time scale of the physical process being modelled. Although precise figures cannot be given for all cases, a Courant number derived from this criterion is typically in the range 100 to 500. The user should inspect the time steps derived in these two ways and select the smallest one for use in the analysis. Solution procedure outline The overall task of setting up parameters for a load-step based transient calculation can be divided into the following steps: Solution controls Step 1 Start up the STAR GUIde system and then define the type of problem you are solving by selecting Transient from the Time Domain pop-up menu in the “Select Analysis Features” panel. Step 2 Go to the Solution Controls folder and open the “Solution Method” panel. Select the numerical algorithm to be used (see topic “Transient problems”). Step 3 Display the “Primary Variables” panel and check that the solution parameter settings are appropriate for your case. If there is any need for alterations, consult Chapter 1, “Transient flow calculations with PISO” or “Transient flow calculations with SIMPLE” in this volume for information and advice. Note that this information is stored for each load step in file case.trns. Therefore, if any changes are needed to these parameters after your load steps have been defined, you will need Version 4.02

5-9

CONTROL FUNCTIONS

Chapter 5


to retrieve the load step information, make the changes and then save the information back in the .trns file (see the description of Step 4 and Step 5 below) Load step controls Step 4 Choose Modules > Transient from the menu bar to activate the Advanced Transients dialog shown below. Select option Advanced Transients On by clicking the action button at the top right-hand side of the dialog. Type the maximum load step number that will be specified in the text box provided and then click Initialize to set up a file (case.trns) for storing all transient history information (i.e. changes in boundary conditions, distribution and length of time steps, etc.). This is a binary file that works very much like the normal pro-STAR problem description (.mdl) file, but is used only in transient problems. The file’s name is entered in the Transient File text box. For a restart run, click the Connect action button to retrieve existing load step information. Note that a number of different files can be utilised in a given run, by first clicking Disconnect to release the current file and then connecting to a new one, as specified in the Transient File text box. pro-STAR’s built-in file browser may be used to locate the required file(s). If necessary, a revised maximum load step number should be typed in the box provided.

Commands: 5-10

TRFILE

LSTEP

LSLIST

LSSAVE Version 4.02

Chapter 5


LSCOMPRESS MVGRID TDSCHEME CDTRANS

LSRANGE CPRINT CPOST

LSGET CPRANGE SCTRANS

LSDELETE WPRINT WPOST

Step 5 If the SIMPLE solution algorithm has been chosen, select a time differencing scheme from the Temporal Discretization pop-up menu. Option Euler Implicit selects the (default) first-order Euler implicit scheme while Three Time Level Implicit selects the second-order three-time-level implicit scheme. The latter gives more accurate solutions but requires more computer time and memory. Your choice of the time differencing scheme should be confirmed by clicking Apply. Step 6 Supply in a sequential manner all information needed to completely define each load step. The current load step should be indicated by highlighting it in the scroll list with the mouse. The required information depends on the time-varying character of the problem and can consist of: 1. Basic parameters of the load step — type these in the text boxes underneath the load step list. The available parameters are: (a) Load step identifying number. (b) Number of time steps. (c) Time increment per time step — if the option button next to this text box is selected, pro-STAR will look for time increment definitions in user subroutine DTSTEP. Any number typed in the text box will be available to the subroutine as a default value. (d) A choice of step or ramp setting for changes in the boundary conditions (note that the ramp setting cannot be chosen if the User option is already selected in step (c) above). (e) Output frequency of print and post-processing data (see “Output controls” below). 2. Redefinition of the boundary type, e.g. changing from wall to outlet boundary conditions and vice versa to simulate the operation of an exhaust valve in a reciprocating engine — see “Boundary Region Definition” on page 4-5. 3. Modification of selected boundary values, without changing the boundary type, as shown in Figure 5-1 — see page 4-7 in the section on “Boundary Region Definition”. 4. Unusual boundary value changes, i.e. other than step-wise or ramp-wise — see option User in the section on “Boundary Region Definition” on page 4-7. The desired variation should be calculated in the appropriate user subroutine (BCDEFI, BCDEFO, BCDEFS, BCDEFP, BCDEFF, BCDEFT, or BCDEFW, see “Boundary condition subroutines” on page 14-5). These routines should supply the required values at every time step and for all boundary regions affected. Any region not covered in this way will take on the usual ramp or step variation specified during the basic load step parameter setting. Remember that in cases where the boundary conditions are to vary linearly from the Version 4.02

5-11

CONTROL FUNCTIONS

Chapter 5


start of the calculation, it is necessary to supply the boundary conditions at the start of the calculations. This is achieved by introducing a dummy first load step with ramp setting. With the exception of the boundary condition, all other data for such a load step are ignored. Output controls The output to be produced by a transient run is chosen in a similar manner to that for steady-state problems. However, since the volume of data that can be generated is potentially very large, additional controls are provided to limit the amount to what is absolutely essential. These controls are implemented in the Advanced Transients dialog and can be sub-divided into a number of basic steps as described below. Note that they are part of the definition for a given load step and can be repeated as necessary during subsequent load steps to achieve the desired fine control over the type and volume of output. Step 7 Decide whether printed output is required. If so, specify: • •

•

•

The printout frequency (in terms of a time step interval) by typing a suitable value in the Print Freq. text box. The cell variables (e.g. velocities, pressure, temperature, etc.) to be printed — click the appropriate Cell Print selection button underneath the desired variable(s). The part of the mesh over which the above quantities will be printed — type a suitable cell range in terms of starting, finishing and increment cell number in the text boxes provided. The wall variables (e.g. shear forces, heat fluxes, etc.) to be printed — click the appropriate Wall Print selection button underneath the desired variable(s).

If some of the cell or wall variables to be printed are additional scalar variables such as chemical species mass fraction, they are specified via the Scalars Select selection button (see Chapter 13, “Multi-component Mixing”, Step 8). Step 8 Decide whether post-processing information is required. If so, specify: • •

•

The output frequency (in terms of a time step interval) by typing a suitable value in the Post Freq. text box. The cell variables (e.g. velocities, pressure, temperature, etc.) to be stored — click the appropriate Cell Post selection button underneath the desired variable(s). The wall variables (e.g. shear forces, heat fluxes, mass fluxes, etc.) to be stored — click the appropriate Wall Post selection button underneath the desired variable(s).

If some of the cell or wall variables to be written are additional scalar variables such as chemical species mass fraction, they are specified via the Scalars Select button (see Chapter 13, “Multi-component Mixing”, Step 8). All the above information is written to a special transient post data (.pstt) file. 5-12

Version 4.02

Chapter 5


The difference between this and the usual solution (.ccm) file is as follows: •

•

File case.ccm contains the calculation results of only the last time step. These form a complete set of all cell data and the file can therefore be used to restart the analysis. File case.pstt, on the other hand, contains user-selected data, such as cell pressures, wall heat fluxes, etc. written at predetermined points in time defined by the parameter typed in the Post Freq. text box. The file is therefore suitable for post-processing runs but cannot be used to restart the analysis.

The Print and Post Freq. parameters above must be used with care since they, together with their associated print and post file operations, may cause excessively large data files to be written. On the other hand, they must not be used too sparingly as they may fail to record important data. If the analysis is split into several stages, as is usually the case with large models and/or lengthy transients, it is advisable to give the .pstt file produced at the end of each stage a unique file name. This helps to spread the output produced amongst several files and thus eases the data management and manipulation processes. Other load step and general solution controls Step 9 Store each completed load step definition in the transient history (.trns) file by clicking on the Save action button. The parameters of the saved definition are displayed in the Load Step scroll list. Step 10 Once all the necessary load steps have been defined, set the total number of load steps to be performed during the next STAR analysis by typing the starting and finishing load step number in the text boxes provided. Confirm by clicking the Apply button. Note that all the above operations have an immediate effect on the transient settings, reflected by immediate changes to what is displayed in the dialog box. However, any subsequent changes made outside this box, e.g by issuing commands via the pro-STAR I/O window, will not be shown. To display these changes, you will need to click the Update button at the bottom of the dialog. Step 11 In addition to the load-step specific information described above, you may also request additional, detailed information that applies to the run as a whole. This includes: •

•

•

Version 4.02

Values of the field variables at a monitoring cell location at each time step. The desired location is specified in the “Monitoring and Reference Data” STAR-GUIde panel. One monitoring cell must be selected for each different material present in the model. Various types of engineering data, as selected from the Monitor Engineering Behaviour panels for specified grid and/or boundary regions. These are also produced at each time step. Input data, boundary conditions and locations, inner iteration statistics, etc. These options are set in the “Monitor Numeric Behaviour” panel. 5-13

CONTROL FUNCTIONS

Chapter 5


Step 12 Specify any other necessary controls in the Sources and Other Controls sub-folders, in the same manner as for steady-state problems. Step 13 The total number of time steps for the run is normally equal to the sum of all time steps in each load step, as defined in Step 6. However, this total may be set independently via command ITER, which may effectively stop the run in the middle of a load step. Step 14 To complete the controls specification, you need to decide whether the analysis is to start from initial conditions or restart from a previous run. Set the appropriate solution controls in the “Analysis (Re)Start” panel. Other transient functions Before initiating a transient run, the user is free to review and modify the existing set of load step definitions. The relevant facilities available in the Advanced Transients dialog are: • • •

Modification — highlight the load step to be changed, type values for the modified parameters and click Save. Deletion — highlight the load step to be deleted and click Delete. Compression of the transient history file — clicking Compress eliminates all deleted steps and renumbers the remaining ones.

Additional points to bear in mind about transient problems are: 1. An analysis can most conveniently be performed in stages, using an initial and several restart runs. When specifying a restart run, you must remember to (a) read in the state of the model as it was when the last run finished, using the “Analysis (Re)Start” STAR-GUIde panel (Standard Restart option) (b) reconnect to the transient history (.trns) file, as described in page 5-10, Step 4 of this section, if additional load steps are to be specified. 2. Along with time-varying boundary values and boundary conditions, you may also elect to vary the geometry of his model, e.g. by moving the mesh in a cylinder-and-piston problem. This can be done by selecting On in the Moving Grid Option pop-up menu at the top of the Advanced Transients dialog. This operation also requires either (a) a user-defined subroutine (NEWXYZ) to calculate the vertex coordinates as a function of time, or (b) the use of special commands provided in the EVENTS module (see Chapter 12, “Moving Meshes”). These permit changes to both vertex locations and cell connectivities. The modified vertex coordinates are also written to the transient post data (.pstt) file and can be loaded and plotted during post-processing.

5-14

Version 4.02

Chapter 5

CONTROL FUNCTIONS Solution Control with Mesh Changes

Solution Control with Mesh Changes The discussion so far in this chapter assumes the usual condition of identical mesh geometry between restart runs. However, it sometimes becomes apparent that changes in mesh geometry applied part-way through the solution process will improve the quality of the final result. For example, inspection of the current solution file may reveal that mesh refinement is needed in some part of the mesh to resolve the flow pattern adequately. Rather than beginning a new analysis from scratch with a new, refined mesh, STAR-CD allows redefinition of the mesh and resumption of the analysis (via a restart run) from the currently available solution. This requires a special mapping operation, called SMAP, that utilises the existing solution data in the .ccm file to create new, approximate solution data that correspond to the re-defined mesh. An example of the result of such an operation is given in Chapter 10, “Solution Mapping” of the Post-Processing User Guide. STAR can read this new file and restart the analysis to obtain a proper solution for the current mesh. Mesh-changing procedure A description of the steps necessary for performing a mesh-changing operation requiring refinement is given below. Note that although restarting with a refined mesh is typical, the same rules apply to any other mesh re-definition, e.g. coarsening, changing cell shapes, or even creating a mesh structure that is physically larger (or smaller) overall than the original configuration. Step 1 Check the directory of your current (coarse-mesh) model to confirm that a pro-STAR model file (say, case-coarse.mdl) and a STAR solution file (say, case-coarse.ccm) exist. Step 2 Start a pro-STAR session and read in the coarse-mesh model from case-coarse.mdl. Then: • • •

Select File > Case Name from the main window, change the case name to, say, case-fine and click Apply Perform whatever mesh refinement operations are necessary (see for example Chapter 3, “Mesh Refinement” in the Meshing User Guide) Select File > Write Geometry File from the main window and save the refined mesh geometry in file case-fine.ccm

Step 3 Signal to STAR that the next run will restart from a different (mapped) solution still to be created: • • • •

Version 4.02

Go to the Analysis Preparation/Running folder in STAR GUIde and open the “Analysis (Re)Start” panel Select option Initial Field Restart from the Restart File Option menu Accept the (default) Restart File name (case-fine.ccm) Select option Restart (Smapped) from the Initial Field Restart menus and click Apply 5-15

CONTROL FUNCTIONS

Chapter 5

Solution Control with Mesh Changes

Step 4 Save all information for the refined mesh, including the restart mode specification above: • •

Select File > Save Model from the main window to save file case-fine.mdl Select File > Write Problem File from the main window to save file case-fine.prob

Step 5 Restore the original coarse-mesh model as follows: • • • •

Select File > Resume From in the main window Input the original mesh by specifying case-coarse.mdl as the model file and click Apply Go to the Post-Processing folder in STAR GUIde and display the “Load Data” panel. Read in the coarse-mesh solution data by specifying case-coarse.ccm as the input file name and then clicking Open Post File

Step 6 Select those coarse-mesh cells that should be used in the mapping process and put them in a cell set (see “Cell set selection facilities” on page 2-46 of the Meshing User Guide). This is because SMAP operates only on cells in the current set. This set may include both fluid and solid cells and will normally contain all cells in the model. The SMAP operation itself is initiated by choosing Utility > Solution Mapping from the main window menu bar to display the Smap/Tsmap dialog shown below:

Commands:

SMAP

TSMAP

The required user input is as follows: 1. Input CCM file — The refined mesh file, case-fine.ccm, created in Step 5-16

Version 4.02

Chapter 5

CONTROL FUNCTIONS Solution-Adapted Mesh Changes

2 above and containing only geometry data at present. If necessary, use pro-STAR’s built-in file browser to locate the file. 2. Output CCM file — The refined mesh solution file, case-fine.ccm. At the end of the mapping operation, this will contain (mapped) solution data as well as geometry data and will therefore be suitable as a restart file for a fine mesh analysis. 3. Instructions on how to assign flow variable values to any fine-grid cells that may lie outside the domain defined by the coarse-grid cells. The available Outside Options are: (a) Default — use default values, as defined in panel “Initialisation” of sub-folder Liquids and Gases in STAR-GUIde (b) Nearest — use values from the nearest cell neighbours (c) Zero — use a value of 0.0 Note that the default mapping algorithm is selected by the Use Smap button. Clicking the Use Tsmap button activates a slightly different algorithm that attempts to enforce global conservation on the fine-grid domain. Other ways in which this option differs from the standard option are as follows: 1. It is not applicable to polyhedral fluid cells 2. Only two Outside Options are available, Nearest and Zero. 3. The volume made up by the fine-grid cells should be fully contained within the volume of the coarse-grid cells. This condition may be satisfied within a tolerance (specified as a volume fraction) entered in the Volume Tolerance box. Step 7 To visualise the outcome of the mapping operation, use the “Load Data”panel in STAR GUIde’s Post-Processing folder. The mapped solution data file just created may be accessed via the “File(s) tab” and field values loaded via the “Data tab”. The data may then be checked by plotting contours but note that only “Cell Data” should be used for this purpose. Step 8 If the mapped results are deemed satisfactory, terminate the pro-STAR session without writing a model file (as this would save the original coarse-grid data) and then run STAR to continue the analysis from the mapped solution. Other noteworthy points are: •

• •

If option Use Tsmap is selected in moving mesh problems containing removed cells (see “Cell-layer Removal/Addition” on page 12-14), the cell set to be mapped should not include removed cells. If any baffles are present in the coarse-grid domain to be refined and mapped, delete the baffles before refinement and redefine them after refinement. Do not change the reference temperature in the restart run.

Solution-Adapted Mesh Changes Section “Solution Control with Mesh Changes” of this chapter shows how to transfer a solution from one mesh to another. In that section, Step 2 simply states Version 4.02

5-17

CONTROL FUNCTIONS

Chapter 5

Solution-Adapted Mesh Changes

that you need to perform whatever mesh refinement operations are necessary. This section aims to show how these changes can be made using the solution from a previous run as a guide. The most frequently used refinement procedures have been assembled in the “Adaptive Refinement” panel of the STAR GUIde system. This caters for mesh refinement based on the results of a previous run. One may employ a refinement operation based on either • •

flow variable gradients, or solution residuals.

Both types of data are stored in the solution (.ccm) file and one may then choose the flow variable and selection method to be employed. A typical refinement session would consist of the following steps: Step 1 Go to the Analysis Preparation/Running folder in the STAR GUIde system and open the “Adaptive Refinement” panel. In the “Refinement Criteria” tab, choose a criterion by selecting the appropriate sub-tab. The flow variable on which to base the refinement depends on the application. For flow-dominated problems, the velocity magnitude or the turbulence kinetic energy have been found to give good results; for chemical reaction- dominated problems, the temperature might be a better choice. Note that: •

Using the Percent of Cells selection method allows you to closely control the number of cells selected for refinement • You may perform multiple selections based on different variables and different criteria; the selection results are accumulated into a compound cell. set • You may abandon your current selection at any stage and start again by clicking the New button Step 2 Go to the “Set Modifications” tab and select set modification options, e.g. •

•

The Near Wall Cell Options may be used to ensure that near-wall cells are left unrefined when limitations on the magnitude of y+ need to be observed. The final set can be ‘grown’, i.e. expanded to include neighbouring cells, to account for inaccuracies in the error estimate and to prevent large differences in refinement level between neighbouring parts of the mesh.

Check the set to be refined visually by plotting it. If necessary, last-minute modifications can be made to this set using the standard pro-STAR cell set utilities (see “Set Manipulation” on page 2-21). Step 3 Once the required cells are finally selected, the “Refine” tab enables you to • • •

5-18

refine them using a simple 2 × 2 × 2 subdivision, recreate the cell connectivity, prepare the resulting new model for the next run. This last step entails mapping the old solution to the new geometry, changing the solution mode to a restart run from the new (mapped) .ccm file and redefining the monitoring Version 4.02

Chapter 5

CONTROL FUNCTIONS Solution-Adapted Mesh Changes

and pressure reference cells, if these were within the area that has been refined. Note, however, that there are many other ways to proceed. Consider filling the volume occupied by the chosen cells with one or more blocks (maybe after a little padding out) and then specifying block factors to build a mesh with progressive, concertina-style refinement. You may also choose to fill the volume with a completely new mesh built by any pro-STAR operation or imported from an external package (see “Importing Data from other Systems” on page 3-1 of the Meshing User Guide). The reverse effect, coarsening the mesh, may by achieved via one of the above methods or by using the CJOIN command.

Version 4.02

5-19

Chapter 6

POROUS MEDIA FLOW Setting Up Porous Media Models

Chapter 6

POROUS MEDIA FLOW The theory behind flow problems of this kind and the manner of implementing it in STAR-CD is given in Chapter 8 of the Methodology volume. The present chapter contains an outline of the process to be followed when setting up a porous media problem and includes cross-references to appropriate parts of the on-line Help system. The latter contains details of the user input required and important points to bear in mind when setting up problems of this kind.

Setting Up Porous Media Models Step 1 Index the cells in the area where distributed resistance exists. This requires use of the cell table (see Chapter 3, “The Cell Table”). As an example, consider the specification of a filter in the pipe shown in Figure 6-1. cell index 1

flow in

cell index 2

cell index 1

flow out

filter

Figure 6-1

Version 4.02

Flow through filter in a pipe

•

For the non-filtered sub-domains (using cell index 1, fluid material property index 12 and porous material index 0) the Cell Table Editor would look as follows:

•

For the filtered sub-domain (using cell index 2, fluid material property index 12 and porous material index 11) the Cell Table Editor would look as follows: 6-1

POROUS MEDIA FLOW

Chapter 6

Setting Up Porous Media Models

The reason for using an identical fluid property index (i.e. 12) is that both sub-domains are part of the same fluid domain. Step 2 Supply property values (resistance coefficients and porosity) for the porous sub-domain using the “Resistance and Porosity Factor” STAR-GUIde panel. If your model contains multiple porous sub-domains possessing different properties, each sub-domain may be selected in turn via the Porous Material # control at the bottom of the panel (see also the “Porosity” Help topic).

Coordinate system 5 x2 (cylindrical) x1 x1 = r 5 x x2 = θ 3 x3 = z

14

z r

θ

z 12 Coordinate system 1 x2 (Cartesian) x1 = x x 1 x1 x2 = y 3 x3 = z

Figure 6-2

y

Honeycombs x

Coordinate system definition in pipe with honeycomb sections

Thus, for the example shown above, the Resistance and Porosity Factor panel 6-2

Version 4.02

Chapter 6

POROUS MEDIA FLOW Setting Up Porous Media Models

settings for the two honeycomb sections should be as follows: First honeycomb section • • • •

Porous material index — 12 Local coordinate system — 1 Flow is along the x- (x1-) direction, hence the value chosen for the resistance coefficients (7) is assigned to Alphax1 and Betax1 The porosity value (0.5) is required only for transient analyses

Second honeycomb section • • • •

Version 4.02

Porous material index — 14 Local coordinate system — 5 Flow is along the θ- (x2-) direction, hence the value chosen for the resistance coefficients (7) is assigned to Alphax2 and Betax2 The porosity value (0.5) is required only for transient analyses

6-3

POROUS MEDIA FLOW

Chapter 6

Useful Points

Step 3 Consider whether, as a consequence of special conditions in your problem, additional input is required for each porous material. Specifically: 1. If turbulence effects are important, specify the relevant parameters using the “Turbulence Properties” STAR-GUIde panel. 2. If there is heat transfer present, specify an effective thermal conductivity and turbulent Prandtl number using the “Thermal Properties” STAR-GUIde panel. 3. If the problem requires calculation of chemical species mass fractions, the effective mass diffusivity and turbulent Schmidt number for each species need to be specified via the “Additional Scalar Properties” STAR-GUIde panel. 4. If you are doing a transient analysis, enter an appropriate value in the Porosity box (see also page 8-2 of the Methodology volume).

Useful Points 1. All porous media properties can be modified by a user subroutine (PORCON, PORDIF, PORKEP, POROS1 or POROS2). 2. α and β should always be positive numbers 3. Excessive values of α and β should be avoided. In cases such as honeycomb structures where cross-flow resistances are much higher than those in the flow direction, the difference in α and β between one direction and the other should be limited to four orders of magnitude. 6-4

Version 4.02

Chapter 6

POROUS MEDIA FLOW Useful Points

4. Avoid setting β = 0 because this can cause K i → 0 as V → 0, leading to a potentially unstable situation. 5. When calculating resistance coefficients from expressions involving pressure drops, remember that the pressure drops are based on unit lengths in each direction. 6. Bear in mind the difference between velocity magnitude V and velocity component u i in your coefficient calculations. 7. Special considerations apply to modelling systems incorporating porous baffles (see “Baffle Boundaries” on page 4-23). Note that baffles may also be used to model a flow resistance at the interface between a fluid and a porous sub-domain, by placing baffles of suitable properties on the faces of the appropriate porous cells. 8. In simulations involving moving meshes, porous media must not be used in areas where there is internal relative mesh motion (cell expansion or contraction). 9. As a result of the particular method used in STAR-CD to calculate pressure gradients at cells on either side of the fluid-porous interface, you need to ensure that porous sub-domains are at least two cell layers thick in any coordinate direction. 10. Tetrahedral meshes should not be used in porous media cases. 11. For examples of porous media flow, refer to the Methodology volume (Chapter 8, “Examples of Resistance Coefficient Calculation”) and to Tutorial 3.1, Tutorial 3.2 and Tutorial 3.3 in the Tutorials volume.

Version 4.02

6-5

Chapter 7

THERMAL AND SOLAR RADIATION Radiation Modelling for Surface Exchanges

Chapter 7

THERMAL AND SOLAR RADIATION The theory behind problems of this kind and the manner of implementing it in STAR-CD is given in Chapter 9 of the Methodology volume. The present chapter contains an outline of the process to be followed when setting up a thermal radiation model and includes cross-references to appropriate parts of the on-line Help system. The latter contains details of the user input required and important points to bear in mind when setting up problems of this kind.

Radiation Modelling for Surface Exchanges Step 1 Open the “Thermal Options” panel in STAR GUIde and select one of the following calculation methods from the Radiation menu: 1. Discrete Transfer - Internal VF Calc, making sure that option Non-Participating is also selected. 2. Discrete Transfer - FASTRAC VF Calc Continue by entering all necessary modelling parameters, as discussed in topic “Thermal Radiation”. Step 2 If present, solar radiation effects can be included by selecting Solar Radiation On and then entering all necessary modelling parameters, as discussed in topic “Solar Radiation”. Note that thermal and solar radiation calculations are independent of each other. A solar-radiation-only analysis may thus be performed without selecting any options from the Radiation menu mentioned in Step 1 above. Solar radiation may enter the solution domain through any open boundary, as well as through transparent walls; see “Solar radiation properties” on page 4-20 for a description of how the latter are specified. Step 3 Inspect the Cell Table Editor entries for cell types assigned to the medium lying between the model’s radiating surfaces and ensure their Radiation option is set to On. Step 4 In the Liquids and Gases folder: 1. Assign thermal properties to the fluid domains via the “Molecular Properties” panel 2. Turn on the temperature solver in the “Thermal Models” panel Step 5 In the “Define Boundary Regions” panel, specify surface radiative properties for all boundaries apart from symmetry and cyclic ones. To do this: 1. If only thermal radiation is modelled: (a) Specify emissivity, reflectivity and transmissivity of all wall, baffle and solid-fluid interface boundaries, as necessary. The description given in “Thermal radiation properties” on page 4-20 (for walls) and on page 4-25 Version 4.02

7-1

THERMAL AND SOLAR RADIATION

Chapter 7

Radiation Modelling for Surface Exchanges

(for baffles) should be read before entering values in this panel. (b) Specify the radiation temperature and emissivity at ‘escape’ surfaces, i.e. boundaries of type “Inlet”, “Outlet”, “Pressure Boundary”, “Stagnation Boundary”, “Free-stream Transmissive Boundary”, “Transient-wave Transmissive Boundary” and “Riemann Boundary”. The required values are entered in the boxes labelled T Radiation and Emissivity. Note that if the FASTRAC method has been chosen, the T Radiation value is not used. Instead, the Surrounding environment temperature specified in the “Thermal Options” panel is used to describe what lies beyond such open boundaries. Note also that the Emissivity value must be set to 0.0 2. For problems involving both thermal and solar radiation, as well as the above parameters, you also need to specify values for the solar reflectivity and transmissivity. These are required at walls, baffles, or solid-fluid interfaces. The description given in “Solar radiation properties” on page 4-20 (for walls) and on page 4-26 (for baffles) should be read before entering such values. 3. For problems involving only solar radiation, the transmissivity of wall boundaries is the only user input required. Step 6 Specify radiation patches unless your problem involves only solar radiation. Tab “Patches” in panel “Create Boundaries” contains most facilities necessary for this task. If you are using the Internal method, you may also create patches via one of the following command-driven options: 1. By specifying the face number that defines the boundary face to be included in the patch — command BDEFINE. 2. By converting a set of shells into a patch — command BSHELL Please also note that: •

Patches generated for use by the Internal method cannot also be used by the FASTRAC method. • The FASTRAC patch specification procedure is different from that for the Internal method. Moreover, the patches are not generated until after the view factor calculation procedure has been initiated (see Step 8 below). • ‘Escape’ surfaces do not need to be patched if the FASTRAC method has been chosen. Step 7 Check the patches created using one of the following methods: 1. Select Patch from the Cell Plot Display Options in the main pro-STAR window 2. Choose Plot > Cell Display > Boundary Patches from the main menu bar 3. Type commands BDISPLAY, PATCH or CDISPLAY, ON, BPATCH in the I/O window. The next cell plot will then display boundaries coloured according to patch number instead of according to boundary type. 7-2

Version 4.02

Chapter 7

THERMAL AND SOLAR RADIATION Radiation Modelling for Participating Media

Step 8 The action here depends on your choice of view factor calculation method: •

•

If you have chosen Discrete Transfer - Internal VF Calc, write the geometry and problem files in the usual way and then run STAR. The view factor and any solar radiation flux calculations are performed at the start of the analysis. In moving mesh cases, view factors are re-calculated at every time step. View factors are saved in a binary file (case.vfs) and are retrieved from that file in a restart run. If you have chosen Discrete Transfer - FASTRAC VF Calc, go to the Analysis Preparation/Running folder, open the “Run Analysis Interactively” panel and start up the external program that calculates the view factors. On completion, the results are stored in file case.nvfs. Subsequent actions are as for the Internal method but using the .nvfs file instead. Note that a re-calculation of the view factors is required if either the solar radiation parameters (Step 2) or boundary transmissivity (Step 5) are altered.

Radiation Modelling for Participating Media This approach is most commonly used to model the radiative effects of a fluid filling the space between radiating solid surfaces. However, STAR-CD is also capable of calculating radiative heat transfer through transparent solid domains, which may then be treated in a similar manner to the intervening fluid. This enables you to make a realistic assessment of, for example, the effect of objects such as windows on the overall heat transfer within an enclosure. The necessary steps for participating media analysis are as follows: Step 1 1. Open the “Thermal Options” panel in STAR GUIde and select one of the following calculation methods from the Radiation menu: (a) Discrete Transfer - Internal VF Calc, making sure that option Participating is also selected. (b) Discrete Ordinates. The participating media radiation option is turned on automatically. 2. Continue by entering all necessary modelling parameters, as explained in topic Thermal Radiation 3. If your problem contains solid domains (including transparent ones) turn on the Solid-Fluid Heat Transfer option Note that inclusion of solar radiation effects is not currently possible for this type of analysis. Step 2 Using the Cell Table Editor: • •

Version 4.02

If transparent solid cells are present, index them to a separate cell type and assign a solid material number to them Select option On from the Radiation menu for all fluid and transparent solid cell types in your model 7-3


Chapter 7

Radiation Modelling for Participating Media

Step 3 Go to the Liquids and Gases folder: 1. Assign thermal properties to the fluid domains via the “Molecular Properties” panel 2. Turn on the temperature solver in the “Thermal Models” panel and click Show Options. In the Participating Media section, specify bulk radiative properties (absorption and scattering coefficients) for the fluid lying between the radiating surfaces. The Conservation and Enthalpy settings in this panel do not affect the radiation solution. Step 4 If transparent solids are present, go to the Solids folder: 1. Assign thermal properties to the solid domains via the “Material Properties” panel 2. Assign radiative properties (absorption and scattering coefficients) to the solid domains via the “Radiative Properties” panel Step 5 In the “Define Boundary Regions” panel, specify surface radiative properties for all boundaries apart from symmetry and cyclic ones. Thus: •

•

Specify emissivity, reflectivity and transmissivity of all wall, baffle and solid-fluid interface boundaries, as necessary. The description given in “Thermal radiation properties” on page 4-20 (for walls) and on page 4-25 (for baffles) should be read before entering values in this panel. Specify the radiation temperature and emissivity at ‘escape’ surfaces, i.e. boundaries by type “Inlet”, “Outlet”, “Pressure Boundary”, “Stagnation Boundary”, “Free-stream Transmissive Boundary” and “Transient-wave Transmissive Boundary”. The required values are entered in the boxes labelled T Radiation and Emissivity.

Note that: •

All boundaries are assumed to be diffuse (i.e. their radiative properties are not dependent on the direction of radiation incident on or leaving the surface). • The absorptivity of the solid-fluid interface (1 - transmissivity - reflectivity) should be consistent with the absorptivity of the solid material defined in Step 4. Step 6 If you have chosen the Discrete Transfer - Internal VF Calc method, create radiation patches for all relevant boundary regions, including external boundaries of solid cells. This process is as described in “Radiation Modelling for Surface Exchanges”, Step 6 and 7. Step 7 Write the geometry and problem files in the usual way and then run STAR. If the initialization stage completes successfully, you will see an echo of the specified modelling parameters in the .info and .run files. • 7-4

If you have chosen the Discrete Transfer - Internal VF Calc method, the Version 4.02

Chapter 7

THERMAL AND SOLAR RADIATION Capabilities and Limitations of the DTRM Method

•

•

view factor calculations are performed at the start of the analysis. In moving mesh cases, view factors are re-calculated at every time step. View factors are also saved in a binary file (case.vfs) and are retrieved from that file in a restart run. Participating media data are stored in another binary file (case.pgr) and then retrieved from it in a restart run. If you have chosen the Discrete Ordinates method, the STAR solver is called every n iterations during the run to solve the radiative transfer equation (where n is the value specified in the “Thermal Options” panel). The solver allocates and frees memory each time, which is reported. In addition, the solver prints out a residual history for the solution of the radiative transfer equation, as well as a summary of the computation. At convergence, the displayed value for the Imbalance quantity should be small compared to heat fluxes of engineering interest. This indicates that the net radiation emission from the medium equals the net absorption into the boundary. If all boundaries are adiabatic and there are no other energy source terms, both the net boundary emission and the net media emission will separately reach very small values.

Capabilities and Limitations of the DTRM Method 1. Lagrangian particle radiation may be modelled by setting Constant 82 to a non-zero value equal to the particle emissivity. For coal combustion cases, this operation may be performed via the “NOx/Radiation” panel in STAR GUIde. 2. Conducting walls (solid-fluid interfaces) should have their transmissivity set to either 1 or 0, depending on whether radiative heat transfer through the solid material is to be considered. If radiation in the solid is on, the transmissivity at the solid-fluid interface must be 1, otherwise it must be 0. 3. The FASTRAC method must be used for thermal/solar radiation problems with transmissive external walls. 4. At present, the FASTRAC method does not apply to problems containing symmetry or cyclic boundary regions. 5. ‘Escape’ or open flow boundaries (inlet, outlet, pressure, etc.) require an assumption regarding the radiation passing through these boundaries and emitted from outside the solution domain. The Internal method assumes that this externally emitted radiation is coming from a surface of given temperature that coincides with the escape boundary surface. The FASTRAC method assumes that a distant ‘environmental’ black body emits radiation at a given temperature. These differing assumptions lead to slightly different results. In addition, the Internal method allows specification of different radiation temperatures at each open boundary whereas FASTRAC assumes that all open boundaries "see" the same environmental surface. 6. Radiation patches cannot be applied to boundaries assigned to the default wall region (region no. 0). If you need to turn on radiation modelling in a problem containing such boundaries, you will need to re-assign them first to a non-zero wall region number. 7. The accuracy of the radiation calculations depends on the patch size since quasi-uniform radiation properties are assumed for a patch. The accuracy of Version 4.02

7-5


Chapter 7

Capabilities and Limitations of the DTRM Method

the view factor calculations depends on both the patch size and the number of beams emitted per patch. For maximum accuracy: (a) Patches should be planar. (b) The aspect ratios of patches should be close to 1.0. (c) If you are using the Internal view factor calculation method, any refinement of patches should be followed by an increase in the number of beams, so that all patches are resolved adequately (see “Patch and beam definition” on page 9-2 of the Methodology volume for a discussion of this point). (d) Patches should not span multiple regions unless the assumption of quasi-uniform radiation properties is valid over those regions. However, acceptable results may be obtained even if one or more of the above conditions are not fully met. 8. If the wrong patch number is assigned to a cell face during the patch definition process (Internal view factor calculations only), the mistake can be rectified either: (a) numerically via the BMODIFY command, or (b) graphically (using the screen cursor) via the BCROSS command. 9. Patches defined as part of the Internal view factor calculation can be stored in a file (case.bnd) and read back from it using the normal boundary export and import facilities provided in panels “Export Boundaries” and “Import Boundaries”, respectively. 10. The default number of beams used in the Internal view factor calculation process (100) may be sufficient for coarse patches. In situations where a patch is created for every boundary cell face, the number of beams may need to be increased (between 1600 and 2500 for typical radiation problems) in order to resolve adequately the patches present in the system. The FASTRAC calculation method uses a fixed number of beams (1,024). 11. The CPU time for Internal view factor calculations increases in proportion to the number of patches multiplied by the number of beams. The CPU time for radiation heat transfer calculations increases in proportion to the number of patches. The FASTRAC view factor calculations are also dependent on the number of patches but the CPU time required is considerably reduced. 12. User subroutine USOLAR cannot be used for solar radiation problems employing the FASTRAC method. 13. STAR-HPC runs are not feasible for problems involving participating media radiation. 14. STAR-HPC runs for problems involving solar radiation are only feasible if the FASTRAC method has been chosen. 15. Surface-exchange problems using the Internal calculation method can be run in STAR-HPC mode, but the view factors have to be calculated in ‘single-processor’ mode. To do this, run the case for zero iterations on a single processor and save the view factor (.vfs) file. 16. As stated on page 7-3, Step 8 above, Internal view factors for moving mesh cases are re-calculated at every time step. Therefore, in view of the previous restriction, STAR-HPC runs for problems involving both radiation and a 7-6

Version 4.02

Chapter 7

THERMAL AND SOLAR RADIATION Capabilities and Limitations of the DORM Method

moving mesh are not feasible. Note also that FASTRAC view factor calculations cannot be used in moving mesh cases at present. 17. Note that command PATCH generates only shell surfaces. It cannot be used to create radiation patches.

Capabilities and Limitations of the DORM Method 1. The discrete ordinates model (DORM) does not need radiation patches and the computational overheads involved in their use. In addition, to facilitate switching from a discrete transfer (DTRM) to a discrete ordinate (DORM) model for the same problem, STAR will accept geometry files with or without patches. 2. Nevertheless, using DORM can still add significantly to the CPU time and memory needed for a given simulation. For this reason, users are encouraged to plan their analyses conservatively until they gain experience with the CPU time and memory requirements of their model. The run-time output for the DORM calculation will echo the memory requirements (see Step 7 above). 3. The memory requirements of the calculation depend on your choice of angular discretization. The table below gives a guide to memory usage per 100,000 cells. Note that this holds for single-precision calculations and a grey medium. .

Table 7-1: Approximate memory required for DORM analysis Ordinates

Angular discretization

Additional memory per 100,000 cells

8

S2

45 MB

24

S4

55 MB

48

S6

75 MB

80

S8

95 MB

4. The model may be run in the normal way under STAR-HPC. However, the solution history for a serial run will be different from that for a parallel run. Although the radiative transfer equation is similar to a normal transport equation, there is no equivalent of the diffusion term and so the equation is not elliptic. To solve this equation efficiently, a specialized solver that follows the directions of each ordinate is used. Thus, in the STAR-HPC environment, some domains may receive the information about certain directions only after it has crossed through the other domains. Nevertheless, converged solutions in serial and HPC calculations are identical. 5. Coupling between the ordinate directions at cyclic and symmetry boundaries approximates such boundaries as diffuse. 6. DORM is fully compatible with all cell shapes supported by pro-STAR (polyhedral cells, baffle cells, etc.) 7. DORM can also be used to model surface-exchange problems (i.e. non-participating media analyses). Since the participating media mode is Version 4.02

7-7


Chapter 7

Radiation Sub-domains

always on, the absorption and scattering coefficients must be set to zero for such cases. 8. At present, DORM does not support cases involving solar radiation.

Radiation Sub-domains In some problems, radiation effects are important only within a restricted sub-domain of the overall solution domain, e.g. when doing a complete continuum mechanics analysis around a car body, where radiation calculations are only necessary under the car bonnet. Under such circumstances, it is possible to confine the radiative heat transfer treatment to the part of the model where it is relevant, thus avoiding the lengthy calculations needed for a full radiation analysis. The following steps are then necessary: Step 1 In the “Thermal Options” STAR GUIde panel, turn on the radiation calculations by selecting either the Discrete Transfer - Internal VF Calc or the Discrete Ordinates option in the Radiation menu and then specify all necessary radiation parameters. Step 2 Using the Cell Table Editor: •

Create a separate cell type for all cells occupying the sub-domain that is subject to the radiative treatment • For this cell type only, turn the Radiation option On Step 3 The action here depends on your choice of method in Step 1 above: •

For the Discrete Transfer - Internal VF Calc method (whether Participating or Non-Participating), create the necessary number of special ‘Radiation’ boundaries so as to completely separate the radiative from the non-radiative part of the domain. • For the Discrete Ordinates method, radiation boundaries are not applied. Step 2 above is all that is required to define the problem properly; no further action is necessary. Step 4 Within the radiative sub-domain, use the “Define Boundary Regions” panel to specify radiation properties for all boundary regions, including the special boundaries created above (see also Chapter 4, “Radiation Boundaries”). Step 5 If you have chosen the Discrete Transfer - Internal VF Calc method, create patches on all boundaries surrounding the radiative sub-domain, including the radiation boundaries, as described in “Radiation Modelling for Surface Exchanges”. Step 6 Write the geometry and problem files in the usual way and then run STAR.

7-8

Version 4.02

Chapter 8

CHEMICAL REACTION AND COMBUSTION Introduction

Chapter 8

CHEMICAL REACTION AND COMBUSTION

Introduction STAR-CD allows for two kinds of chemical reaction: • •

Homogeneous — the reaction occurs within the bulk of the fluid Heterogeneous — the reaction takes place only at surfaces, such as in catalytic converters

Heterogeneous reactions are currently implemented via user-supplied subroutines. Homogeneous reactions are grouped into three distinct types: 1. Unpremixed/Diffusion — reactions of this type occur when the fuel and oxidant streams enter the solution domain separately, as in a Diesel engine. The reactions may be sub-divided into the following groups: (a) Local Source — these include eddy break-up, chemical kinetic, and hybrid models (see “Local Source Models” on page 8-2 for more details) (b) Complex Chemistry — these model the reaction system by including the full reaction mechanism (see “Complex Chemistry Models” on page 8-11 for more details). (c) Presumed Probability Density Function (PPDF) — these include single and multiple fuel implementations and the Laminar Flamelet model (see “Presumed Probability Density Function (PPDF) Models” on page 8-3 for more details) 2. Partially Premixed — combustion of this type is one of the essential features in Gasoline Direct Injection engines, where combustion occurs in a non-uniform mixture. The reactions may be sub-divided into the following groups: (a) Local Source, of the type mentioned above (b) Complex Chemistry, of the type mentioned above (c) Regress Variable, represented by a Flame Area Evolution (FAE) model 3. Premixed — reactions of this type occur when the fluid initially has a uniform composition, as in a spark ignition engine (a) Local Source, of the type mentioned above (b) Complex Chemistry, of the type mentioned above (c) Regress Variable, represented by various eddy break-up and flame-area models (see “Regress Variable Models” on page 8-10 for more details) The theory behind reaction models of the local source, complex chemistry and PPDF type is described in Chapter 10 of the Methodology volume. Regress variable models are normally used in engine combustion simulation and are described separately in Methodology Chapter 11. A set of recently implemented engine combustion models is discussed in the section on “Setting Up Advanced I.C. Engine Models” on page 8-22 of this chapter. In some cases, the model describing the main chemical reaction(s) may need to be supplemented by subsidiary models that describe: Version 4.02

8-1


Chapter 8

Local Source Models

• •

•

Ignition mechanisms Emission of pollutants, typically NOx products. Special considerations apply to modelling NOx-type reactions, as discussed in “NOx Modelling” on page 8-39. Application Specific models, such as engine knock

All these models together constitute a so-called chemical reaction scheme. Note that: • •

•

•

Chemical schemes are defined and numbered individually Chemical scheme definitions can exist independently of any fluid domains or scalar variables. However, they need to be explicitly assigned to a domain before they can be used in your simulation. Each fluid domain may be associated with only one chemical reaction scheme. However, this association may be changed by the user to suit problem requirements or to try out alternative reaction models. Special considerations apply to modelling coal combustion; these are discussed in the section on “Coal Combustion Modelling” on page 8-41.

Local Source Models The main characteristics of this group of models are as follows: 1. Up to 30 chemical reactions may be defined per scheme 2. The reactions are irreversible 3. Each reaction is associated with a single chemical species designated as the leading reactant (equivalent to fuel in a combustion reaction). This species characterises the reaction and is consumed by it. The remaining reacting species are defined as reactants. 4. The products of a reaction are defined as products. However, (a) if a product of a reaction participates as a reactant in a second reaction, it should be specified as a leading reactant or ordinary reactant, as appropriate; (b) if a product is transported into the solution domain from an external source, it also should be specified as a reactant. 5. The distribution of products within the solution domain can be calculated algebraically, provided that the products are generated only within the domain. 6. If all incoming streams consist of identical fuel-to-reactant ratios (in transient cases the initial fields must also have the same ratio), the reaction process is termed premixed (see “Premixed reaction/homogeneous systems” on page 10-4 of the Methodology volume). If this is not the case, the process is either of the diffusion or the partially premixed type and the user needs to solve an additional scalar transport equation for the mixture fraction (total mass fraction of burned and unburnt fuel, see “Diffusion reaction / non-homogeneous systems” on page 10-5). 7. STAR-CD automatically sets up mixture fraction scalars for each leading reactant in diffusion and partially premixed reactions. However, it is the user’s responsibility to ensure that boundary conditions for both leading reactant and 8-2

Version 4.02

Chapter 8

CHEMICAL REACTION AND COMBUSTION Presumed Probability Density Function (PPDF) Models

mixture fraction are specified correctly and that they are the same for both of them. 8. The reactions themselves are defined by specifying the amounts (in kilomoles) of the participating leading reactants, reactants and products. For example, the input required for the following reaction (combustion of methane) CH 4 + 2O 2 → CO 2 + 2H 2 O

(8-1)

is Reaction (1) Leading reactant (fuel) (1) Reactant (1) Product (1) Product (2)

— — — —

CH 4 O2 CO 2 H2 O

kmol 1 2 1 2

9. pro-STAR includes facilities for checking that mass is conserved for each reaction.

Presumed Probability Density Function (PPDF) Models Models of this type are described in Chapter 10, “Presumed-PDF (PPDF) Model for Unpremixed Turbulent Reaction” in the Methodology volume. These fall into two main groups: •

•

Single-fuel PPDF, where only one type of fuel and one type of oxidiser are present, though each of these may enter the combustion system through more than one inlet. Multiple-fuel PPDF, where two types of fuel and one type of oxidiser are present.

Single-fuel PPDF The basic equations solved are for the mean mixture fraction f and its variance g f (see Chapter 10, “Single-fuel PPDF” in the Methodology volume). There is a choice between equilibrium chemistry models (these assume a local instantaneous chemical equilibrium) and a laminar flamelet model that allows for non-equilibrium effects (such as flame stretch). Equilibrium models In these models, the PDF integration may be performed in two ways: 1. By employing a numerical integration technique 2. By expressing all instantaneous values of the variables as polynomials of the mixture fraction and then doing the integration analytically. Polynomial coefficients may be (a) supplied by the user (b) read in from a built-in database stored in file ppdf.dbs (c) calculated by the CEA (Chemical Equilibrium with Applications) program [5, 6]. This is an auxiliary program that computes the chemical Version 4.02

8-3


Chapter 8

Presumed Probability Density Function (PPDF) Models

equilibrium composition of a mixture. This program is included in the STAR-CD suite and is used in conjunction with the built-in PPDF model. There is also a choice between adiabatic and non-adiabatic PPDF: 1. For adiabatic PPDF: (a) The mixture density and temperature are calculated numerically or from polynomials in f. Note that these polynomials are based on molar fractions. (b) Since temperature is calculated independently, the ‘Constant’ specific heat property option with default values may be used 2. For non-adiabatic PPDF, the density is calculated from the ideal gas law and the temperature from the enthalpy transport equation. The mass fractions of all other chemical species related to the reaction are defined as additional scalar variables and calculated numerically or from the user-supplied polynomials in f, as above. Up to forty eight such species can be specified by the user. Laminar flamelet model In this model, the PDF integration is always performed numerically and the results stored in a look-up table which is characterised by its mean mixture fraction, mixture fraction variance and strain rate. There is also a choice between an adiabatic and a non-adiabatic model, as above. The setup procedure for the model is described in the on-line Help for the “Reaction System” STAR GUIde panel. One part of this procedure is to specify the reaction mechanism, stored in a reaction definition file in CHEMKIN format. This is organized in three sections: • • •

Element data Species data Reaction data

The basic data are often supplemented by auxiliary data for special reactions such as third-body reactions. Element Data All chemical species in the reaction mechanism must be composed of chemical elements or isotopes of chemical elements. Each element and isotope must be declared using a one- or two-character symbol. The purpose of the element data is to associate the element atomic weights with their character symbol representations. If an ionic species is used in the reaction mechanism (e.g, OH+), an electron must be declared as the element E. Element data must start with the word ELEMENTS (or ELEM) but, following that, there are minimal restrictions on the formatting of the rest of the section. Any number of element symbols can be written on any number of lines. The symbols may appear anywhere on a line, but those on the same line must be separated by blanks. Any line or portion of a line starting with an exclamation mark (!) is considered a comment and will be ignored. Blank lines are also ignored. If an element is in the list below, then only the symbol identifying it need appear 8-4

Version 4.02

Chapter 8


in the element data. The recognized elements are as follows: H, HE, LI, BE, B, C, N, O, F, NE, NA, MG, AL, SI, P, S, CL, AR, K, CA, SC, TI, V, CR, MN, FE, CO, NI, CU, ZN, GA, GE, AS, SE, BR, KR, RB, SR, Y, ZR, NB, MO, TC, RU, RH, PD, AG, CD, IN, SN, SB, TE, I, XE, CS, BA, LA, CE, PR, ND, PM, SM, EU, GD, TB, DY, HO, ER, TM, YB, LU, HF, TA, W, RE, OS, IR, PT, AU, HG, TL, PB, BI, PO, AT, RN, FR, RA, AC, TH, PA, U, NP, PU, AM, CM, BK, CF, ES, FM, D, E For an isotope, the atomic weight must follow the identifying symbol and be delimited by slashes (/). The atomic weight may be given in integer, floating-point, or “E” format, but internally it will be converted to a floating-point number. For example, the isotope deuterium may be defined as D/2.014/. If desired, the atomic weight of an element in the above list may be altered by including the atomic weight as input just as though the element were an isotope. An acceptable format for element data specification is shown below: ELEMENTS H D /2.014/ O N END! END is optional

Species Data Each chemical species in a reaction must be identified on one or more species line(s). Any set of up to 16 upper or lower case characters can be used, as for species names, which are case sensitive. In addition, each species must be composed of elements that have been identified in the element data section. Species data must start with the word SPECIES (or SPEC) but, as already discussed, subsequent formatting of this section is not particularly important. An acceptable format for species data specification is shown below: SPEC H2 O2 H O OH HO2 H2O

Reaction Data The reaction mechanism may consist of any number of chemical reactions involving the species named in the species section. A reaction may • • •

be reversible or irreversible; be a three-body reaction with an arbitrary third body and/or enhanced third-body efficiencies; have one of several pressure-dependent formulations.

The rate of each reaction is defined by specifying A R , β R and E R from the general Arrhenius rate equation for the forward reaction, see equation (10-65) in the Methodology volume. Reaction data must start with the word REACTIONS (or REAC). On the same line, you may specify units of the activation energies to follow by including the word CAL/MOLE, KCAL/MOLE, JOULES/MOLE, KJOULES/MOLE, KELVINS, or EVOLTS. The default units for E R are cal/mole and the default units for A R are cm, mole, sec and K. Including the word MOLECULES on the REACTIONS line changes the units of A R to cm, molecules, sec and K. The lines following the REACTIONS line contain reaction definitions together with their Arrhenius rate coefficients, as described in Table 8-1. The description is Version 4.02

8-5


Chapter 8


composed of reaction data and optional auxiliary information data. Table 8-1: Species Symbols Each species in a reaction is described by a unique sequence of characters, as they appear in the species data (e.g. CH4). Coefficients A species symbol may be preceded by an integer or real coefficient. The coefficient’s meaning is that there are that many moles of the particular species present as either reactants or products; e.g. 2OH, is equivalent to OH + OH. Non-integer coefficients are allowed, but the element balance in the reaction must still be maintained. Delimiters +

A plus sign is the delimiter between each reactant species and each product species.

=

An equality sign is the delimiter between the last reactant and the first product in a reversible reaction.

=>

An equality sign with an angle bracket on the right is the delimiter between the last reactant and the first product in an irreversible reaction.

Special Symbols

8-6

+M

An M as a reactant and/or product stands for an arbitrary third body. It should appear as both a reactant and a product. In a reaction containing an M, certain species can be specified as having enhanced third-body efficiencies; in which case auxiliary data (described below) must follow the reaction line. If no enhanced third-body efficiencies are specified, all species act equally as third bodies and the effective concentration of the third body is the total concentration of the mixture.

(+M)

An M as a reactant and product surrounded by parentheses indicates that the reaction is pressure-dependent, in which case auxiliary information line(s) (described below) must follow the reaction to identify the fall-off formulation and parameters. A species may also be enclosed in parentheses. For example, (+H2O) indicates that water is acting as the third body in the fall-off region, not the total concentration M.

!

An exclamation mark means that all following characters on the reaction line are comments. For example, the comment may be used to give a reference to the source of the reaction and rate data. Version 4.02

Chapter 8


The second field of each reaction line is used to define the Arrhenius rate coefficients A R , β R and E R , in that order. At least one blank space must separate the first number and the last symbol in the reaction. The three numbers must be separated by at least one blank space and be given in integer, floating point, or “E” format (e.g., 123, 123.0 or 123E1). Their units are as specified in the REACTIONS line above. An example of reaction data for a simple mechanism is shown below: REACTIONS CAL/MOLE H2 + O2 = OH 1.70E+13 0.000 47780 OH + H2 = H2O + H 1.17E+9 1.300 3626 H + O2 = HO2 3.61E+17 -0.720 0.000 O + H2 = OH + H 5.06E+04 2.670 6290 HO2 + H2 = H2O2 + H 1.25E+13 0.000 0 0.5H2O2 + 0.5H2 = H2O 1.60E+12 0.000 3800 ! example of real coefficients END ! END statement is optional;

The basic rules for specifying reaction data are summarised below: 1. The first line must start with the word REACTIONS (or REAC), and may include units definition(s). 2. The reaction description can begin anywhere on the line. All blank spaces, except those between Arrhenius coefficients, are ignored. 3. Each reaction description must have = or => between the last reactant and the first product. 4. Each reaction description must be contained within one line. 5. Three Arrhenius coefficients ( A R , β R and E R ) must appear in order on each line, separated from each other and from the reaction description by at least one blank space; no blanks are allowed within the numbers themselves. 6. No more than six reactants or six products are allowed in a reaction. 7. Comments are any characters following an exclamation mark. Auxiliary Reaction Data The format of an auxiliary information line is a character-string keyword followed by a slash-delimited (/) field containing an appropriate number of parameters (in either integer, floating point, or “E” format). Different types of auxiliary reaction data are described below, followed by an example: 1. Third-Body and Pressure-Dependent Reaction Parameters If a reaction contains M as a reactant and/or product, auxiliary information lines may follow the reaction line to specify enhanced third-body efficiencies of certain species. The keyword defining an enhanced third-body efficiency is the species name of the third body, and its single parameter is its enhanced efficiency factor. A species that acts as an enhanced third body must be declared as a species. If a pressure-dependent reaction is indicated by a (+M) or by a species contained within parentheses, say (+H2O), one or more auxiliary information lines must follow to define the pressure-dependence parameters. For all pressure-dependent reactions, an auxiliary information line must follow to specify either the low-pressure limit Arrhenius parameters (for fall-off reactions) or the high-pressure limit Arrhenius parameters (for chemically Version 4.02

8-7


Chapter 8


activated reactions). For fall-off reactions, the keyword LOW must appear on the line, with three rate parameters A 0 , β 0 and E 0 . For chemically activated bimolecular reactions, the keyword HIGH must appear on the line, with the three rate parameters A ∞ , β ∞ and E ∞ .There are then three possible interpretations of the pressure-dependent reaction: (a) Lindemann formulation - no additional parameters are defined (b) Troe formulation - in addition to the LOW or HIGH parameters, the keyword TROE followed by three or four parameters must be included in the following order: a, b, c and d, as defined in equation (10-77) of the Methodology volume. The fourth parameter is optional and if omitted, the last term in equation (10-77) is not used. (c) To define an SRI pressure-dependent reaction, in addition to the LOW or HIGH parameters, the keyword SRI followed by three or five parameters must be included in the following order: a, b, c, d and e, as defined in equation (10-79) of the Methodology volume. The fourth and fifth parameters are optional. If only the first three are stated, then by default d = 1 and e = 0 . 2. Landau-Teller Reactions To specify Landau-Teller parameters, the keyword LT must be followed by two parameters — the coefficients B R and C R from equation (10-81) in the Methodology volume. The Arrhenius parameters A R , β R , and E R are taken from the numbers specified on the reaction line itself. If reverse parameters are specified in a Landau-Teller reaction via REV (see item 4 below), the reverse Landau-Teller parameters must also be defined, with the keyword RLT and two coefficients B R and C R for the reverse rate. 3. Logarithmic Interpolation of Pressure-Dependent Rates This generalized way of describing the pressure dependence of a reaction rate is indicated by the PLOG keyword in auxiliary lines. In this case, the reaction description should not include (+M) in it, although this is used to indicate that the reaction is pressure dependent in other cases. This particular option for describing pressure-dependent reactions cannot be combined in any given reaction with other options for describing pressure dependence. One supplementary line starting with the PLOG keyword needs to be supplied for each pressure in the set. The keyword is followed by slash-delimited values for the pressure (in atmospheres) and the rate parameters for that pressure. The supplementary lines need to be in order of increasing pressure. If the rate expression at a given pressure cannot be described by a single set of Arrhenius parameters, more than one set may be provided. Each of these should be followed by the keyword DUPLICATE, meaning the sum of the sets of rates provided for a given pressure will be used. The units of the rate parameters provided with the PLOG keyword should match the units used for the overall reaction description. Note that, in this case, although rate parameters need to be supplied on the main reaction line to prevent an error, those values are superseded by the ones provided on the supplementary lines. 4. Reverse Rate Parameters For a reversible reaction, auxiliary data may follow the reaction to specify Arrhenius parameters for the reverse-rate expression. Here, the three 8-8

Version 4.02

Chapter 8


Arrhenius parameters ( A R , β R , and E R ) for the reverse rate must follow the keyword REV. This option overrides the reverse rates that would be normally computed by satisfying microscopic reversibility through the equilibrium constant, as described by equation (10-66) in the Methodology volume. 5. Reaction Order Parameters Auxiliary data may be included to override the reaction order for a species, using the auxiliary keywords FORD or RORD, for forward and reverse reaction descriptions, respectively. Each occurrence of these keywords must be followed by the species name and the new reaction order. This option overrides the stoichiometric coefficients for the species included in the auxiliary data. 6. Reaction Units It is sometimes convenient to specify units for a particular reaction rate fit that differ from the default units specified for other reaction expressions in the chemistry mechanism. In such a case, you should employ the auxiliary keyword UNITS. This keyword must be followed by one or more of the following unit descriptors: MOLECULE, CAL, KCAL, JOULE, KJOULE, KELVIN, or EVOLTS. The inclusion of MOLECULE would indicate that the reaction rate expression is in units of molecules/cm3 rather than mole/cm3. The remaining unit descriptors specify the energy units in the rate expression. Note that the temperature units in the rate expression are always in Kelvin. An example of the use of auxiliary reaction data for a three-parameter Troe fall-off reaction with enhanced third-body efficiencies is shown below: CH3+CH3(+M)=C2H6(+M) 9.03E16 -1.18 654 LOW / 3.18E41 -7.03 2762 / TROE / 0.6041 6927 132. / H2/2/ CO/2/ CO2/3/ H2O/5/

Multiple-fuel PPDF 1. Four equations are solved, for the progress variables f p (primary fuel mixture fraction), f s (secondary fuel mixture fraction), g f (primary fuel variance) and g ξ (variance of variable ξ, see Chapter 10, “Multiple-fuel PPDF” in the Methodology volume). 2. Only an equilibrium chemistry model is available in this case 3. The PDF integration is always performed numerically Other noteworthy points about PPDF models are: 1. In order to increase the efficiency of combustion systems by increasing the temperature of incoming oxidisers, the use of vitiated air containing combustion products is a viable option. The basic PPDF model, which assumes that only fuel and air enter the system, cannot be used for this kind of problem. However, STAR-CD’s implementation has been extended to allow up to four dilutants to enter the combustion system. The basic setup is the same as that used for the standard PPDF model. However, additional transported scalars are defined to represent the dilutants; therefore, additional boundary conditions need to be defined for them. Version 4.02

8-9


Chapter 8

Regress Variable Models

It is emphasised that PPDF with dilutants can only be used in conjunction with the single-fuel, equilibrium chemistry model plus the non-adiabatic PPDF option. 2. The multiple-fuel PPDF option may also be used to model a system containing only one type of fuel but two different types of oxidiser.

Regress Variable Models Models in this group solve a transport equation for a regress variable representing the combustion process and are described in the Methodology volume, Chapter 11. Their main features are: 1. The regress variable b defined by equation (11-4) in the Methodology volume is the transported variable and is a passive scalar 2. All physical scalar variables participating in such schemes are linearly related to b 3. Regress variable models may be classified into two groups: (a) Flame-area models, discussed in sections “Premixed Combustion in Spark Ignition Engines” and “Partially Premixed Combustion in Spark Ignition Engines” of the Methodology volume: i) “The Weller flame area model” — makes use of the wrinkling factor Ξ, which is either obtained from an algebraic relationship given by equation (11-36) or from the solution of a transport equation ii) “The CFM-ITNFS model” — employs a transport equation for the flame area density Σ, given by equation (11-10) iii) “The Weller 3-equation model” — requires the solution of equations for both wrinkling factor and mixture fraction iv) All have their own ignition models (b) Eddy break-up models, used in a manner similar to that described above under “Local Source Models”. 4. The one-step reaction representing the combustion process is associated with a single chemical species designated as the leading reactant (or fuel). This species characterises the reaction and is consumed by it. The remaining reacting species are defined as reactants. 5. The reaction is irreversible and is defined by specifying the amounts (in kilomoles) of the participating leading reactants, reactants and products. 6. pro-STAR includes facilities for checking that mass is conserved. 7. If all incoming streams consist of identical fuel-to-reactant ratios (in transient cases the initial fields must also have the same ratio), the reaction process is termed premixed (see “Premixed reaction/homogeneous systems” on page 10-4 of the Methodology volume). If this is not the case, i.e. the process is of the partially premixed type, an additional scalar transport equation for the mixture fraction needs to be solved. The only regress variable model that may be used in partially premixed systems is the Weller 3-equation model. 8. STAR-CD automatically sets up mixture fraction scalars for each leading reactant in diffusion and partially premixed reactions. However, it is the user’s 8-10

Version 4.02

Chapter 8

CHEMICAL REACTION AND COMBUSTION Complex Chemistry Models

responsibility to ensure that boundary conditions for both leading reactant and mixture fraction are specified correctly and that they are the same for both of them. 9. Exhaust gases present in EGR (Exhaust Gas Recirculation) systems are taken into account by defining active scalars for each exhaust gas species and solving additional transport equations for their mass fraction (see also Chapter 11, “Exhaust Gas Recirculation” in the Methodology volume).

Complex Chemistry Models The complex chemistry model supports two types of format for reaction mechanism definition. One of them is the CHEMKIN format, the other the STAR-CD native format described below. In order to use STAR-CD’s complex chemistry model, a reaction mechanism file called cplx.inp&& has to be created by the user for each chemical scheme in which a complex chemistry model is applied. The characters ‘&&’ at the end of the file name represent the chemical scheme number in which the complex chemistry model is applied. For example, if such a model is applied in chemical scheme no. 2, the reaction mechanism file should be called cplx.inp02. STAR will write an echo file cplx.inp&&-echo for each cplx.inp&& file it has read, so that users can ensure settings have been correctly applied. File cplx.inp&& contains the reaction formula, chemical kinetic data and keywords and extra parameters for special reactions, as outlined below: Reaction formula definition The general form of a reaction formula is given by n 1 ′R 1 m 1 ′ + n 2 ′R 2 m 2 ′ + … = n 1 ″P 1 m 1 ″ + n 2 ″P 2 m 2 ″ + … A β E Here, n 1 ′ , n 2 ′ , …, n 1 ″ , n 2 ″ , … are the stoichiometric coefficients which could be integer or real numbers, R 1 , R 2 , …, P 1 , P 2 , … are species names, m 1 ′, m 2 ′, …, m 1 ″ , m 2 ″ , … are the mass fraction exponentials, A is the pre-exponential factor (in units of cm-mole-sec-K), β the temperature exponent and E the activation energy of the Arrhenius rate constant (in cal/mol). If the mass fraction exponentials are equal to 1, they are not written into the corresponding echo file (cplx.inp&&-echo). Rules: • •

• • • • • Version 4.02

There are no spaces between stoichiometric coefficients n i ′ , n i ″ and species names. If n i ′ or n i ″ are equal to 1, they can be omitted. m i ′ and m i ″ must be separated by at least one space from the species name. If the value of m i ′ or m i ″ is not specified, it will be assumed that m i ′ = n i ′ or m i ″ = n i ″ . Character ‘=’ is used for reversible reactions; ‘⇒’ for irreversible reactions. There is no ‘+’ character between the pre-exponential factor, A, and the nearest species name. A, β and E are separated by at least one blank space. Everything following the ‘!’ character is treated as a comment The ‘+’ character should not be used in a real number expression. For example, 1.2E+05 should be written as 1.2E05. The maximum number of reactants or products in a single reaction must not exceed 5 8-11


Chapter 8

Complex Chemistry Models

Three-body reaction definition To define a three-body reaction, add a line starting with the keyword M after the reaction formula, i.e. /M/ A/α 1 / B/α 2 / … Rules: • •

Keyword M must be enclosed by two ‘/’ characters and is not case sensitive A, B, … are species names and α 1 , α 2 , … are the corresponding efficiency factors. They are separated by at least one blank space.

The Landau-Teller reaction To define a Landau-Teller reaction, add a line starting with the keyword RLT after the normal reaction formula, i.e. /RLT / B C Rules: • • •

Keyword RLT must be enclosed by two ‘/’ characters and is not case sensitive B and C are the Landau-Teller parameters and are separated by at least one blank space If the reaction is a three-body reaction as well, a new line is added starting with ‘ /M/ ’ and the third body efficiency factors

The Lindemann fall-off reaction To define a Lindemann fall-off reaction, add a line starting with the keyword LOW after the reaction formula, i.e. /LOW / A L β L E L Rules: • •

• •

Keyword LOW must be enclosed by two ‘/’ characters and is not case sensitive A L , β L , and E L are the pre-exponential factor, temperature exponent and activation energy, respectively, of the low pressure limit and are separated each from each other by at least one blank space The corresponding values for the high pressure limit are assumed to be those given above as part of the reaction formula definition If the reaction is a three-body reaction as well, a new line is added starting with ‘ /M/ ’ and the third body efficiency factors

The Troe fall-off reaction To define a Troe fall-off reaction, add two lines starting with keywords LOW and TROE, respectively, after the reaction formula, i.e. /LOW / A L β L E L 8-12

Version 4.02

Chapter 8

CHEMICAL REACTION AND COMBUSTION Complex Chemistry Models

/TROE/ a b c d Rules: • •

• • •

The definition of keyword LOW is the same as above The pre-exponential factor, temperature exponent and activation energy values for the high pressure limit are assumed to be those given above as part of the reaction formula definition Keyword TROE must be enclosed by two ‘/’ characters and is not case sensitive a, b, c and d are the corresponding Troe parameters (d is optional) If the reaction is a three-body reaction as well, a new line is added starting with ‘ /M/ ’ and the third body efficiency factors.

The SRI fall-off reaction To define a SRI fall-off reaction, add two lines starting with the keywords LOW and SRI, respectively, after the reaction formula, i.e. /LOW / A L β L E L /SRI / a b c d e Rules: • •

• • •

The definition of keyword LOW is the same as above The pre-exponential factor, temperature exponent and activation energy values for the high pressure limit are assumed to be those given above as part of the reaction formula definition Keyword SRI must be enclosed by two ‘/’ characters and is not case sensitive a, b, c, d and e are the corresponding SRI parameters and are separated from each other by at least one blank space. If the reaction is a three-body reaction as well, a new line is added starting with ‘ /M/ ’ and the third body efficiency factors.

The Eddy Break-up reaction To define an eddy break-up reaction in turbulent combustion, add a line starting with keywords EBU after the reaction formula, i.e. /EBU / A ebu B ebu IOP f i Rules: • •

Version 4.02

Keyword EBU must be enclosed by two ‘/’ characters and is not case sensitive A ebu and B ebu are constansts appearing in the standard eddy break-up model, see equation (10-8). If B ebu is not zero, the product will be included in the reaction rate calculation. IOP is an integer determining which EBU model is being used: IOP = 1 : Standard EBU model, reaction rate determined by equation (10-8) 8-13


Chapter 8

Setting Up Chemical Reaction Schemes

IOP = 2 : Combined time scale model, reaction rate determined by equation (10-15) IOP = 3 : Hybrid kinetic/EBU model, reaction rate determined by equation (10-11)

•

For IOP = 2 or 3, the corresponding pre-exponential factor, temperature exponent and activation energy are defined as usual. f i is optional and represents the burnt fuel mass fraction used in ignition modelling (see Chapter 10, “Ignition” in the Methodology volume)

In an eddy break-up reaction defined as A1 + A2 + … → B1 + B2 + … species A1 will be treated as the fuel, A2 as the oxidizer and B1 as the product. The mass fractions corresponding to these species are denoted in equation (10-8) as Y F ,Y O ,Y P , respectively. The eddy break-up reaction is also assumed to be irreversible. An example reaction mechanism file is shown in Table 8-2. Table 8-2 2.24E4 0. H + O2 = OH + O O + O = O2 2.62E16 –0.84 /M/ H2/2.40/ H2O/5.40/ CH4/2.00/ CO/1.75/ CO2/3.60/ HCO = CO 1.2 + H 5.00E12 0. /M/ CO + O = CO2 1.80E10 0. 6.020E14 0. /LOW / H + CH2 = CH3 6.0E14 0. 1.04E26 –2.76 /LOW / .7830 74.0 2941.0 /TROE/ CH4 + 2O2 = CO2 + 2H2O 0. 0. /EBU / 4. 0. 1 CH + N2 = HCNN 3.1E12 0.150 /M/ H2/2.0/ H2O/6.0/ CH4/2.0/ CO/1.5/ CO2/2.0/ /LOW / 1.3E25 –3.16 /TROE/ 0.667 235.0 2117.0

16795 0 19208

! modified

2385 3000 0. 1600 6964 0. 0. 0.0 740 4536.0

Setting Up Chemical Reaction Schemes Step 1 Go to the “Select Analysis Features” panel and choose option Chemical Reaction from the Reacting Flow menu. Click Apply. The Reacting Flow sub-folder will appear in the NavCenter tree, nested inside folder Thermophysical Models and Properties.

8-14

Version 4.02

Chapter 8

CHEMICAL REACTION AND COMBUSTION Setting Up Chemical Reaction Schemes

Step 2 Open the Reacting Flow sub-folder to display a second sub-folder called Chemical Reactions. This contains all panels needed to fully define a chemical reaction scheme. Step 3 Go to the Chemical Reactions sub-folder, open panel “Scheme Definition” and select a free scheme number using the Chemical Scheme # scroll bar at the bottom of the panel. You must then: •

Specify the basic reaction type (Unpremixed/Diffusion, Premixed, Partially Premixed, or Heterogeneous/Surface) by choosing an option from the Reaction Type menu • Select the most appropriate reaction model for your problem from the Reaction Model menu. The menu options depend on the reaction type specified above. • For some models, you will also need to specify the form of their Implementation or the method of calculating the Unburnt Gas Temperature, as explained in the “Scheme Definition” Help topic. Step 4 In the “Reaction System” panel, use the on-line help provided to assist you in specifying the relevant chemical reaction definitions, control settings and model parameters. pro-STAR associates all chemical species defined in this panel with additional scalar variables of the same name and also does a stoichiometric check for every reaction. The required scalars and their properties are retrieved from pro-STAR’s built-in database. Note that: •

If a species cannot be mapped to a material in a database, a warning is displayed in the Output window and a fresh scalar of that name (but with undefined properties) is created and added to the scalars list. You should therefore go to the “Molecular Properties (Scalar)” panel to specify the missing properties before proceeding further. It is also important that definition of all domain (material) properties via panel “Molecular Properties” has already been completed before any scalar properties are defined. • If the mass fraction of a non-reacting species is to be included in the calculations, assign a scalar variable to the species via the “Molecular Properties (Scalar)” panel and put it at the end of the existing scalars list. • The parameters of a reaction can be redefined at any time by selecting its parent scheme via the Chemical Scheme # scroll bar and then making the necessary changes. Step 5 In the “Ignition” panel, choose an ignition model or ignition start-up scheme, depending on the chemical scheme type defined in Step 3 Step 6 If required, go to panel “Emission” and activate the built-in pollutant emission models for NOx and/or soot

Version 4.02

8-15


Chapter 8


Step 7 Some schemes allow the inclusion of knock modelling as part of the overall chemical reaction simulation process. Parameters for this model may be specified in panel “Knock”. Step 8 If the Coupled Complex Chemistry model is in use, go to panel “Solution Controls” to select the appropriate solution method controls and to perform the necessary species-to-scalar mapping. Step 9 Go back to Step 3 and repeat the above process until all schemes have been defined. Step 10 Assign a reaction scheme to every fluid domain in your model using the “Scheme Association” panel. Note that it is not necessary to assign every available scheme to one of the domains. This allows you to define redundant schemes and then experiment with different schemes for the same domain, by performing separate analyses for each combination. In multi-domain problems where each domain has a different scalar composition, the “Additional Scalars” panel (Equation Behaviour sub-folder) enables you, in effect, to select which scalars exist in what domain. It is strongly recommended to make use of pro-STAR Constants 64, 89 and 90 when running combustion cases. Their effect is as follows: •

• •

Setting Constant 64 = 2 will constrain calculated values for all active scalar mass fractions to the range 0.0 — 1.0. Thus, numerical under- and over-shoots that can destabilize the solution process may be avoided Constant 89 can be assigned to the minimum allowable temperature calculated by STAR Constant 90 can be assigned to the maximum allowable temperature

Useful general points for local source and regress variable schemes 1. You are strongly recommended to perform stoichiometric checks for every reaction, especially if Step 4 above found missing scalars that were subsequently defined manually. To do this, click the Check Stoichiometry button in the Reaction System tab when you have finished setting up the model and before writing data to the problem (.prob) file. 2. For steady-state problems involving reactions that use a hybrid model, experience so far has shown that the best practice is to obtain a converged solution first, using only the eddy break-up model for all reactions. The chemical kinetic model should then be employed by selecting the Combined/User option and the analysis continued using the hybrid model until the final solution is obtained. 3. The steady-state under-relaxation factors for temperature T and all scalar variables representing transported mass fraction, mixture fraction, etc. should be identical. The recommended range is 0.3 to 0.7. Note that this factor has no effect for scalars calculated by other means, e.g. by an internal algebraic equation. 4. The residual error tolerance for temperature and all scalar variables can be 8-16

Version 4.02

Chapter 8


5. 6. 7.

8.

9.

10.

tightened from the default value of 0.01 to 0.001. This will increase the number of sweeps per PISO iteration but will improve the accuracy. The turbulent Prandtl and Schmidt numbers for all scalar variables should be identical. For premixed flames, the value of mixture fraction is known and remains constant throughout the analysis. When defining domain material properties via the “Molecular Properties” panel in STAR-GUIde, you are recommended to choose option Polynomial in the “Specific Heat” pop-up menu. This will load suitable polynomials from the CHEMKIN or CEC thermodynamic databases [1, 2]. A polynomial variation for molecular viscosity and thermal conductivity can be specified in the same way. For mass diffusivity, set via the “Diffusivity” panel in STAR-GUIde, the Constant option is recommended for maximum efficiency, particularly in the case of turbulent combustion. If the same reaction appears in more than one scheme, user input can be reduced by employing command RSTATUS to copy the reaction definition from a previous scheme to the current one. If modelling considerations demand it, individual reactions in multi-step reaction systems can be turned on or off at appropriate points in the simulation. This may be done by selecting Off in the Status pop-up menu corresponding to the reaction concerned. Chemical reactions (especially those for combustion) commonly take place in a domain where air is the background material. Given that the nitrogen component is often chemically inert and therefore does not appear in a chemical reaction equation, it is convenient to include N2 as a separate scalar to represent the background material. Therefore: (a) If N2 does not appear in a reaction definition, pro-STAR will automatically set up an extra active scalar called N2. By default, its physical properties are those for nitrogen and the solution method is set to Internal (see panel “Additional Scalars”). The value of the N2 mass fraction returned by STAR is such as to make the mass fractions at every cell sum to 1.0 (b) If N2 is present in a reaction definition, N2 will be set up like any other scalar and its solution method will be set to Transport.

11. If you are modelling an EGR system, the recirculated gases must be explicitly defined as active transported scalars within STAR Guide’s “Additional Scalars” folder. These must also be given names that are different from those of the parent species participating in the chemical reaction and make sure that their properties (as defined in the “Molecular Properties (Scalar)” panel) are correct. STAR will then be able to distinguish between species representing products of the chemical reactions and the ones coming from the EGR stream. 12. Complex chemistry models must be run in double precision. 13. If a regress variable is employed by a combustion model, its initial value must be set to 1 for correct model operation. You must therefore ensure that the regress variable scalar in your model is initialised properly before proceeding with the simulation.

Version 4.02

8-17


Chapter 8


Chemical Reaction Conventions The following conventions should be observed when typing reaction definitions in the “Reaction System” panel: 1. Enter the ‘→’ symbol as two consecutive characters ‘->’ 2. Specify the leading reactant as the first chemical substance on the left-hand side of the reaction equation. Its name will appear in the Leading Reactants list at the bottom of the panel, once the reaction details are confirmed. 3. Specify up to three ordinary reactants taking part in the reaction(s). Their names will appear in the Reactant Parameters list, once the reaction details are confirmed. 4. If a reaction constituent only occurs on the right-hand side of all reaction equations, it will be assumed to be a product and its name will appear in the Products list. However, if you wish this constituent to be a reactant (see, for example, point no. 4. on page 8-2), type the symbol [R] immediately after its name. 5. In multiple reaction schemes, the normal rule for what may appear as a product is as follows: (a) Reaction 1 is allowed to produce leading reactants 2 to 30 as products (b) Reaction 2 is allowed to produce leading reactants 3 to 30 as products (c) Reaction 3 is allowed to produce leading reactant 4 to 30 as products . . . (d) Reaction 29 is allowed to produce leading reactant 30 as a product (e) Reaction 30 is not allowed to produce any leading reactants For example, the two equations in the following scheme CH 4 + 1.5 O2 → CO + 2H 2 O [ r ] CO + 0.5 O 2 → CO 2 should be defined in the order shown above and not the other way round in order to satisfy this rule. The system in this example also includes an influx of H 2 O from an external source so that both O 2 and H 2 O are reactants in this case. Therefore, the symbol [R] needs to be entered after the latter’s name. 6. Note that, point no. 5 above notwithstanding, STAR will still allow one reaction only to create a product that has already been defined as the leading reactant of a previous reaction. Useful points for PPDF schemes 1. In single-fuel PPDF models, the quantities f and g f are automatically assigned by STAR-CD as scalar numbers 1 and 2. For the multiple-fuel model, the quantities f p , g f , f s and g ξ become scalar numbers 1 to 4, respectively. 2. Any additional variables are assigned to further scalars, beginning with scalar 8-18

Version 4.02

Chapter 8


number 3 (single-fuel) or 5 (multiple-fuel). This can be confirmed by displaying a STAR-GUIde panel that contains a Scalar list (for example, “Initialisation” in the Additional Scalars folder). 3. In adiabatic PPDF applications: (a) Remember that only the quantities given in item 1 above are calculated from transport equations. Temperature, density and all other variables are calculated internally. However, if any additional non-reacting scalars are defined (see “Setting Up Chemical Reaction Schemes”, Step 4) these are solved in the normal way. (b) pro-STAR provides a reminder that density is no longer calculated by one of the normal options. Thus the density setting in the “Molecular Properties” panel is automatically changed to read PPDF. (c) Polynomial coefficients should be supplied in terms of molar fractions (kmol/kmol). However, scalar concentrations for initial and boundary conditions should be specified as mass fractions. (d) If the molecular weights of all scalar species are correctly specified, STAR will output the calculated species concentrations in terms of mass fractions. However, if all species molecular weights are assigned the same value, the output will be in terms of species mole fractions. 4. In non-adiabatic PPDF applications, check the information displayed by the STAR-GUIde interface to ensure that: (a) Option Active is selected from the Influence pop-up menu for all chemical species (“Molecular Properties (Scalar)” panel in folder Additional Scalars) (b) Option Chemico-Thermal is selected from the Enthalpy pop-up menu (“Thermal Models” panel in folder Liquids and Gases) (c) The Ideal-f(T,P) option is used for density (see topic “Density”) (d) The Polynomial option is used for specific heat (see topic “Specific Heat”) (e) The scalar species concentrations are specified in terms of mass fractions 5. If the PDF is to be calculated by numerical integration, a number of control parameters should be specified. These are illustrated in the Figure below:

Version 4.02

8-19


Chapter 8


φ

}

MF ×× ×

N1

N2

fs

Ni

Ni+1

0

Figure 8-1

NF

f

1

Control parameters for PDF integration

The quantities shown in Figure 8-1 are defined as follows: (a) f s — stoichiometric mass fraction (b) N F — mixture fraction points. This is the total number of locations where chemical equilibrium calculations are performed. (c) MF — multiplying factor. This is the number of points added between any two adjacent points, such as N i and N i + 1 . These extra points are used for improving the resolution of the calculation and their values are extrapolated from those at N i and N i + 1 . The total number of points Nt used in the integration is given by (8-2) N t = ( MF + 1 ) × ( N F – 1 ) + 1 . (d) P F — integration partition. This parameter represents the percentage of points used to resolve the region between 0 and f s in the mixture fraction space, i.e. the number of points in this region is given by ( P F ⁄ 100 ) N t 6. When using the laminar flamelet model, the following points should be borne in mind: (a) Each flamelet library refers to a different strain rate. A typical example might be to have 6 flamelet libraries at strain rates of 0, 25, 50, 200, 400 and 1000 s-1. (b) Calculating flamelet libraries may be very time consuming. Therefore, when creating a new library, you should consider restarting the calculation from the nearest available strain rate wherever possible. However, if the difference in strain rate is quite large and convergence becomes difficult, it will be necessary to specify a new set of initial conditions and start again. (c) STAR-CD provides an option for either specifying the inlet strain rate or 8-20

Version 4.02

Chapter 8


calculating it via a built-in code. For simplified reaction mechanisms, try the first alternative combined with the restart option from a previously converged strain rate. For more complex mechanisms, you may want to try the second alternative, check what strain rate the code calculates, and then change the initial conditions accordingly. When the initial conditions are sufficiently close to the desired strain rate, you may be able to select the first alternative with a restart option to achieve a solution. (d) The results of each flamelet library calculation are printed out in a separate output panel. You should always inspect that panel to ensure the displayed values are reasonable. (e) If your problem setup contains multiple reaction scheme definitions, any laminar flamelet model(s) should appear at the top of the reaction scheme list. Useful points for complex chemistry models 1. The distinction between premixed, partially premixed and unpremixed combustion made in the pro-STAR GUI is irrelevant for complex chemistry models, since transport equations are solved for all species (or one of them is calculated as 1 – ΣY i ). Hence, this model is available for all the above reaction types. 2. The calculation of reaction rates can be very time-consuming. Users may therefore specify, via Constant 173, a temperature limit below which reaction rates will not be calculated. The default value of this limit is 300 K but may be re-set as necessary. 3. The steady-state complex chemistry solver employs an internal sub-timestep whose default value is 10–5. Users may change this value via Constant 154. Normally, a very small sub-timestep value will result in the calculation of large reaction rates, which could in turn make the solution of the steady-state transport equations unstable. On the other hand, if the value is too large, the chemistry solver will become very time-consuming. 4. For very stiff problems, the maximum number of sub-timesteps may need to be increased beyond its default value, currently set at 500. This is done via Constant 192. Users can also change the chemistry solver’s relative and absolute convergence tolerance via Constants 123 and 124, respectively. The default values for these are set at 10–4 and 10–10, respectively. 5. There is a balance between robustness and convergence rate. The latter may be increased by higher values of the species under-relaxation factor, but users should be careful that the stability of the solution is not sacrificed at the same time. 6. For steady-state cases, it is recommended that the initial species distribution should correspond to a non-combustible mixture, such as air. Useful points for ignition models 1. Shell and 4-step ignition models: Option Use Heat of Reaction in the “Reaction System” STAR GUIde panel is valid only when the pro-STAR-defined specific heat polynomial coefficients are used. When the reaction is exothermic, the heat of reaction value is negative. For an Version 4.02

8-21


Chapter 8

Setting Up Advanced I.C. Engine Models

endothermic reaction, the value is positive. 2. CFM ignition model: In problems where only a part of the solution domain is being simulated, you need to specify (via Constant 142) a geometrical factor whose value is the fraction of the flame kernel area in the partial (simulation) domain relative to the entire domain. For example, for a wedge-shaped solution domain in a cylindrical system and with the ignition point lying on the axis, this value should be ∆θ ⁄ 360 , where ∆θ is the angular extent of the wedge. The default value of the above factor is 1.

Setting Up Advanced I.C. Engine Models The notes below describe the set up of combustion simulations that employ the CFM, ECFM, ECFM-3Z or Level Set models. Note that: • •

The GUI facilities presented are available only when running the Auto Mesh version of pro-STAR (prostar -amm). Use of ECFM, ECFM-3Z and their attendant ignition models requires a special licence obtainable from CD-adapco.

Note also that if you are resuming from an .mdl file in which a combustion model has been defined, it is important to delete this model, its submodels (such as NOx, Soot and Knock) and the associated scalars before selecting an alternative model. This requires issuing the following pro-STAR commands (or performing the equivalent GUI operations): SCDEL,ALL — delete all scalars SOOT,n,OFF — turn off soot modelling, where n stands for every curently-defined chemical scheme number NOX,n,OFF — turn off Nox modelling KNOCK,n,OFF — turn off knock modelling CRDEL,ALL — delete all chemical reaction schemes CHSCHEME,m,NONE — remove chemical scheme associations with STAR domains (streams), where m stands for every currently-defined domain CHER,OFF — turn off chemical reaction calculations To set up a case: Step 1 In the Select Analysis Features panel, select all general features required for an engine combustion simulation (these parameters will be selected in advance if the es-ice engine simulation expert system is used): •

Time Domain > Transient (select option Angle and enter values for RPM and Initial Position) • Reacting Flow > Chemical Reaction • Multi-Phase Treatment > Lagrangian Multi-phase (if modelling sprays) Step 2 Select folder Thermophysical Models and Properties > Liquid and Gases and then enter appropriate values in the following panels:

8-22

Version 4.02

Chapter 8

CHEMICAL REACTION AND COMBUSTION Setting Up Advanced I.C. Engine Models

• • •

Molecular Properties Turbulence Models Thermal Models (select Temperature Calculation On and then choose options Conservation > Static Enthalpy and Enthalpy > Thermal) • Initialization • Monitoring and Reference Data • Buoyancy (where applicable) Step 3 At this stage, the special IC set-up panel can be used, accessed by selecting Advanced > IC Setup from pro-STAR’s main menu. The panel shown below will then pop up:

•

The panel will initially display the Analysis setup sub-panel. Check that the Combustion option is selected and then choose the type of combustion model required from the drop-down menu underneath. • Fuel parameters: Select the desired fuel from the second drop-down menu. Depending on the model type, i.e. spark or compression, the panel will display a default octane or cetane number in the text box on the right. You should replace this with an appropriate value if necessary. The corresponding chemical reaction formula will also be displayed below the fuel name. Step 4 Click the Combustion tab button on the left to display the Combustion sub-panel. Its contents depend on the combustion model selected, as described below.

Version 4.02

8-23


Chapter 8


Coherent Flame model (CFM)

The chemical reaction for the chosen fuel is displayed at the top. •

•

Modelling parameters — enter values for coefficients α and β to be used in the source term of the Σ equation (see equation (11-10) in the Methodology volume). Ignition parameters: (a) In the Spark time box, input the time (in degrees crank angle or in seconds) at which the spark is to be discharged (b) Enter the ignition Location in terms of X, Y, Z coordinates relative to

8-24

Version 4.02

Chapter 8


coordinate system no. 1 (in model units) (c) In the Delay box, enter the time delay (or transition time t1 , see Chapter 11, “Ignition treatment for the CFM-ITNFS model” in the Methodology volume) (d) In the Kernel diameter box, enter the appropriate value in mm • • •

Enter the Mixture fraction for the premixed air-fuel mixture Specify whether exhaust gas recirculation (EGR) should be On or Off Specify whether emissions (NOx) and/or Knock is to be modelled by selecting On or Off from their respective drop-down menus.

Apart from the above input, the panel will also execute commands to effect the following changes (which will overwrite any property settings specified previously in pro-STAR’s Molecular Properties panel): • • •

Set up appropriate property definitions for material #1 and change the molecular viscosity setting to MultiComponent Define the specific heat setting of the background fluid as Polynomial Create chemical species scalars and assign appropriate physical properties to them, imported from the built-in property database props.dbs

Note that there is an alternative method of setting up a CFM model using pro-STAR’s Chemical Reactions panels in STAR GUIde. The main difference between the two is that • •

pro-STAR specifies the ignition location in terms of the centroid of an ‘ignition cell’ the IC Setup panel specifies the location in terms of its X, Y, Z coordinates and passes them on to STAR via an Extended Data segment delimited by the keywords BEGIN SPARK_DATA and END SPARK_DATA and appended to the .prob file.

It is most important that the two approaches should not be mixed.

Version 4.02

8-25


Chapter 8


Extended Coherent Flame model (ECFM)


•

8-26

Modelling parameters — enter values for coefficients α and β to be used in the source term of the σ equation (see equation (11-90) in the Methodology volume). The PSDF Moments soot model may be used by selecting On from the Mauss Soot Model drop-down menu (see “Soot Modelling” on page 8-39) Version 4.02

Chapter 8


•

Ignition parameters — there are two ignition model choices: (a) Standard: i) In the Spark time box, enter the time (in degrees crank angle or in seconds) at which the spark is to be discharged ii) Enter the ignition Location in terms of X, Y, Z coordinates relative to coordinate system no. 1 (in model units) (a) Aktim — see “The Arc and Kernel Tracking ignition model (AKTIM)” on page 8-33

•

Enter a Sector angle value if you want to perform a “Sector Mesh” analysis.

The panel also executes additional commands to effect the following changes: • •

Version 4.02

Turn on the Transient setting and associate the ECFM chemical scheme with material #1 Define 17 scalars and their material properties

8-27


Chapter 8


Extended Coherent Flame model 3Z (ECFM-3Z) — spark ignition


•

8-28

Modelling parameters — enter values for coefficients α and β to be used in the source term of the σ equation (see equation (11-90) in the Methodology volume). Select the Multiple Cycles option if you wish to run a simulation over multiple engine cycles Version 4.02

Chapter 8


• •

•

The PSDF Moments soot model may be used by selecting On from the Mauss Soot Model drop-down menu (see “Soot Modelling” on page 8-39) Parameter Start of ECFM-3Z is the time at which fuel / air mixing and combustion will start to be calculated. The user should set this time equal to the time when fuel injection starts, in degrees crank angle or seconds. Ignition parameters — there are two ignition model choices: (a) Standard: i) In the Spark time box, enter the time (in degrees crank angle or in seconds) at which the spark is to be discharged ii) Enter the ignition Location in terms of X, Y, Z coordinates relative to coordinate system no. 1 (in model units) (b) Aktim — see “The Arc and Kernel Tracking ignition model (AKTIM)” on page 8-33

•

• •

Multiple ignition locations — the number of locations can be increased/decreased by clicking the up/down # location arrows. Coordinates for each ignition location can be entered by selecting the particular location from the drop-down menu. Specify whether Knock is to be modelled by selecting On or Off from the drop-down menu Enter a Sector angle value if you want to perform a “Sector Mesh” analysis.

In addition, the panel defines 25 appropriate scalars and their material properties. Extended Coherent Flame model 3Z (ECFM-3Z) — compression ignition


• •

Version 4.02

Modelling parameters — enter values for coefficients α and β to be used in the source term of the σ equation (see equation (11-90) in the Methodology volume). Select the Multiple Cycles option if you wish to run a simulation over multiple engine cycles The PSDF Moments soot model may be used by selecting On from the Mauss Soot Model drop-down menu (see “Soot Modelling” on page 8-39) 8-29


Chapter 8


•

•

Parameter Start of ECFM-3Z is the time at which fuel / air mixing and combustion will start to be calculated. The user should set this time equal to the time when fuel injection starts, in degrees crank angle or seconds. There is an additional option to turn On the Tabulated Double-Delay Autoignition Model (see “The Double-Delay autoignition model” on page 8-37).

Useful points for ECFM models 1. The information entered in the ECFM and ECFM-3Z panels is passed on to STAR via an Extended Data segment delimited by the keywords BEGIN ECFM_DATA and END ECFM_DATA. This segment is created automatically and may be inspected by selecting Utilities > Extended Data from the main pro-STAR window’s menu bar but the user does not need to add any further information to it. The segment is appended to the end of the .prob file when the later is saved at the end of the current pro-STAR session. 2. All ECFM models must be run in double precision. 3. Although the list of fuels for use in these models is limited, users can supply their own fuel definition by selecting option User Defined from the fuel selection drop-down menu and then specifying the number of C and H atoms. The specific heat ‘cp’ should be changed accordingly. 4. ECFM models cannot be used in conjunction with the k-ω or the Spalart-Allmaras turbulence models. 5. Additional scalars may be added but they must be Inactive. 6. It is important to remember that species N, O, H, OH cannot be present as part of EGR gases. 7. Model parameters should be specified via the panels described in this document. Users should not attempt to supply any parameters via the standard pro-STAR STAR GUIde panels. 8. If ECFM models are applied in materials other than no. 1, make sure that the solution method for scalars in these materials is the same as for material 1. The correct setting is Transport, except for scalar RVB (the progress variable ‘c’) which should be Internal. 9. The ECFM model is currently emulated by the ECFM-3Z model. Tracers for species NO, CO, H2 and SOOT are turned off, along with O2UM and FUM, and the corresponding mass fractions are set to 0.0 (i.e. the system is always “mixed”). Comparisons with results from earlier STAR-CD implementations will show slightly differences as different sub-models for the post-flame regime are used in the current version. 10. For unburnt gases, the initial mass fractions of fuel and oxygen tracers (TF, TO2) plus any other applicable tracers (TCO, TH2, TNO, TSOOT) must be set equal to the corresponding initial mass fractions of species Fuel and O2 plus, if applicable, species CO, SOOT, NO, H2.

8-30

Version 4.02

Chapter 8


Level Set model

The chemical reaction for the chosen fuel is displayed at the top. • • •

Re-initialization method — select an option from the adjacent pop-up menu (see Chapter 9, “Re-initialisation” in the Supplementary Notes volume) Select the Multiple Cycles option if you wish to run a simulation over multiple engine cycles Ignition parameters (a) In the Spark time box, enter the time (in degrees crank angle or in seconds) at which the spark is to be discharged (b) Enter the ignition kernel Location in terms of X, Y, Z coordinates relative to coordinate system no. 1 (in model units) and ignition kernel Radius (in metres) (c) Enter the ignition Duration (in degrees crank angle or in seconds)

•

Multiple ignition kernel locations — the number of locations can be increased/decreased by clicking the up/down # location arrows. Data for each ignition kernel can be entered by selecting the particular location from the drop-down menu.

The information entered in the Level Set panel is passed on to STAR via an Extended Data segment delimited by the keywords BEGIN LEVELSETDATA and END LEVELSETDATA. This segment is created automatically and may be inspected by selecting Utilities > Extended Data from the main pro-STAR window’s menu bar. The user does not normally need to add any further information to it unless advanced features of the model need to be implemented, as described in Chapter 9 of the Supplementary Notes volume. The Extended Data segment is appended to the end of the .prob file when the later is saved at the end of the current pro-STAR session. Version 4.02

8-31


Chapter 8


Step 5 Click the Write data tab button on the left to display the Write data sub-panel, shown below. Write Data sub-panel

In this panel: •

• •

Click Save parameters to save the combustion parameters to a file called star.ics. This file is important as it contains combustion data and should be kept together with the .mdl file. The file name can be changed if necessary. Click Write and close to write the problem settings to a model (.mdl) file and close the panel Click Close to close the panel without writing anything to the .mdl file

Step 6 Initialize certain scalars, as shown below. Only those scalars included in a list should be initialised; all others should have 0 as their initial value. •

For ECFM: (a) (b) (c) (d) (e) (f) (g)

•

For ECFM-3Z: (a) (b) (c) (d) (e)

8-32

Fuel → set the fuel initial mass fraction O2 → set the O2 initial mass fraction CO2 → set the CO2 initial mass fraction H2O → set the H2O initial mass fraction N2 → set the N2 initial mass fraction TF → set the unburnt fuel initial mass fraction TO2 → set the unburnt O2 initial mass fraction Fuel → set the fuel initial mass fraction O2 → set the O2 initial mass fraction CO2 → set the CO2 initial mass fraction H2O → set the H2O initial mass fraction N2 → set the N2 initial mass fraction Version 4.02

Chapter 8


TF → set the unburnt fuel initial mass fraction TO2 → set the unburnt O2 initial mass fraction TCO → set the unburnt CO initial mass fraction TH2 → set the unburnt H2 initial mass fraction TNO → set the unburnt NO initial mass fraction TSOOT → set the unburnt soot initial mass fraction FUM → set the initial mass fraction of the Unmixed Fuel (default value is 0) (m) O2UM → set the initial mass fraction of the Unmixed Oxygen

(f) (g) (h) (i) (j) (k) (l)

Step 7 Define the time step size and any load steps in the appropriate transient settings panel (where applicable). The Arc and Kernel Tracking ignition model (AKTIM) AKTIM may be used only with one of the ECFM options, as follows: • • • • •

• • •

Version 4.02

Select the required ECFM model Choose option Aktim from the Ignition menu In the Spark time box, enter the time (in degrees crank angle or in seconds) at which the spark is to be discharged Enter the ignition Location in terms of X, Y, Z coordinates relative to coordinate system no. 1 (in model units) If the electrodes are represented as distinct entities in the mesh, select option Regions from the Electrode Model menu and input the boundary region numbers corresponding to the anode and cathode. Otherwise, select Area and temperature from the Electrode Model menu and input the surface area and temperature of the anode and cathode. The two alternatives are illustrated below. Specify the Diameter of the cathode and anode electrodes Input values for parameters E 2 ( t SI ) , L , R in the Secondary Circuit section of the panel Multiple ignition locations — the number of locations can be increased/ decreased by clicking the up/down # location arrows. Coordinates for each ignition location can be entered by selecting the particular location from the drop-down menu.

8-33


Chapter 8


8-34

Version 4.02

Chapter 8


Useful points for the AKTIM model Use of Extended Data The information contained in the above panels is passed on to STAR via the ECFM Extended Data segment (see “Useful points for ECFM models” on page 8-30). The contents of this segment do not need to be altered by the user except in the case of two-dimensional (x-y) problems. For such problems: • Version 4.02

Select Utilities > Extended Data from the main pro-STAR window’s menu 8-35


Chapter 8


• •

bar to display the current ECFM_DATA segment in the Extended Data panel. Enter a new line containing the keyword TWODS anywhere within the segment (typically at the end) Save and then Close the Extended Data panel

Entering the above keyword enforces the turbulent fluctuation contribution to the calculation of the flame kernels’ positions in the z-direction to be zero. Post and track data Two new files are written by STAR when AKTIM is in use: 1. casename.strk, the track file for spark particles 2. casename.ktrk, the track file for flame kernel particles Track files The .strk and .ktrk files can be used in the same way as .trk files for droplets (see Chapter 7, “The Particle Track File” in the Post-Processing User Guide). They can be loaded via the Plot Droplets/Particle Tracks panel by choosing option Track File in the Load Droplet Data section and then specifying the appropriate file name and extension. Note that: • This action will erase all current droplet track data • Only option Constant is supported in the Droplet Radius menu • Only option Color is supported in the Fill Color menu Post files The .ccm file contains all solution data and can be used to plot spark particles and flame kernel particles. These can be loaded into pro-STAR via command GETD,POST,SPARK for spark particles or GETD,POST,FLAME for flame kernel particles. Note that: •

•

After the .ccm file is loaded, the wrinkling factor, progress variable, mass and burnt gas mass in the flame kernel particle can be plotted by selecting Diameter, Temperature, Mass and Count, respectively, from the Fill Color menu in STAR GUIde’s Plot Droplets/Particle Tracks panel No other options in the Fill Color menu are supported

There are no files equivalent to .pstt or .ccm_timestep for the spark and flame kernel particles. HPC issues For HPC calculations, the electrodes must be meshed or specified in only one processor domain. In addition, spark particles will not be permitted to pass between different processor domains. Mesh quality It is important to use good quality meshes when the electrodes are resolved. This is because the source terms might be very high during the ignition process. Thus, localized and big numerical errors due to poor quality meshes may lead to wrong results, instabilities and/or divergences.

8-36

Version 4.02

Chapter 8


The Double-Delay autoignition model This model may be employed by selecting ECFM-3Z, compression and then turning On the Tabulated Double-Delay Autoignition Model option, as shown below. A new scalar YIG2 is created to track the primary autoignition progress. Scalar YIG monitors the progress of the main autoignition.

STAR requires three types of data for this model: 1. The primary ignition delay time tdelay1 2. The secondary ignition delay time tdelay2 3. The burnt fuel mass fraction ffrac Each of the above is tabulated at nT values of temperature T, nP values of pressure p, nE values of equivalence ratio E and nX values of residual gas mole fraction X. Hence, a 4-dimensional array structure is employed for each data item. By default, the model uses ignition delay data read from pre-computed tables. Users may also use their own data by storing them in three separate text files called delay1.dat, delay2.dat and ffrac.dat for the primary delay, secondary delay and burnt fuel mass fraction, respectively. The files must be placed in the case working directory. User data creation procedure The following steps are necessary to create user-generated data files: • • • •

Select Utilities > Extended Data from the main pro-STAR window’s menu bar to display the current ECFM_DATA segment in the Extended Data panel. Enter a new line containing the keyword LU2DATA in the line after LAUTO2 Save and then Close the Extended Data panel Create the three data files described above in a format suitable for STAR input. The Fortran program listed below shows how these files should be written.

Program Example_UserData c c nT, nP, nE, nX, Tmin, Tstep, p, E, X, tdelay1, tdelay2, ffrac are user input data. c Tmin, Tstep are the minimum temperature and the temperature step (K) in the table, c respectively, so that c T(1)=Tmin, T(2)=T(1)+Tstep, T(3)=T(2)+Tstep, ... c p(1),...,p(nP) are the nP pressure (bars) points in the table in ascending order.

Version 4.02

8-37


Chapter 8

Setting Up Advanced I.C. Engine Models c c c c c c c... c

E(1),...,E(nE) are the nE equivalence ratio points in the table in ascending order. X(1),...,X(nX) are the nX residual gases points in the table in ascending order. tdelay1(iT,iP,iE,iX), tdelay2(iT,iP,iE,iX), ffrac(iT,iP,iE,iX) are user-defined 4-dimensional arrays for the primary delay, secondary delay and burnt fuel mass fraction. File "delay1.dat" Double Precision Tmin,Tstep,p,E,X,tdelay1,tdelay2,ffrac Dimension p(nP), E(nE), X(nX) Dimension tdelay1(nT,nP,nE,nX) Dimension tdelay2(nT,nP,nE,nX) Dimension ffrac(nT,nP,nE,nX)

c open(iunit,file=’delay1.dat’) write(iunit,*) nT, nP, nE, nX write(iunit,*) Tmin, Tstep do ip=1,nP write(iunit,*) p(ip) end do do ie=1,nE write(iunit,*) E(ie) end do do ix=1,nX write(iunit,*) X(ix) end do do ix=1,nX do ie=1,nE do ip=1,nP do it=1,nT write(iunit,*) tdelay1(it,ip,ie,ix) end do end do end do end do close(iunit) c c... File "delay2.dat" c open(iunit,file=’delay2.dat’) do ix=1,nX do ie=1,nE do ip=1,nP do it=1,nT write(iunit,*)tdelay2(it,ip,ie,ix) end do end do end do end do close(iunit) c c... File "ffrac.dat" c open(iunit,file=’ffrac.dat’) do ix=1,nX do ie=1,nE do ip=1,nP do it=1,nT write(iunit,*)ffrac(it,ip,ie,ix) end do end do end do end do close(iunit) c end

Note that if any of the above files is not present, STAR will abort the simulation with warning messages. For verification purposes, STAR will also output three text files at the beginning of the simulation containing the parameters and data specified by the user. The files are called check_data1 for the primary delay, check_data2 for the 8-38

Version 4.02

Chapter 8

CHEMICAL REACTION AND COMBUSTION NOx Modelling

secondary delay and check_dataf for the burnt fuel mass fraction and are written into the working directory for sequential runs or into the first processor’s directory for HPC runs.

NOx Modelling NOx concentration is usually low compared to other species in combustion systems. As a result, it is generally agreed that NOx chemistry has negligible influence and can be decoupled from the main combustion and flow field calculations. The recommended procedure for performing a NOx analysis is as follows: Step 1 Set up the combustion model as usual. Step 2 In the Chemical Reactions folder of STAR GUIde, open the “Emission” panel and then go the “NOx” section. Select option On from the NOx Model menu to activate STAR-CD’s built-in NOx subroutines. Step 3 Turn on the appropriate NOx production mechanism from the Thermal, Prompt or Fuel menus (see Chapter 10, “NOx Formation” in the Methodology volume). Option User in any of these menus enables you to perform the necessary calculations via subroutine NOXUSR. If option On is selected for Thermal NOx, specify values for the required constants as explained in the on-line help topic for “NOx”. Step 4 Check that pro-STAR has created an extra passive scalar variable called NO, by opening the “Molecular Properties (Scalar)” panel in the Additional Scalars folder and inspecting the currently defined scalars. If the problem requires the prediction of fuel NOx (this is only applicable to nitrogen-containing fuels, e.g. coal), check that an additional passive scalar called HCN has also been created. Step 5 If your model provides for the calculation of OH and H mass fractions, their values will be used in equation (10-84) of the Methodology volume to implement the extended Zeldovich mechanism. Step 6 For steady-state problems, make sure that a sufficient number of iterations has been performed for the solution of NO and (if present) HCN to have converged.

Soot Modelling The Flamelet Library soot model is applicable only to unpremixed and partially premixed reactions and is activated via the Emission panel’s “Soot” section in STAR GUIde. The only user input required is four scaling factors, see equation (10-120) and (10-121), that determine the magnitude of the contribution from each source term. For example, a decrease in the value of the scaling factors for positive source Version 4.02

8-39


Chapter 8

Soot Modelling

terms (surface growth and particle inception) results in slower formation of soot. In diffusion flames, this can shift the point of maximum soot volume fraction further downstream. A typical range for these factors is 0.5 — 10.0 and their default value is 1. The PSDF Moments soot model may be used in conjunction with any ECFM combustion model (see “Setting Up Advanced I.C. Engine Models”) and is accessed via the GUI facilities presented when running the Auto Mesh version of pro-STAR (prostar -amm). The relevant panel is shown below:

To use this model: • • •

Select the required ECFM model Select option On from the Mauss Soot Model menu Choose the number of Moments to be solved for (0, 2, 3, or 4). The effect of EGR (if present) will be taken into account. If 0 is selected, “The Flamelet Library method” described in Chapter 10 of the Methodology volume will be used. Enter scaling factors (see “The method of moments” on page 10-29 of the Methodolgy volume for definitions) for the following quantities:

•

(a) (b) (c) (d)

Surface Growth Fragmentation Particle Inception Oxidation Rate

Use of this method with models other than ECFM is possible only within the pro-STAR environment. The steps required for set-up in this case are as follows: 1. 2. 3. 4.

8-40

Select a combustion model that allows Soot to be employed Select option Soot on in the Emissions panel If no moments need to be solved, bypass steps 4, 5 and 6 Define 2, 3 or 4 additional passive scalars with names M0, M1, M2, M3, depending on how many moments are to be solved for. These represent the quantities Mr ⁄ ρ with Mr the r-th moment and units of [mol/m3]. STAR will Version 4.02

Chapter 8

CHEMICAL REACTION AND COMBUSTION Coal Combustion Modelling

identify them as the moments for the soot model. 5. Select option Transport in the Analysis Control > Solution Control > Equation Behaviour > Additional Scalars panel as the method of solution for these scalars 6. Deactivate the SOOT scalar, as in this case the soot mass fraction is obtained from the soot PSDF moments solution 7. If EGR is present (up to 40% in concentration), then: (a) Define 3 new active scalars with names EGR_CO2, EGR_N2, EGR_H2O. These distinguish the CO2, N2, H2O arising from the EGR from the CO2, N2, H2O arising from the combustion (b) Assign the same molecular properties to them as for scalars CO2, N2, H2O in the Molecular Properties (Scalar) panel. in addition, the properties in the Binary Properties panel should be the same as for scalars CO2, N2, H2O. (c) Select option Transport in the Analysis Control > Solution Control > Equation Behaviour > Additional Scalars panel as the method of solution of the EGR scalars. Note that EGR can be set up for soot cases even if no moments are solved for. The EGR for soot calculation is not supported for PPDF models. If at least two moments are solved for, the following (mass averaged) data will be produced at each time-step: Soot Volume, SootMass, Number Density, Mean Diameter, Dispersion Size Distribution, Variance of the Size Distribution, Surface Density These are added after the scalar data in the .spd file (see Chapter 9, “Engine Combustion Data Files” in this volume).

Coal Combustion Modelling Coal combustion models involve two-phase flow with complex solid and gas phase chemical reactions. To reduce CPU time, it is recommended to run such a simulation as a two-stage process using the STAR GUIde system. Thus, initially the problem is run isothermally. Then, once a reasonably converged solution is obtained, the problem is re-run with coal combustion turned on. An outline of the steps involved and recommendations on model set-up at each stage of the process is given below: Stage 1 Run the model as an isothermal (non-reacting) problem and obtain a converged solution which effectively serves as an initial condition for the flow field. Step 1 Generate a mesh for the problem as usual and check that the steady-state analysis mode has been chosen in the “Select Analysis Features” panel. Step 2 Check that the temperature calculation is switched on in the “Thermal Models” panel (Liquids and Gases sub-folder) and select an appropriate turbulence model in the “Turbulence Models” panel.

Version 4.02

8-41


Chapter 8

Coal Combustion Modelling

Step 3 Define all boundaries and set up boundary conditions throughout, including appropriate temperature distributions at inlet boundaries. Step 4 Run the case until a reasonable flow field is established (see Note 1 below). Stage 2 Turn on coal combustion and generate the final solution as follows: Step 1 Go to the “Select Analysis Features” panel and choose option Coal Combustion from the Reacting Flow menu and Lagrangian Multi-Phase from the Multi-Phase Treatment menu. Click Apply. The Coal Combustion sub-folder will appear in the NavCenter tree, nested inside folders Thermophysical Models and Properties > Reacting Flow. An additional sub-folder, Lagrangian Multi-Phase, will also open in the NavCenter tree. At the same time, pro-STAR will set up your model automatically for this type of analysis, using the ‘Constant Rate’, ‘1st-Order Effect’ and ‘Mixed-is-burnt’ sub-models as defaults for volatiles, char and gas combustion, respectively. Step 2 If radiation effects are to be modelled, go to the “Thermal Options” panel in folder Thermophysical Models and Properties and select the appropriate radiation model, as described in Chapter 7 of the CCM User Guide (see also Note 2 below). Step 3 Go to the Coal Combustion sub-folder and supply or modify data in each of its panels in turn: •

In the “Coal Composition” panel, enter coal composition data and click Apply. The data supplied in this panel can be stored in a file called coal.dbs by entering a coal name and clicking Save to D/B. Alternatively, you can read the coal composition from an existing file by clicking Open D/B. • •

•

8-42

Data entered in the Proximate analysis tab should be supplied on an air dried basis. Data in the Ultimate analysis tab should be supplied on a ash-free basis, where C + H + O + N + Chlorine (Cl) + Sulphur (S) = 1. The Cl and S components are then assumed to be ash. If Cl and S need to be included in the calculations, these components can be modelled separately via user subroutine PARUSR at a later point. Details of this method are given in Chapter 3 of the Supplementary Notes volume describing coal blend modelling. This chapter also covers all other aspects of this type of model, such as specifying the components and reaction rates for the different coals in the blend. Supply the coal Q factor, which is an adjustment for volatile matter, in the Miscellaneous tab. Studies have shown that under certain heating conditions, a significantly higher amount of volatile matter can be devolved than that measured by the standard proximate analysis test. This Version 4.02

Chapter 8


effect is accounted for by the Q factor which is defined by (V*/VM) = Q •

•

• •

•

Version 4.02

where V* is the volatiles yield and VM the proximate analysis matter. Also in the Miscellaneous tab, enter the net calorific value of the fuel and the fraction of total nitrogen in the volatiles. Finally, choose the volatiles specific heat option (see Note 3 below).

In the “Sub Models” panel, select the desired models for volatiles, char and gas combustion. •

•

(8-3)

To speed up convergence, the Constant Rate scheme should be selected initially in the Volatiles tab. The initial devolatilisation temperature should be set to that of the primary inlet flow containing the coal particles (this will help with initialising the temperature field and instigating ignition). After the rest of the coal particle parameters have been set, the problem should be run using an Initial Field Restart (panel “Analysis (Re)Start”) from the isothermal solution obtained in Stage 1 and run for several hundred iterations or until a reasonably stable solution is reached. The devolatilisation temperature should then be raised to a more realistic level (e.g. 550 K) and the model run once again using a Standard Restart until a stable solution is reached. The Single Step or 2-Competing steps model should then be chosen. These require values for pre-exponential factors and activation energy that should be determined experimentally or taken from the available literature (as is the case with the default values used by STAR). The solution should then be re-run with a standard restart (see Notes 4 and 5 below regarding changing parameters or submodels in this panel). In the Char tab, select one of the three char models available. Char rates can be determined experimentally or taken from the literature. In the Gaseous Combustion tab, select either Mixed-is-Burnt (fast chemistry approach) or the 1-step or 2-step EBU models. When this is set, pro-STAR will create the appropriate scalars and select the transport or internal solution method for them, depending on the model chosen. Char oxidation products can be changed for specific gaseous combustion models using Constant 120, as explained in “Switches and constants for coal modelling” below.

In the “NOx/Radiation” panel, turn on the NOx generation and/or coal particle radiation options, as required. Note that, if the latter is chosen, you should already have set up your model for radiation calculations as described in Chapter 7. Enter the particle emissivity value. The NOx model can be turned on near the end of the solution as it has only a small effect on the overall flow field. In the “Control/Printout” panel, specify the required solution control and printout parameters. It is sometimes necessary to initially reduce the under-relaxation factor of the particle source term (to as low as 0.1) in order to achieve a stable solution. The factor may be increased later on in the solution. The iteration number at which particle source term averaging starts 8-43


Chapter 8


should be set to a high value to ensure a converged solution can be reached (see Note 6 below). Step 4 Go to Chemical Reactions sub-folder and open the “Emission” panel. Turn on the appropriate NOx models, i.e. thermal, fuel and prompt NOx. Then go to the “Scheme Association” panel and click Apply to assign the chemical reaction schemes defined above to the current fluid domain. Step 5 Go to the “Initialisation” panel (Additional Scalars sub-folder) and set up an appropriate initial mass fraction for the carrier fluid (e.g. for air YO2 = 0.233, YN2 = 0.767). Step 6 Go the “Lagrangian Multi-Phase” folder, check the settings for the Lagrangian two-phase modelling scheme and make any changes/additions necessary for defining coal particle initial positions, entrance behaviour and physical properties (panel “Droplet Physical Models and Properties”): •

Turbulent dispersion can be turned on in the Global Physical Models tab to predict a realistic particle track behaviour. • In the Droplet Properties tab, ensure that all values of Hfg in the Component Properties list are set to 0. Step 7 Switch off the heat and mass transfer time scale calculation by going to the “Switches and Real Constants” panel (Other Controls sub-folder) and setting constants C71 and C72 to 1.0. Step 8 Go to the “Thermal Models” panel and check that options Static Enthalpy and Chemico-Thermal have been selected for the enthalpy equation. Step 9 Go to the “Scalar Boundaries” panel (Define Boundary Conditions folder) and adjust the scalar mass fractions at the inlet boundaries. Step 10 Go to the “Analysis (Re)Start” panel (Analysis Preparation/Running folder) and set up the analysis as a restart run, beginning from the solution obtained in Stage 1. Step 11 Run the case until the solution converges or reasonably small residuals are achieved. Useful notes 1. If the coal model is turned off to run the case in non-reacting mode, it may be necessary to first turn off /delete the chemical scheme definition already set-up in the “Scheme Definition” panel. 2. To avoid solution instabilities and reduce computer time in radiation cases, it is advisable to run the simulation for several hundred iterations with radiation turned off before switching it back on to complete the simulation. 8-44

Version 4.02

Chapter 8


3. In the “Coal Composition” panel (Miscellaneous tab) you have three options for setting the volatiles specific heat: (a) Coal CV (the recommended option) uses the enthalpy of volatiles calculated from the heat balance of the coal combustion for both char and volatiles. (b) CH4 assumes that all volatiles are methane. (c) Mass Weighted considers the mixture of volatile components and calculates the overall volatile enthalpy as the sum of the products of mass fraction and enthalpy for each individual volatile component, i.e. H = ΣY i h i 4. If you change any parameter in the “Coal Composition” panel, the chemical reaction scheme is changed and therefore the gas combustion scheme in the “Sub Models” panel must be reset. When you re-apply the gas composition scheme, this also resets the scalars involved. Therefore, you must (a) re-apply the scheme association, (b) initialise the additional scalars for the background fluid, (c) reset Hfg = 0 for all components in the Lagrangian “Droplet Physical Models and Properties” panel, and (d) define scalar boundary conditions for the inlet. 5. Accurate modelling can be achieved through input of appropriate values for devolatilisation and char rates. Manipulating these values can increase the solution accuracy, while changes in the turbulence models employed can lead to more accurate prediction of the flow aerodynamics. 6. The iteration number at which to begin source term averaging should be set to a high value so as to ensure that a stable initial flow field has been achieved (and also to economize on computer time expended). This number has a default setting of 50 iterations and should be altered to a value suitable for establishing a stable flow field. This may be done by setting Constant 24 to the desired value. 7. When discretising the coal particle size distribution, it is important to include some sub-5 micron particles. This enables a stable flame to be established in the immediate vicinity of the burner inlets. 8. When starting the coal combustion calculations in Stage 2, it is important to use the constant rate devolatilisation option for all particles, and to make the devolatilisation temperature equal to the particles’ initial temperature. This is the numerical equivalent of ‘lighting up’ the combustion system in a real-life situation. Switches and constants for coal modelling It is sometimes advisable to define some of the following pro-STAR Constants and Switches when setting up a coal model: • • • • • Version 4.02

Constant 24 — see Note 6 above Constant 64 = 2 — constrain active scalar values to the range 0.0 — 1.0 Constant 71 = 1 — deactivates the mass transfer time scale Constant 72 = 1 — deactivates the heat transfer time scale Constant 82 — specifies the coal particle emissivity in radiation problems 8-45


Chapter 8


• • • •

Constant 88 — specifies the maximum coal particle temperature Constant 89 — specifies the minimum carrier fluid temperature limit Constant 90 — specifies the maximum carrier fluid temperature limit Constant 120 — specifies special options for the char reaction, as follows: (a) Constant 120 = 0 — use the default V3.26 settings (b) Constant 120 = 1 — the char reaction product is CO (c) Constant 120 = 2 — the char reaction products are CO and CO2, where Y CO –T * ----------- = A exp ⎛ --------- ⎞ ⎝ T ⎠ Y CO2

(8-4)

and the default values are A = 3.0×108 and T* = 30,193 • • •

Constant 70 — used for changing the value of A in conjunction with Constant 120 Constant 74 — used for changing the value of T* in conjunction with Constant 120 Switch 71 — specifies an implicit calculation of the source terms in the particle energy equation (can improve algorithm stability)

Special settings for the Mixed-is-Burnt and Eddy Break-Up models When Constant 120 is used, the following settings are also required depending on the combustion model that has been chosen: For Mixed-is-Burnt: • •

Constant 120 = 0 — no extra scalars need to be defined; the product of the char reaction is CO2 Constant 120 = 1 — two extra scalars are needed, to be defined in the Additional Scalars > Molecular Properties panel: (a) MIX_CO — this is a passive scalar representing the CO mixture fraction and is to be solved by a transport equation. The latter is specified in panel Analysis Controls > Solution Controls > Equation Behavior > Additional Scalars panel by choosing option Transport from the Solution Method menu. (b) CO — this is an active scalar to be solved algebraically. This is specified in panel Analysis Controls > Solution Controls > Equation Behavior > Additional Scalars panel by choosing option Internal from the Solution Method menu.

•

Constant 120 = 2 — three extra scalars are needed, to be defined in the Additional Scalars > Molecular Properties panel: (a) MIX_CO — this is a passive scalar representing the CO mixture fraction and is to be solved by a transport equation. The latter is specified in panel Analysis Controls > Solution Controls > Equation Behavior > Additional Scalars panel by choosing option Transport from the Solution Method menu. (b) MIX_CO2 — this is a passive scalar representing the CO2 mixture fraction and is to be solved by a transport equation. The latter is specified

8-46

Version 4.02

Chapter 8


in panel Analysis Controls > Solution Controls > Equation Behavior > Additional Scalars panel by choosing option Transport from the Solution Method menu. (c) CO — this is an active scalar to be solved algebraically. This is specified in panel Analysis Controls > Solution Controls > Equation Behavior > Additional Scalars panel by choosing option Internal from the Solution Method menu. For Eddy Break-Up: One-step model — Constant 120 cannot be used because the char reaction product can only be CO2 Two-step model • •

Version 4.02

Constant 120 = 0 — no extra scalars are needed; the char reaction product is CO Constant 120 = 2 — no extra scalars are needed; the char reaction products are CO and CO2

8-47

Chapter 9

LAGRANGIAN MULTI-PHASE FLOW Setting Up Lagrangian Multi-Phase Models

Chapter 9

LAGRANGIAN MULTI-PHASE FLOW The theory behind Lagrangian multi-phase problems and the manner of implementing it in STAR-CD is given in the Methodology volume, Chapter 12. The present chapter contains an outline of the process to be followed when setting up a Lagrangian multi-phase simulation, including details of the user input required and important points to bear in mind when setting up problems of this kind.

Setting Up Lagrangian Multi-Phase Models Step 1 Go to panel Select Analysis Features in STAR GUIde and choose option Lagrangian Multi-Phase from the Multi-Phase Treatment menu. Click Apply. The Lagrangian Multi-Phase folder will appear in the NavCenter tree, containing a number of panels that are appropriate to this type of analysis. Step 2 In the first panel, “Droplet Controls”, set various solution control parameters (see the on-line Help text for more details). The same panel also defines how droplet parcel initial conditions (entrance behaviour and location) are to be specified. The available options are: •

Spray injection with atomization — use one of the built-in nozzle and atomisation models (see Chapter 12, “Nozzle flow models” and “Atomisation models” in the Methodology volume). These are especially useful in internal combustion engine studies. • Explicitly defined parcel injection — explicit (‘manual’) setting of all required quantities. This option also allows the use of distribution functions for the droplet diameters. • User Subroutine — specify everything via a user subroutine Step 3 The second panel, “Droplet Physical Models and Properties”, defines dispersed phase heat, mass and momentum transport mechanisms (including inter-droplet and wall collisions), plus droplet physical properties. Several different droplet types may coexist in your model, so properties are specified for each individual type. Step 4 The folder’s remaining panels relate to splitting droplets into parcels for modelling purposes and defining the latter’s entrance behaviour (initial velocities and entrance properties). How this is done depends on the option chosen in Step 2; the folder will display the appropriate panel for each choice: 1. Spray injection with atomization — opens a single panel, “Spray Injection with Atomization”, in which you specify the fuel mass flow rate entering the solution domain through an injection nozzle. The liquid fuel is converted into droplets whose injection velocity depends on the nozzle model characteristics. In addition, a number of atomisation models are employed to determine the distribution of droplet diameters and velocity directions on exit from the nozzle. 2. Explicitly defined parcel injection — opens the following two panels: Version 4.02

9-1

LAGRANGIAN MULTI-PHASE FLOW

Chapter 9

Setting Up Lagrangian Multi-Phase Models

(a) “Injection Definition” sets up parcel entrance conditions, in terms of either velocity and rotation components or nozzle parameters (b) “Injection Points” defines parcel entrance locations The association between conditions and locations is made by first dividing parcels into injection groups that share the same entrance conditions. All entrance locations defined subsequently are then assigned to one of these groups. The concept is illustrated by the example shown in Figure 9-1 below: Droplet Type 1 Momentum ON Heat transfer ON Properties of Heptane

Injection Group 1

Droplet Type 2 Momentum ON Heat transfer OFF Properties of Water

Injection Group 2

Injection Group 3

Injection Definition Constant Diam. Wi = –5 m/s mT = 0.05 kg/s 2 parcels/point

Injection Definition Rosin-Ram PDF Vi = 2 m/s mT = 0.02 kg/s 3 parcels/point

Injection Definition Normal PDF Wi = 7 m/s mT = 0.05 kg/s 1 parcel/point

Injection Points Set 1: Line, 3 pts Set 2: Circle, 6 pts

Injection Points Set 1: Single point Set 2: Line, 8 pts

Injection Points Set 1: Boundary, 12 pts

Set 2, 6pts

Set 2, 8pts

Set 1, 12 pts

Set 1, 3pts Single Parcel Set 1, 1pt

Figure 9-1

Injection Point

Illustration of terminology for explicitly defined parcel injection

3. User Subroutine — opens a single panel, “Droplet User Subroutine”, that calculates all parcel initial conditions through user coding 9-2

Version 4.02

Chapter 9

LAGRANGIAN MULTI-PHASE FLOW Setting Up Lagrangian Multi-Phase Models

Note that the Spray injection and Explicitly defined options are mutually exclusive. Thus, if you change your mind about which method to use for specifying initial conditions, you will need to go back to panel “Droplet Controls”, pick the other method and overwrite the previous definitions. On the other hand, User Subroutine may be used in conjunction with either of the above options, i.e. STAR will take the definitions supplied in subroutine DROICO into account as well as the spray or explicit definitions. Step 5 Check the result of the parcel initialisation process graphically by displaying the parcels in the context of a plot of the domain into which they are launched, as illustrated in Figure 9-2:

Figure 9-2

Plot of droplet initial conditions

This is done by going to the Post-Processing folder, panel “Plot Droplets/Particle Tracks” and using the plotting facilities of the “Droplets” tab, as explained in the on-line Help text. Alternatively, choose Post > Get Droplet Data from the main window menu bar to display the Load Droplet Data dialog shown below and perform the same function from there.

Version 4.02

9-3


Chapter 9

Data Post-Processing

Data Post-Processing pro-STAR provides special facilities for visualising the results of a Lagrangian multi-phase flow analysis. These facilities fall into the following two categories: 1. Static displays — these show the location of one or more droplets at a given point in time. Alternatively, they may also be used to show successive positions of a given droplet as it progresses through the solution domain. The droplets are represented by small circles, as shown in Figure 9-3. The circle size and colour can be made to depend on a variety of local droplet properties.

Figure 9-3

Static display illustration

2. Trajectory displays — these show droplet tracks, either as continuous trajectories or as animated streaks, whose rate of progress through the solution domain can be controlled by the user, as illustrated in Figure 9-4.

9-4

Version 4.02

Chapter 9

LAGRANGIAN MULTI-PHASE FLOW Data Post-Processing

Figure 9-4

Trajectory display illustration

Static displays Steady-state problems Step 1 Read the required droplet data from the track (.trk) file generated automatically by the STAR-CD solver for Lagrangian flow problems. To do this, use panel “Plot Droplets/Particle Tracks”, tab “Droplets”. Step 2 If necessary, use command DTIME to specify a time range over which you want droplet track data to be plotted. The display will then include only locations visited by droplets during this time interval. Step 3 Use the “Droplets” tab controls to choose options appropriate to the plot you want to create. Note that a droplet display may be superimposed on a post data plot by choosing Plot > Cell Display > Droplets from the main window menu (or by issuing command CDISPLAY, ON, DROPLET) before the cell plotting operation. If the plot is a contour plot and the droplet fill colour varies according to a physical property, a secondary scale will be displayed for that droplet property. If the droplets are filled with a single arbitrary colour, and droplet velocity vectors are displayed, the secondary scale will correspond to droplet velocity magnitude, as represented by the vector colours.

Version 4.02

9-5


Chapter 9


Step 4 Select a set of parcels whose progress through the solution domain is to be displayed. The selection procedure is analogous to that described in Chapter 2 of the Meshing User Guide regarding sets of cells, vertices, splines, etc. Thus, sets may be selected by • • •

a coloured button marked D -> on the left-hand-side of the main window a similar button labelled Dp in the “Droplets” tab typing command DSET in the I/O window. This provides the most extensive range of selection options.

The set selection facilities available via the D -> or Dp buttons are as follows: 1. 2. 3. 4. 5. 6. 7.

All — puts all parcels in the set None — clears the current set Invert — selects all unselected parcels and clears the current set New — replaces the current set with a new set of parcels Add — adds new parcels to the current set Unselect — deletes parcels from the current set Subset — selects a smaller group of parcels from those in the current set

For the last four items, the target parcels may be assembled by choosing an option from a secondary drop-down list, as described below. In every case, what constitutes a valid option depends on how droplet data were read into pro-STAR: 1. For all loading choices, option Cell Set selects parcels that are contained within the physical space occupied by the current cell set. If the choice was Track File (see Step 1 on page 9-5), all droplet tracks whose initial positions fall within the current cell set are selected. 2. If the loading choice was Droplet Initial Conditions (see Step 5 on page 9-3) or Current Post Data File (see Step 2 on page 9-8), the following options are available: (a) Cursor Select — click on the desired parcels with the cursor; complete the selection by clicking the Done button on the plot (b) Zone — use the cursor to draw a polygon around the desired parcels. Complete the polygon by clicking the last corner with the right mouse button (or click Done outside the display area to let pro-STAR do it for you). Abort the selection by clicking the Abort button. 3. If the loading choice was Current Post Data File, the following options are available: (a) Active — select all active parcels (b) Stuck — select all parcels that have stuck to a wall and become immobilised Note that droplet set information is not saved in the restart (.mdl) file on completion of the post-processing run Step 5 Display the selected parcels as a series of droplet circles by clicking Droplet Plot in the “Droplets” tab 9-6

Version 4.02

Chapter 9

LAGRANGIAN MULTI-PHASE FLOW Data Post-Processing

The locations of the circles represent the points where a parcel intersects cell boundaries as it travels from the beginning to the end of its path through the mesh. Step 6 If detailed numerical information is required on the selected parcels, choose Lists > Tracks from the main window menu bar to open the Particle/Droplet Track Data dialog. Select the track file and then click Load Data to read in and display all available information in that file, as shown below:

The required information is displayed by clicking the appropriate parcel number (shown in the Track column) with the mouse. The same information (but in a different format) can also be displayed on the I/O window by typing command PTPRINT. Special data requirements In some situations, the user may require the following additional information: 1. The position of a range of parcels at a given point in time, as opposed to a specified parcel at a series of time points. The data needed for such a display may be obtained by interpolation of the available data at the time point in question using command PTREAD. Continue by specifying the appropriate parcel set and then use the “Droplets” tab in STAR GUIde to display the required droplet distributions. Note that the time specified in PTREAD is independent of any time information specified via command DTIME (see Step 2 above) 2. The ‘age’ of all currently-loaded parcels, given by command DAGE. A parcel’s age is defined as the interval between the time when the first parcel entered the solution domain and the time when the parcel in question hits a wall or exits from the solution domain. Age is calculated from data in the Version 4.02

9-7


Chapter 9


track (.trk) file and may be used as the basis for selecting a parcel set, via command DSET. This information may be listed in the I/O window using command DLIST. Transient problems Step 1 Decide which time step is to be inspected and then load the corresponding data (from file case.pstt), using STAR GUIde’s “Load Data” panel (“File(s) tab”). If more than one transient file is available, pro-STAR will locate the right one automatically. Step 2 Open panel “Plot Droplets/Particle Tracks” (“Droplets” tab) and read the contents of the transient file by selecting option Current Post Data File from the pop-up menu at the top. Step 3 Choose appropriate options in the Droplet Plot Options section of the same tab, as for “Steady-state problems”. Step 4 Select the desired parcel set using the most appropriate of the methods described under “Steady-state problems”. Step 5 Plot droplets by clicking the Droplet Plot button. Step 6 Information about a range of parcels at the current time step can also be displayed in the I/O window using command DLIST. For example, DLIS,1,50,2,OTHER will list the density, diameter, mass, droplet count and temperature of every second parcel between 1 and 50. Information on parcel ‘age’ is also obtainable with this command (having first executed command DAGE). In transient problems, age is defined as the interval between the time when the first parcel entered the solution domain and the current time. Trajectory displays Trajectory displays are basically droplet track displays. These are plotted as continuous trajectories or animated streaks, using the options provided in panel “Plot Droplets/Particle Tracks” (“Droplets” tab). As for particle tracks generated at the post-processing phase, the data required for such plots are stored in file case.trk. This file is generated automatically during the Lagrangian multi-phase analysis for both transient and steady-state calculations. Note that: • • 9-8

It is also possible to print position, velocity and other droplet data stored in case.trk for each track using command PTPRINT. The data in this file will be overwritten if the user generates post-processing Version 4.02

Chapter 9

LAGRANGIAN MULTI-PHASE FLOW Engine Combustion Data Files

particle tracks without first saving the droplet data.

Engine Combustion Data Files In addition to the normal results files, engine combustion cases also produce additional output data (.spd) files, written by STAR if the Lagrangian multi-phase and/or combustion simulations options are in use. One such file is produced for every fluid domain in your model and contains both fuel droplet data (represented as globally averaged quantities) and general engine data. The information in this file may also be displayed in graphical form using the utilities provided in STAR GUIde’s Graphs folder (see panel “External Data”). The meaning of the quantities appearing in the file is as follows: Name

Meaning

T-Step

Time step number

Time

Elapsed time at this time step [s]

Crank_Ang.

Crank Angle [degrees]

Average_P

Cylinder absolute average pressure [pa]

Average_T

Cylinder absolute average temperature [K]

Average_d

Cylinder average density [kg/m3]

Cylinder_Mass In-cylinder mass [kg]

Version 4.02

Tot_Inj_Lqui

Total injected mass [kg]

Cur_mas_Fue

Total mass of liquid phase [kg]

Evaporated

Total evaporated mass [kg]

Evaprt_%

Ratio of total evaporated mass to the total injected mass [%]

Leading_par

Unused

Distance

Unused

Velocity

Unused

V_mag

Unused

Idr

Unused

Sauter_D

Sauter mean diameter [m]

AngMom_X

Fluid angular momentum w.r.t. the X-axis of the local coordinate system used in the model [kg/m2s]

AngMom_Y

Fluid angular momentum w.r.t. the Y-axis

AngMom_Z

Fluid angular momentum w.r.t. the Z-axis

Mass_Burnt

Burnt fuel mass [kg]

9-9


Chapter 9

Useful Points

%Evap_Burnt

Burnt fuel as a percentage of fuel evaporated

Heat_Release_R Heat release rate [J/s] ate Scalar

Mass of scalar no. i [kg]

Note that, depending on the model, some of the above data may have no meaning.

Useful Points 1. The above treatment is strictly valid only for droplets whose physical dimensions are appreciably smaller than those of a typical mesh cell through which they travel. It is recommended that the total droplet volume (i.e. volume of a typical droplet times the number of droplets in the parcel) should not exceed 40% of this cell volume. 2. If a convergent solution cannot be easily obtained in steady-state models using the coupled approach, it may be beneficial to start the analysis by obtaining a solution that does not include the dispersed phase. The latter should then be introduced into the calculated flow field and the analysis restarted using the Initial Field Restart option to produce the final, complete solution. 3. In steady-state models using the uncoupled approach, the computer time required may again be reduced by obtaining the solution in two stages. First, a converged solution without the dispersed phase should be calculated. The dispersed phase should then be introduced and the analysis restarted using the Initial Field Restart option to obtain the desired solution in one iteration only. 4. In transient analyses involving droplets that move faster than their surrounding fluid, the Courant number used for estimating a reasonable time step size (see Chapter 5, “Load step definition”) should be based on the droplet rather than the fluid velocity. 5. STAR-CD’s default treatment for heat transfer coefficients can be combined with user-calculated mass transfer coefficients and vice-versa. In practice, however, the user will most probably want to use the same calculation procedure for both of them. 6. Complex or unusual physical conditions relating to momentum, heat and mass transfer between droplets and the continuous phase can be accommodated by supplying user subroutines DROMOM, DRHEAT and DRMAST that describe each transfer process, respectively. Similarly, special conditions relating to the momentum, heat and mass transfer behaviour of droplets at wall boundaries can be specified by supplying the required relationship via subroutine DROWBC.

9-10

Version 4.02

Chapter 10

EULERIAN MULTI-PHASE FLOW Introduction

Chapter 10

EULERIAN MULTI-PHASE FLOW

Introduction The theory behind problems of this kind is given in the Methodology volume, Chapter 13. This chapter contains an outline of the process to be followed when setting up an Eulerian multi-phase analysis. Also included are cross- references to appropriate parts of the on-line Help system, containing details of the user input required.

Setting up multi-phase models Step 1 Switch on the Eulerian multi-phase simulation facility using the “Select Analysis Features” panel in STAR GUIde: • •

Select Eulerian Multi-Phase from the Multi-Phase Treatment menu Click Apply. pro-STAR checks if another multi-phase simulation option (Lagrangian, Free Surface, Cavitation) is already on. If so, it issues a warning message and turns it off. • An additional sub-folder called Eulerian Multi-Phase now appears in the NavCenter tree, within the Thermophysical Models and Properties folder. Step 2 Set up the mesh and define the boundary region locations as usual. At present, only part of the full STAR-CD boundary type set is available for this kind of analysis. The permissible options are: 1. 2. 3. 4. 5. 6. 7. 8. 9. 10.

Inlet Outlet Pressure Wall Non-porous baffle Cyclic Symmetry Degassing Attachment Monitoring

Note that: •

The above list contains an additional boundary type, ‘Degassing’, valid only for Eulerian multi-phase flows. This permits dispersed phase mass to escape into the media surrounding the solution domain (see also Chapter 4, “Phase-Escape (Degassing) Boundaries” in this volume). Your problem should not contain more that one boundary of this type. • Only the currently available boundary types, as listed above, can be set up via the “Create Boundaries” panel. Step 3 Open the Thermophysical Models and Properties folder and use each of its sub-folders to provide relevant information about your problem. Note that: Version 4.02

10-1


Chapter 10

Setting up multi-phase models

• •

Thermal/solar radiation is not supported in this version of the code. Use the Liquids and Gases panels to specify physical properties and special flow conditions in your model. Note that: (a) Only a single domain (or material) is allowed at present, so the Material # slider in each panel remains set to 1. (b) Where appropriate, data are entered per phase, with the number of phases currently restricted to two. Of these, no. 1 is treated as the continuous and no. 2 as the dispersed phase. (c) “Molecular Properties” — compared to single-phase problems, only a restricted range of options is available for evaluating physical properties. The specification process and permissible options are common to both phases. (d) “Turbulence Models” — if turbulent flow conditions prevail, specify a method for calculating the turbulence characteristics of both phases and also the turbulence-induced drag (e) “Thermal Models” — if heat transfer is present in the analysis, turn on the temperature solver for each phase as required (f) “Initialisation” — specify initial conditions for each phase (g) “Monitoring and Reference Data” — supply a reference pressure and temperature and the cell location corresponding to the reference pressure. The values specified apply to both phases. (h) “Buoyancy” — if buoyancy effects are important, specify a datum location and reference density. Again, these values apply to both phases.

•

The current version does not support the following features:

(a) Multi-component mixture problems requiring the presence of additional scalar variables in either phase. Therefore, STAR GUIde does not display the Additional Scalars sub-folder. (b) Porous media flow, therefore the Porosity sub-folder is not displayed. (c) Chemical reactions of any kind, including coal combustion and the STAR/KINetics package. Therefore, the “Select Analysis Features” panel does not permit the above options to be turned on. (d) Liquid films of any kind. Again, the “Select Analysis Features” panel does not allow this option. Step 4 In the Eulerian Multi-Phase folder: •

Open the Interphase Momentum Transfer sub-folder to specify appropriate models and related parameters for this part of the analysis. The information is supplied in two separate panels: (a) “Drag Forces” — define a model for calculating drag forces directly or via the drag coefficient (b) “Other Forces” — define models for calculating other interphase forces (e.g. virtual mass and/or lift force)

•

10-2

If heat transfer is present in the analysis, use the “Interphase Heat Transfer” panel to specify the method of calculating the Nusselt number (and hence the heat transfer coefficient). Version 4.02

Chapter 10

EULERIAN MULTI-PHASE FLOW Setting up multi-phase models

•

Specify the size of the particles making up the dispersed phase using the “Particle Size” panel. At present, all particles are assumed to be of equal size. Step 5 If required by problem conditions, use the “Source Terms” panel in folder Sources to specify mass sources or additional source terms for the momentum, turbulence or enthalpy equations of either phase. At present, multi-phase sources may only be specified via user subroutines. Step 6 Specify boundary conditions using the “Define Boundary Regions” panel. The permissible range of boundary types is shown in Step 2. Note that for inlet, pressure, wall/baffle and cyclic regions, separate boundary conditions are needed for each phase. When pro-STAR’s boundary display facilities are used to check the various boundary region definitions (see Chapter 4, “Boundary Visualisation”), inlet phase velocities will be displayed according to the setting of the Phase # slider in panel “Define Boundary Regions”. Step 7 In the Analysis Controls folder: •

Select Solution Controls > Equation Behavior, open the “Primary Variables” panel and make any necessary adjustments to the current settings • If you wish to monitor the value of any flow variable(s), as a function of iteration or time step, select Output Controls > Monitor Engineering Behavior and then open panel “Monitor Boundary Behaviour” and/or panel “Monitor Cell Behaviour”. The choice depends on whether you wish to monitor values at a boundary region or within a cell set. Note that the choice of which variables to monitor is phase-dependent. • If you are running a transient problem, use the “Transient tab” in the “Analysis Output” panel to select which variables you wish to store in the transient post data file (.pstt). Note that the choice of such variables is phase-dependent. Step 8 Run STAR in double precision mode. There are two reasons for this: •

•

Solving the volume fraction equation in this manner gives rise to a smaller truncation error, especially in parts of the mesh where the volume fraction is close to 1 or 0. This is sometimes essential for convergence of the solution. Double precision cases have been more extensively tested

Step 9 Post-processing the analysis results follows the same rules as single-phase problems. Note that: •

•

Version 4.02

Analysis data are stored in the .ccm file per phase. A phase slider in the “Data tab” of panel “Load Data” enables you to select the precise data required. Likewise, phase-specific data may be plotted in a graph. The types of graph available are described in topics “Residual / Monitored History Data”, 10-3


Chapter 10

Setting up multi-phase models

“Engineering Data” and “Analysis History Data”. Useful points on Eulerian multi-phase flow 1. The momentum under-relaxation factors should be the same for both continuous and dispersed phases. The pressure under-relaxation factor should also be equal to the volume fraction factor. Suggested values for these parameters are 0.3 on momentum for both phases and 0.1 on pressure and volume fraction. 2. To ensure satisfactory convergence for steady and pseudo-transient cases, a maximum residual error tolerance of 1.0 × 10-6 is recommended.

10-4

Version 4.02

Chapter 11

FREE SURFACE AND CAVITATION Free Surface Flows

Chapter 11

FREE SURFACE AND CAVITATION

Free Surface Flows The theoretical description of free-surface flow models is given in Chapter 14 of the Methodology volume. This section contains an outline of the procedure to be followed when setting up free-surface flow problems. Also included are cross-references to appropriate parts of the on-line Help system, which contains details of the user input required. Setting up free surface cases Step 1: Define the mesh and boundary regions Set up an appropriate mesh and define its boundary regions in the usual way. All standard STAR mesh features are applicable to free-surface flows but this is not also the case for all types of boundary region. The boundary types currently supported in free-surface problems are: • • • • • • • • •

Inlet Outlet Slip and no-slip impermeable walls Symmetry planes Static and piezometric pressure (with both environmental and mean options deactivated) Baffles Cyclic boundaries (except for partial cyclics in which the mass flow rate is specified) Attachment boundaries Monitoring boundaries

Step 2: Activate the free surface model Turn on the free-surface option using the “Select Analysis Features” panel of the STAR GUIde system: • •

•

Select On from the Free Surface menu Select option Transient from Time Domain menu. Free-surface flows have to be computed in a time-marching manner, even if the final solution is steady. In the latter case, one can choose larger time steps or only one iteration per time step to save on computing time, as described in Step 7 below. Click Apply. An additional folder called Free Surface will now appear in the NavCenter tree.

Note that: 1. A passive scalar named VOF is required for free-surface problems. pro-STAR will automatically define such a scalar (if it has not been defined already) on pressing the Apply button in the Select Analysis Features folder. This scalar stores the volume fraction of the ‘heavy’ fluid in the solution domain (see “Mathematical model” on page 14-2 of the Methodology volume) and requires definition by the user of appropriate boundary conditions. 2. Certain combinations of the free surface model with other STAR-CD features Version 4.02

11-1


Chapter 11

Free Surface Flows

are not currently supported. Such features are: (a) (b) (c) (d)

Eulerian and Lagrangian multi-phase flow Reacting flow Radiative heat transfer Aeroacoustic analysis

pro-STAR will issue a warning message if an attempt is made to switch on any of the above features and the free surface model will then be switched off. Step 3: Define model control parameters In the Free Surface folder, open the “Controls” panel and specify appropriate settings for the following parameters: 1. Differencing Scheme — this defines the differencing scheme to be used for the solution of the VOF transport equation: (a) The default scheme is HRIC, which stands for ‘High-Resolution Interface Capturing’. As suggested by the name, this scheme should be employed if a sharp interface between the heavy and light fluids is to be resolved. There is also a blending factor associated with the scheme. The default value for this factor is appropriate for most situations; higher values provide a sharper interface but there is a danger of interface alignment with grid lines under unfavourable flow direction. In such a case, the blending factor should be reduced. (b) The Upwind scheme will not provide a sharp interface but it may be used on coarse or poor-quality meshes or when a sharp resolution of the interface is not an issue. 2. Surface Tension — determines whether the surface tension effect across the heavy-light fluid interface is to be included in the calculations. The effect is excluded by default as it plays an important role only in small-scale problems, where the mesh is fine enough to resolve the interface curvature on a scale that results in an appreciable pressure difference. The latter is proportional to σ/R, where σ is the surface tension coefficient and R the radius of interface curvature. Note that the HRIC scheme must be selected if you choose to include surface tension in your model. Step 4: Define material properties The fluid medium in a free-surface problem is defined as a single fluid material possessing two components: a ‘heavy’ and a ‘light’ one. To define their respective material properties, go to the Free Surface folder and open the “Molecular Properties” panel. For each of the heavy and light components, fill in the relevant property values or select materials from pro-STAR’s built-in property database and then click Apply. Step 5: Define thermophysical models In the Thermophysical Models and Properties folder: 1. If your application involves solid-fluid heat transfer: 11-2

Version 4.02

Chapter 11

FREE SURFACE AND CAVITATION Free Surface Flows

(a) Open the “Thermal Options” panel and then select Heat Transfer On in the Solid-Fluid Heat Transfer section (b) Click Apply (c) As a result of the above, a sub-folder called Solids will appear in the STAR GUIde tree. Use this sub-folder to define solid material properties. More than one solid domains may be defined in such models. Note that radiative heat transfer is not currently supported in free-surface problems. 2. If gravitational effects are important in your application, open the “Gravity” panel and define the gravitational acceleration and its direction with respect to the coordinate system of the solution domain. 3. The overall flow conditions should be specified by entering the Liquid and Gases sub-folder. This is designed to supply relevant information for the following aspects of the model: (a) Open the “Turbulence Models” panel and choose an appropriate turbulence model for your case, or select the Laminar flow option if applicable. (b) If thermal effects are important, open the “Thermal Models” panel and select the Temperature Calculation On option. In the Show Options section, choose the enthalpy formulation and transport equation to be solved for it. Please note that the following are not supported in free-surface cases: i) Stagnation Enthalpy option in the Conservation menu ii) Chemico-Thermal option in the Enthalpy menu (c) Initialise the flow field and turbulence quantities in the “Initialisation” panel. Only the Constant and User options are supported for free-surface cases. (d) Use the “Monitoring and Reference Data” panel to specify the locations of the monitoring and reference cells, as well as the reference pressure and temperature. (e) Use the “Buoyancy” panel to specify whether buoyancy effects are to be included in the calculation. Select the On button if this effect is important. It is advisable to use a Datum Density value corresponding to the ‘light’ fluid density and, if possible, to choose the Datum Location at a cell that is likely to be always occupied by the ‘light’ fluid. 4. Define ‘active’ and ‘passive’ scalars in the Additional Scalars sub-folder. A passive scalar named VOF should already be defined at this stage. Note that: (a) When the Cavitation option is turned On (see “Cavitating Flows” on page 11-5), an active scalar named CAV also needs to be defined (b) Apart from CAV, no other active scalar can be defined for free-surface problems (c) You may define as many ‘passive’ scalars as are necessary for your model (d) The diffusion term in the transport equation for all scalars defined in a free-surface model will be switched off Version 4.02

11-3


Chapter 11

Free Surface Flows

5. In the Additional Scalars sub-folder, initialize the distribution of the VOF scalar by opening the “Initialisation” panel and choosing one of the available options. In most cases, either Constant values, User coding, or assignment according to cell type is applied. In the latter case, click CTAB on the main pro-STAR window to open the Cell Table Editor and use it to set the Initial Free Surface Material switch to Light (for those cell types initially occupied by the ‘light’ fluid) or to Heavy (for cell types initially occupied by the ‘heavy’ fluid). 6. For free-surface problems involving porous media, use the Porosity sub-folder to define properties for the porous materials in the normal way. 7. Use the Sources sub-folder to define external source terms for momentum, turbulence, enthalpy and scalars. Note that: (a) User Coding is the only supported option in this case (b) The VOF transport equation does not accept additional source terms Step 6: Define boundary conditions Go to the Define Boundary Conditions folder • •

Open the “Define Boundary Regions” panel and specify appropriate boundary conditions in the usual manner Open the “Scalar Boundaries” panel and specify boundary conditions for the VOF scalar

Valid boundary types for free-surface problems are listed under Step 1. If a pressure boundary condition is specified, make sure that both the Envir Press (environmental pressure) and Mean (pressure profile mean value) options are set to Off. Step 7: Define analysis control parameters Go to the Analysis Controls folder: 1. Select the Solution Controls sub-folder and then open the “Solution Method” panel to set/adjust the solution algorithm parameters. Note that: (a) Only the SIMPLE algorithm is applicable to free-surface problems (b) Both CG and AMG solvers are applicable and the desired one may be selected from the Solver Type menu. AMG is recommended since it usually leads to shorter computing times. (c) Select option Euler Implicit from the Temporal Discretisation menu. This is the only option supported for this type of problem. (d) If a steady-state solution is expected, one can limit the number of outer iterations per time step (see topic “Transient problems” in the STAR GUIde on-line Help) to 1, in which case a pseudo-transient marching towards the steady state is obtained. This is applicable to both cavitating and free-surface flows, but one needs to be certain that a steady-state solution can be reached. 2. Select the Equation Behavior sub-folder and then open the “Primary Variables” panel. Make any necessary adjustments to the current or default settings in the Equation Status, Solver Parameters and Differencing Schemes 11-4

Version 4.02

Chapter 11

FREE SURFACE AND CAVITATION Cavitating Flows

tabs. 3. In the “Additional Scalars” panel: (a) Adjust the under-relaxation factor for the VOF scalar, if necessary (the default value is 0.8). (b) Ignore the section concerning the differencing scheme because the latter has already been set in the free-surface “Controls” panel. (c) Select an appropriate differencing schemes and other control parameters for all scalars other than VOF. (d) Click Apply before changing to another scalar or exiting from the panel. 4. Select the Output Controls sub-folder and use the “Monitor Numeric Behaviour” panel to print additional information such as convergence residuals and conservation checks (optional). 5. Select the Monitor Engineering Behavior sub-folder and, if you wish, use the “Monitor Cell Behaviour” panel to save selected cell data for subsequent plotting against iteration or time step number. 6. The Analysis Output sub-folder enables you to specify the frequency of saving solution results in the .ccm file. 7. Use the “Switches and Real Constants” panel in the Other Controls sub-folder to set switches and constants for any ‘non standard’ practices. Please check carefully the meaning of each switch and constant and use it only when absolutely necessary. Step 8: Define the time step size and run duration There are two ways to define the time step size and the run time length: 1. Go to the Analysis Preparation/Running folder and open the “Set Run Time Controls” panel. Fill in appropriate values for simulation time and time step size in the relevant boxes. This is the recommended way of defining time steps. 2. If your application involves a moving mesh defined by an events file, you will need to use the Advanced Transients panel by choosing Modules > Transient from the main pro-STAR window. In this panel, you can define load steps, each of which contains the time step size and number of time steps to be used for each load step. Please note that you need to go through Step 1 to Step 7 before defining load steps using the Advanced Transients panel.

Cavitating Flows The theoretical description of cavitation models is given in Chapter 14 of the Methodology volume. This section contains an outline of the procedure to be followed when setting up cavitating flow problems. Also included are cross-reference to appropriate parts of the on-line Help system, which contains details of the user input required. Setting up cavitation cases Step 1: Define the mesh and boundary regions Set up an appropriate mesh and define its boundary regions in the usual way. All standard STAR mesh features are applicable to cavitation but this is not also the Version 4.02

11-5


Chapter 11

Cavitating Flows

case for all types of boundary region. The boundary types currently supported in cavitation problems are: • • • • • • • •

Inlet Slip and no-slip impermeable walls Symmetry planes Static and piezometric pressure (with both environmental and mean options deactivated) Baffles Cyclic boundaries (except for partial cyclics in which the mass flow rate is specified) Attachment boundaries Monitoring boundaries

Step 2: Activate the cavitation model Turn on the cavitation option using the “Select Analysis Features” panel of the STAR GUIde system: • •

•

Select On from the Cavitation menu Select option Transient from Time Domain menu. Cavitating flows have to be computed in a time-marching manner, even if the final solution is steady. In the latter case, one can choose larger time steps or only one iteration per time step to save on computing time, as described in Step 7 below. Click Apply. An additional folder called Cavitation will now appear in the NavCenter tree.

Please note: 1. An active scalar named CAV is required for cavitation problems. pro-STAR will automatically define such a scalar (if it has not been defined already) on pressing the Apply button in the Select Analysis Features folder. This scalar stores the volume fraction of vapour generated during the cavitation process (see “Mathematical model” on page 14-6 of the Methodology volume) and requires definition by the user of appropriate physical properties and boundary conditions. 2. Certain combinations of the cavitation model with other STAR-CD features are not currently supported. Such features are: (a) (b) (c) (d)

Eulerian and Lagrangian multi-phase flow Reacting flow Radiative heat transfer Aeroacoustic analysis

pro-STAR will issue a warning message if an attempt is made to switch on any of the above features and the cavitation model will be switched off. 3. Combinations of the cavitation and free surface models are supported. This typically occurs in applications requiring resolution of a sharp interface between a cavitating liquid and a gas, in which case you may select both the Free Surface and Cavitation options. If both are selected, a passive scalar called VOF is defined automatically by pro-STAR (unless this definition already exists in the model). 11-6

Version 4.02

Chapter 11


4. When the cavitation and free surface options are combined, physical properties for the gas component are defined using the “Molecular Properties” panel in the Free Surface folder. You will also need to define the differencing scheme for the VOF scalar, as discussed in Step 3 of the “Free Surface Flows” section. Step 3: Define material properties The fluid medium in cavitating flows is defined as a single fluid material consisting of two or three components. Thus, for cavitation without a free surface, there is a heavy component and a vapour component; for cavitation with a free surface, there is a heavy component, a light component and a vapour component. •

•

•

For cases involving cavitation only, go to the Cavitation folder and open the “Molecular Properties” panel. Two tabs labelled Light Fluid and Heavy Fluid will appear, of which the Light Fluid one is always inactive. Use the Heavy Fluid tab to define molecular properties for the cavitating liquid. Please note that displayed values for surface tension and contact angle will not be used for cases involving only cavitation. When cavitation is combined with a free surface, the definition of molecular properties for the heavy and light fluids is identical to that for free surface flows, as described in the previous section. Vapour molecular properties are defined using the “Molecular Properties (Scalar)” panel, as for any other active scalar.

Step 4: Define model parameters Go to the Cavitation folder and open the “Cavitation Model” panel. The STAR-CD default is currently the Rayleigh model but you may also define your own model by choosing the User option from the Model Selection menu. Three parameters are needed for the Rayleigh model: the Saturation Pressure, Average Nuclear Radius and Number of Nuclei contained in 1 m3 of liquid. Of these, the saturation pressure may be either constant or user-defined in subroutine CAVPRO; the other two parameters are constants. Please note that the number of nuclei per m3 of liquid has a strong influence on the amount of vapour generated and therefore requires your own knowledge as to its likely value. Although only limited measurement data are available, it is well known that liquid purity (affected by filtering, degassing and possibly other treatment) strongly affects the cavitation process. The following recommendations can be used in the absence of more specific information: • •

For small-scale, high-pressure systems such as engine injectors, a value in the range 1011 — 1014 was found to be adequate. For large-scale, low-pressure systems such as ship propellers and large pumps, smaller values in the range 106 — 1010 may be more appropriate.

However, the best choice is always the one based on your own experience. Step 5: Define thermophysical models In the Thermophysical Models and Properties folder: 1. If your application involves solid-fluid heat transfer: Version 4.02

11-7


Chapter 11

Cavitating Flows

(a) Open the “Thermal Options” panel and then select Heat Transfer On in the Solid-Fluid Heat Transfer section (b) Click Apply (c) As a result of the above, a sub-folder called Solids will appear in the STAR GUIde tree. Use this sub-folder to define solid material properties. More than one solid domains may be defined in such models. Note that radiative heat transfer is not currently supported in cavitation problems. 2. If gravitational effects are important in your application, open the “Gravity” panel and define the gravitational acceleration and its direction with respect to the coordinate system of the solution domain. 3. The overall flow conditions should be specified by entering the Liquid and Gases sub-folder. This is designed to supply relevant information for the following aspects of the model: (a) Open the “Turbulence Models” panel and choose an appropriate turbulence model for your case, or select the Laminar flow option if applicable. (b) If thermal effects are important, open the “Thermal Models” panel and select the Temperature Calculation On option. In the Show Options section, choose the enthalpy formulation and transport equation to be solved for it. Please note that the following are not supported in cavitation cases: i) Stagnation Enthalpy option in the Conservation menu ii) Chemico-Thermal option in the Enthalpy menu (c) Initialise the flow field and turbulence quantities in the “Initialisation” panel. Only the Constant and User options are supported for cavitation cases. (d) Use the “Monitoring and Reference Data” panel to specify the locations of the monitoring and reference cells, as well as the reference pressure and temperature. (e) Use the “Buoyancy” panel to specify whether buoyancy effects are to be included in the calculation. Select the On button if this effect is important. 4. Define ‘active’ and ‘passive’ scalars in the Additional Scalars sub-folder. An active scalar named CAV should already be present at this stage, defined automatically by pro-STAR. Note that: (a) If the Free Surface option is turned On as well (see “Free Surface Flows” on page 11-1), a passive scalar named VOF is also defined automatically by pro-STAR. This tracks the distribution of the ‘heavy’ fluid volume fraction. (b) Apart from CAV, no other active scalar can be defined in cavitation problems. (c) The default molecular properties of the CAV scalar are those of water vapour. You may therefore need to define alternative properties if your vapour corresponds to a different fluid. (d) You may define as many ‘passive’ scalars as are necessary for your 11-8

Version 4.02

Chapter 11


model (e) The diffusion term in the transport equation for all scalars defined in the cavitation model will be switched off 5. In the Additional Scalars sub-folder, initialize the distribution of the VOF scalar by opening the “Initialisation” panel and choosing one of the available options. In most cases, either Constant values, User coding, or assignment according to cell type is applied. In the latter case, click CTAB on the main pro-STAR window to open the Cell Table Editor and use it to set the Initial Free Surface Material switch to Light (for those cell types initially occupied by the ‘light’ fluid) or to Heavy (for cell types initially occupied by the ‘heavy’ fluid). 6. For cavitation problems involving porous media, use the Porosity sub-folder to define properties for the porous materials in the normal way. 7. Use the Sources sub-folder to define external source terms for momentum, turbulence, enthalpy and scalars. Note that: (a) User Coding is the only supported option in this case (b) The VOF transport equation does not accept additional source terms (c) When the default Rayleigh model is used, you cannot define additional source terms for the CAV scalar. Step 6: Define boundary conditions Go to the Define Boundary Conditions folder • •

Open the “Define Boundary Regions” panel and specify appropriate boundary conditions in the usual manner Open the “Scalar Boundaries” panel and specify boundary conditions for the CAV and (if applicable) VOF scalars, the latter representing the heavy fluid volume fraction.

Valid boundary types for cavitation problems are listed under Step 1. If a pressure boundary condition is specified, make sure that both the Envir Press (environmental pressure) and Mean (pressure profile mean value) options are set to Off. Step 7: Define analysis control parameters Go to the Analysis Control folder: 1. Select the Solution Controls sub-folder and then open the “Solution Method” panel to set/adjust the solution algorithm parameters. Note that: (a) Only the SIMPLE algorithm is applicable to cavitation problems (b) Both CG and AMG solvers are applicable and the desired one may be selected from the Solver Type menu. AMG is recommended as it usually leads to shorter computing times. (c) For transient cases, select option Euler Implicit from the Temporal Discretisation menu. This is the only option supported for this type of problem. (d) If a steady-state solution is expected, one can limit the number of outer iterations per time step (see topic “Transient problems” in the STAR GUIde on-line Help) to 1, in which case a pseudo-transient marching Version 4.02

11-9


Chapter 11

Cavitating Flows

towards the steady state is obtained. This is applicable to both cavitating and free-surface flows, but one needs to be certain that a steady-state solution can be reached. 2. Select the Equation Behavior sub-folder and then open the “Primary Variables” panel. Make any necessary adjustments to the current or default settings in the Equation Status, Solver Parameters and Differencing Schemes tabs. 3. In the “Additional Scalars” panel, select a differencing scheme and under-relaxation parameter for the CAV scalar and for scalars other than VOF (the differencing scheme for the latter is set in the free-surface “Controls” panel; however, you may want to adjust its under-relaxation parameter). 4. Select the Output Controls sub-folder and use the “Monitor Numeric Behaviour” panel to print additional information such as convergence residuals and conservation checks (optional). 5. Select the Monitor Engineering Behavior sub-folder and, if you wish, use the “Monitor Cell Behaviour” panel to save selected cell data for subsequent plotting against iteration or time step number. 6. The Analysis Output sub-folder enables you to specify the frequency of saving solution results in the .ccm file. 7. Use the “Switches and Real Constants” panel in the Other Controls sub-folder to set switches and constants for any ‘non standard’ practices. Please check carefully the meaning of each switch and constant and use it only when absolutely necessary. Step 8: Define the time step size and run duration There are two ways to define the time step size and the run time length:

11-10

•

Go to the Analysis Preparation/Running folder and open the “Set Run Time Controls” panel. Fill in appropriate values for simulation time and time step size in the relevant boxes. This is the recommended way of defining time steps.

•

If your application involves a moving mesh defined by an events file, you will need to use the Advanced Transients panel by choosing Modules > Transient from the main pro-STAR window. in this panel, you can define load steps, each of which contains the time step size and number of time steps to be used for each load step. Please note that you need to go through Step 1 to Step 7 before defining load steps using the Advanced Transients panel.

Version 4.02

Chapter 12

ROTATING AND MOVING MESHES Rotating Reference Frames

Chapter 12

ROTATING AND MOVING MESHES The theory behind rotating and moving mesh problems and the manner of implementing it in STAR-CD is given in the Methodology volume, Chapter 13. The present chapter contains an outline of the process to be followed when setting up a rotating or moving mesh simulation, including details of the user input required and important points to bear in mind when setting up problems of this kind.

Rotating Reference Frames Models for a single rotating reference frame Step 1 Go to the “Select Analysis Features” STAR GUIde panel and select option On from the Rotating Reference Frame Status pop-up menu. This activates an additional folder in the NavCenter tree called Rotating Reference Frames. Step 2 In the above folder, open the “Rotating Reference Frames” panel and select option Single Frame. This enables you to define spin parameters for the material in your model. The required parameters are angular velocity and a local coordinate system whose Z-axis defines the axis of rotation, see Figure 12-1.

ω = 200 rpm

Figure 12-1

Solid body rotation

Useful points on single rotating frame problems 1. The angular velocity can vary with time, with the variation specified in (a) user subroutine UOMEGA, or (b) a user-defined table, or (c) by giving it a different value at each load step of a transient run (see Chapter 5, “Load-step based solution mode”). 2. The boundaries of the rotating domain are also assumed to be rotating. To model stationary walls, it is necessary to specify an equal and opposite spin velocity in the Omega text box of the Boundary Region dialog for walls (see the STAR GUIde “Wall” Help topic). Similarly, to model axial inflow, it is necessary to specify a spin velocity in the dialog for Inlet regions. 3. When a stagnation boundary condition is used, an option is provided to specify whether the direction cosines are based on relative or absolute velocities. Stagnation quantities are also defined using either relative or Version 4.02

12-1

ROTATING AND MOVING MESHES

Chapter 12

Rotating Reference Frames

absolute velocities. 4. When turbulence is specified as an intensity (at inlet or pressure boundaries), the turbulence kinetic energy is computed on the basis of static coordinate frame velocities. For stagnation boundaries, the specified intensity uses the same velocity as the stagnation quantities. 5. Boundary velocities are computed in the local rotating coordinate system. This is important in interpreting the information passed to the user subroutines. 6. When post processing results, you may view velocities in either the relative or the absolute reference frame (see the “Coord System tab”, located in the “Load Data” STAR GUIde panel). Models for multiple rotating reference frames (implicit treatment) Step 1 Go to the “Select Analysis Features” STAR GUIde panel and select option On from the Rotating Reference Frame Status pop-up menu. This activates an additional folder in the NavCenter tree called Rotating Reference Frames. Step 2 •

•

Decide how many reference frames are required to model the problem adequately. For example, the two-dimensional mixer problem shown in Figure 12-2 requires two rotating frames. Generate the mesh.

Baffle r = 15 cm

Sub-domain 2 Spin index = 2

ω = 0 rpm

Sub-domain 1 Spin index = 1

r = 10 cm ω = 500 rpm

r = 5 cm

Figure 12-2

12-2

Baffle

Multiple rotating frame illustration

Version 4.02

Chapter 12


Step 3 Display the Cell Table Editor by clicking the CTAB button on the main pro-STAR window. Define cell index numbers to correspond to each of the rotating mesh blocks (sub-domains) (see “The Cell Table” on page 3-1). Assign different spin and colour table indices to each cell type, as shown below, for the two rotating sub-domains of Figure 12-2. Note that the table entries for both sub-domains have the same material property reference number since the sub-domains belong to the same fluid domain. Sub-domain 1

Sub-domain 2

Version 4.02

12-3


Chapter 12


Step 4 Assign all cells within a sub-domain in turn to each of the cell types created above (see “Cell indexing” on page 3-3). Step 5 In the Rotating Reference Frames folder, open the “Rotating Reference Frames” panel and select option Multiple Frames - Implicit. This enables you to specify spin parameters (angular velocities and axes of rotation) for each of the spin indices already defined. In terms of the example of Figure 12-2, zero rotational speed needs to be assigned explicitly to sub-domain no. 2 since its local coordinate system is used in transforming velocities across the sub-domain interface. Useful points on multiple implicit rotating frame problems 1. When modelling multiple rotating reference frame (m.r.f.) problems, it is advisable to check the results carefully and see if they are reasonable and within the limitations of this approach. If this is not the case, one may need to resort to moving mesh methods, such as those described in the section on “Regular sliding interfaces”. Note, however, that a result obtained via the m.r.f. method can always be used as an initial field for a transient moving mesh simulation. This will reduce the time needed to reach a periodic state solution. 2. It is important to ensure that the interface between the different m.r.f. sub-domains is a smooth surface (i.e. a constant-radius surface). This point needs particular attention in all-tetrahedral mesh cases. 3. An angular velocity can vary with time, with the variation specified in (a) user subroutine UOMEGA, or (b) a user-defined table, or (c) by giving it a different value at each load step of a transient run (see Chapter 5, “Load-step based solution mode”). 4. The boundaries of a rotating domain are also assumed to be rotating. To model stationary walls, it is necessary to specify an equal and opposite spin velocity in the Omega text box of the Boundary Region dialog for walls (see the STAR GUIde “Wall” Help topic). Similarly, to model axial inflow, it is necessary to specify a spin velocity in the dialog for inlets. 5. When a stagnation boundary condition is used, an option is provided to specify whether the direction cosines are based on relative or absolute velocities. Stagnation quantities are also defined using either relative or absolute velocities. 6. When turbulence is specified as an intensity (at inlet or pressure boundaries), the turbulence kinetic energy is computed on the basis of static coordinate frame velocities. For stagnation boundaries, the specified intensity uses the same velocity as the stagnation quantities. 7. In cases where the mesh structure changes across the interface between two sub-domains (for example, between two axial turbomachinery stages, with the blades swept in opposite directions): (a) Build each sub-domain separately with its own ‘best fit’ mesh structure, and cell types with different spin indices 12-4

Version 4.02

Chapter 12


(b) Create a continuous mesh by coupling together the cells layers on either side of the interface using the Couple tool (Create Couples option; see also Chapter 3, “Couple creation” in the Meshing User Guide). (c) Use the Couple tool’s Couple Transform option to replace the coupled cells with polyhedral cells that have a one-to-one cell face correspondence at the interface. 8. Boundary velocities are computed in the local rotating coordinate system. This is important in interpreting the information passed to the user subroutines. 9. When post processing results, you may view velocities in either the relative or the absolute reference frame (see the “Coord System tab”, located in the “Load Data” STAR GUIde panel). 10. The present version of STAR-CD does not support the use of rothalpy (see “Rothalpy” on page 1-5 of the Methodology volume) in combination with the implicit solution technique. Models for multiple rotating reference frames (explicit treatment) Step 1 Go to the “Select Analysis Features” STAR GUIde panel and select option On from the Rotating Reference Frame Status pop-up menu. This activates an additional folder in the NavCenter tree called Rotating Reference Frames. Step 2 • •

Version 4.02

Decide how many rotating frames of reference are required to model the problem adequately, and the locations of the interfaces. Generate the mesh. The interface between adjacent rotating mesh blocks is defined by pairs of adjacent (but spatially coincident) boundaries, as shown in Figure 12-3. The coincident boundaries are first defined as independent boundary regions using separate sets of vertices and then coupled together as described in Step 7 below. Note that the interface must be either a plane perpendicular to the axis of rotation or a conical section, i.e. a surface generated by rotating a straight line around that axis.

12-5


Chapter 12

Rotating Reference Frames Boundary Regions no. 6 no. 5 no. 7 (pressure) (inlet) (pressure)

circumferential direction 4

ω = 100 rpm

36

3

35

2

34

1

IMAT = 1

33

8 12 7 11 6 10 5 9

cell number

40

ω = 500 rpm

64

39

63

38

62 IMAT = 2

37

61

no. 8 (inlet)

16 20 15 19 14 18 13 17

68

ω = 1000 rpm

100

67

99

66

98

65

IMAT = 3

97

boundary number

(a)

1134 134

33

135 1034

34

1135

1035

37

35

(b) Figure 12-3

Coupled boundary illustration

Step 3 Display the Cell Table Editor by clicking CTAB on the main pro-STAR window. Define cell index numbers to correspond to each of the rotating domains (see “The Cell Table” on page 3-1). Assign different material property and colour table indices to each cell type but ignore the spin index. In the above example, cell and material indices 1, 2 and 3 are defined to correspond to each domain. Step 4 Assign all cells within a domain in turn to each of the cell types created above (see “Cell indexing” on page 3-3). Also ensure that separate monitoring cell and reference pressure locations are specified for each domain. Step 5 Go to panel “Create Boundaries” in STAR GUIde, open tab “Regions” and use its facilities to create separate boundary regions at either side of each interface between domains, as shown in Figure 12-3. Step 6 Specify boundary conditions for both sides of an interface using panel “Define Boundary Regions” (only inlet and pressure boundary types are allowed). Example dialog boxes for boundary regions 5 and 6, making up the first interface in the above example, are shown below:

12-6

Version 4.02

Chapter 12


Step 7 Go back to panel “Create Boundaries” and use tab “Couples” to join the interface Version 4.02

12-7


Chapter 12


boundaries together. In doing so, you also need to: 1. Specify whether to join individual boundaries from each region on a one-to-one basis, or to couple the two regions to each other as a whole. If the latter is chosen, the value to be imposed on the couple’s pressure boundary is found by an averaging process. For example, the average of the values assigned to boundary region no. 5 in Figure 12-3 is 8

P region 5

8

⎛ ⎞ ⎛ ⎞ = ⎜ ∑ p i s i⎟ ⁄ ⎜ ∑ s i⎟ ⎝i = 5 ⎠ ⎝i = 5 ⎠

(12-1)

where p is the pressure and s the area of each boundary face. 2. If necessary, place region couples (as defined above) into separate groups. This enables you to identify boundary faces across which mass must be conserved and is only necessary in problems that have only inlet boundary couples. Such domains are recommended for solving closed loop problems where the flow rate needs to be determined as part of the solution. The groups to balance are specified in the “Rotating Reference Frames” panel (see Step 8 below). Step 8 In the Rotating Reference Frames folder, open the “Rotating Reference Frames” panel and select either option “Multiple Frames - Explicit” or option “Multiple Frames - NR-Explicit”. This enables you to specify: 1. Spin parameters (angular velocities and axes of rotation) for each of the mesh domains already defined. In the above example, domains 1, 2 and 3 have angular velocities of 100, 500 and 1000 r.p.m., respectively. The spin axis is normally common to all domains. 2. Control parameters required by the explicit solution algorithm and, if required, the coupled region groups mentioned in Step 7 above. Useful points on multiple explicit rotating frame problems 1. When modelling multiple rotating reference frame (m.r.f.) problems, it is advisable to check the results carefully and see if they are reasonable and within the limitations of this approach. If this is not the case, one may need to resort to moving mesh methods, such as those described in the section on “Regular sliding interfaces”. Note, however, that a result obtained via the m.r.f. method can always be used as an initial field for a transient moving mesh simulation. This will reduce the time needed to reach a periodic state solution. 2. It is important to ensure that the interface between the different m.r.f. domains is a smooth surface (i.e. a constant-radius surface). This point needs particular attention in all-tetrahedral mesh cases. 3. An angular velocity can vary with time, with the variation specified in (a) user subroutine UOMEGA, or (b) a user-defined table, or (c) by giving it a different value at each load step of a transient run (see 12-8

Version 4.02

Chapter 12

ROTATING AND MOVING MESHES Moving Meshes

Chapter 5, “Load-step based solution mode”). 4. The boundaries of a rotating domain are also assumed to be rotating. To model stationary walls, it is necessary to specify an equal and opposite spin velocity in the Omega text box of the Boundary Region dialog for walls (see the STAR GUIde “Wall” Help topic). Similarly, to model axial inflow, it is necessary to specify a spin velocity in the dialog for inlets. 5. When a stagnation boundary condition is used, an option is provided to specify whether the direction cosines are based on relative or absolute velocities. Stagnation quantities are also defined using either relative or absolute velocities. 6. When turbulence is specified as an intensity (at inlet or pressure boundaries), the turbulence kinetic energy is computed on the basis of static coordinate frame velocities. For stagnation boundaries, the specified intensity uses the same velocity as the stagnation quantities. 7. Interfaces between differentially-rotating mesh domains are best placed at positions that do not lie inside recirculating flow fields. 8. Caution should be exercised when using this approach because of the explicit coupling at the special boundaries. The method is most suitable for problems involving strong outflow across the coupled interface. 9. The NR-Explicit option should be chosen over the Explicit option for configurations where the turbomachinery blades are closely packed and/or if a shock wave is expected to hit either of the two coupled boundaries at the interface. 10. Boundary velocities are computed in the local rotating coordinate system. This is important in interpreting the information passed to the user subroutines. 11. When post processing results, you may view velocities in either the relative or the absolute reference frame (see the “Coord System tab”, located in the “Load Data” STAR GUIde panel).

Moving Meshes Basic concepts The moving mesh feature is activated by command MVGRID. Changes in mesh geometry can be specified either by pro-STAR commands (i.e. the Change Grid operation in the EVENTS command module), or by user coding included in subroutine NEWXYZ. In this subroutine, the user can vary the geometry of a model by defining vertex coordinates as a function of time. The deformed coordinates are written to the transient post data (.pstt) file and can be loaded and plotted during post-processing. As an alternative, the Change Grid (CG) operation can be used to alter the vertex positions with time. Its distinguishing features are as follows: •

• Version 4.02

The operation is initiated at an ‘event step’ specified by the user and remains active at all subsequent time steps, until the CG operation is explicitly turned off by a termination event, or a new set of CG commands are provided as part of another event step. The main body of the operation consists of a set of pro-STAR commands that 12-9


Chapter 12

Moving Meshes

are used while STAR is running (as part of a STAR/pro-STAR interaction process). • The above commands utilise a set of both program-defined and user-defined parameters that can store anything that is of relevance to the problem description. The parameters used by the CG command set are: 1. Program-defined (a) (b) (c) (d) (e) (f) (g)

ITER — current time step number TIME — current solution time LSTP — current load step (see Chapter 5, “Load step definition”) EVEX — last executed event number EVNO — event number to be executed next ETIM — time at which the next event is scheduled YPST — piston position; a special parameter for piston engine problems, calculated on the basis of other parameters supplied by command EVPARM (see “Setting up models” on page 12-15). 2. User-defined These are specified by the user in subroutine UPARM to provide additional parameters. They are of two kinds: (a) Integer parameters in the range 0-999 (b) Real parameters in the range 0-999 Note that pro-STAR restricts the number of active parameters to 99. The CG operation uses all the standard pro-STAR facilities and is therefore more flexible and powerful for mesh geometry changes than user coding supplied in subroutine NEWXYZ. Note that STAR-CD also provides other special operations related to moving meshes, as follows: • • •

Cell removal/addition — (see “Cell-layer Removal/Addition” on page 12-14) Sliding mesh — (see “Sliding Meshes” on page 12-18) Conditional cell attachment and change of fluid type — (see “Cell Attachment and Change of Fluid Type” on page 12-22)

Setting up models The main steps for setting up a moving mesh model are outlined below. For more detailed information, refer to Tutorial 11 in the Tutorials volume. Step 1 Generate the mesh at time t = 0 and issue the following command: TIME,TRANS

(turn on the transient solution option)

followed by either MVGRID,ON

(turn on the moving-grid option, when using subroutine NEWXYZ only)

or

12-10

Version 4.02

Chapter 12


MVGRID,ON,EVENT,PROSTAR (turn on the moving-grid option, when using the EVENTS command module) Step 2 (Skip this step only if mesh changes are input through the user subroutine NEWXYZ) Define an event step data file, e.g. EVFILE,INITIAL,case.evn (initialise the events file) EVSTEP,1,TIME,0.0 (define an event) EGRID,READ,case.cgrd (get the description of mesh operations from file case.cgrd, in coded form) EVSAVE,1 (save this information as event no. 1) The contents of file case.cgrd mentioned above for the problem shown in Figure 12-4 are as follows: ! Comments like this are allowed by starting the line with “!” VSET,NONE VSET,ADD,VRANGE,1,2,1 *SET,YBOT,TIME VMOD,VSET,F,YBOT

VFILL,1,11,4,3,2,2,1

Y

(clear the vertex set) (add vertices 1 and 2 to the set) (set parameter YBOT equal to the current time) (change the y-coordinates of the vertex set so that they follow the bottom boundary movement) (re-position the mesh vertices between the two boundaries)

11

12

9

10

7

8

5

6

3

4 1 m/s

X

Figure 12-4 Version 4.02

1

2

Moving mesh illustration 12-11


Chapter 12

Moving Meshes

Note that: 1. An event step can be (a) deleted, if necessary, with command EVDELETE and remaining event steps re-numbered via command EVCOMPRESS; (b) modified with command EVGET; (c) listed on the screen with command EVLIST. 2. Command EVUNDELETE restores a previously deleted event step. 3. User-specified offsets can be applied to the actual event time via command EVOFFSET. Step 3 • • •

If using the method described in Chapter 5, “Load-step based solution mode”, define the load step for the transient run. Check the validity of specified events and prepare the events data file for subsequent use via command EVPREP. Save the problem’s data files using commands GEOMWRITE, PROBLEMWRITE, etc. or their equivalent GUI operations accessible from the File menu.

Note that the events data file can be • •

written in coded form to a (.evnc) file with command EVWRITE, typically in order to transfer data to another computer read in coded form from a (.evnc) file with command EVREAD, typically when transferring data from another computer

Step 4 Exit from pro-STAR and then run STAR from your session’s X-window, as described in Chapter 2, “Running a STAR-CD Analysis”, Step 6. Step 5 Post-process the data. For example, the commands needed to process time step no. 10 are: SUBTITLE Results at time step 10 Velocity field EVFI CONN case.evn (connect the event file) TRLOAD case.pstt (load the transient post data file) STORE ITER 10 (the appropriate events are loaded and executed automatically) GETC ALL (get the cell data) POPT VECT PLTY NORM CSET NEWS FLUID CPLOT QUIT,NOSAVE 12-12

Version 4.02

Chapter 12


Be very careful not to save problem information to file case.mdl as the current geometry corresponds to the state of the mesh at time step no. 10. Useful points 1. STAR can be run in ‘mesh preview’ mode only, which is very useful for checking out the mesh set-up. To do this, a hidden switch has to be set up in pro-STAR as follows: RCONSTANT, 4, 1.

(set constant number 4 to 1.)

The message “MESH PREVIEW RUN” should appear both on the screen and in the run-time output (.run) file when running STAR. Note that this facility is not available for parallel runs. 2. Moving grid events normally describe a continuous motion and will therefore remain operational throughout the run. If, however, the grid motion needs to be stopped for whatever reason, this can be done via a termination event as follows: EGRID,NONE 3. The transient post data (.pstt) file is usually very large, so care must be taken when specifying the post data output frequency. If the analysis is split into several stages, it is also advisable to give the .pstt file produced at the end of each stage a unique filename. This helps to spread the output produced amongst several files and thus ease the data management and manipulation processes. 4. Porous media should not be used in areas of the mesh where there is relative internal movement (i.e. cell expansion or contraction). 5. You are strongly advised to set the pressure correction under-relaxation factor to a value less than 1.0 (e.g. 0.8) before starting the analysis. 6. Flow boundary conditions on boundaries that have moving vertices may result in mass flux into / out of the domain, caused by the displacements of the boundaries. 7. The only valid option for restart runs is Standard Restart (see the “Analysis (Re)Start” panel in STAR GUIde. Automatic Event Generation for Moving Piston Problems pro-STAR provides a special command, MMPISTON, which may be used in an engine model to automatically generate the moving part of the piston mesh, the Change Grid (.cgrd) command file and the event (.evn) file. More specifically, the moving mesh commands accomplish the following: Starting from a basic mesh, they create the cells, vertices, boundaries, events and moving grid commands to completely specify the mesh motion for the STAR solver. These commands are designed to be used in sequence, so that all entities are created at the end of the current model and do not compromise earlier events. As the output is standard pro-STAR events and EGRID commands, an advanced user can readily modify these to suit specific problems.

Version 4.02

12-13


Chapter 12

Cell-layer Removal/Addition

Cell-layer Removal/Addition Basic concepts A cell is removed by collapsing intervening faces between two opposite sides in a given direction. This is done by moving together the vertices making up the faces. Cells can be collapsed at the beginning of a given time step or prior to the start of the calculations. The latter case is treated as a special mesh set-up operation and does not affect the solution in any way. Normally, entire layers of cells are removed at a given event step. However, it is also possible to remove part of a layer, in which case cells at the edge of the retained section collapse into prisms. A cell layer (or partial layer) has the following properties: • • • • • • •

It is defined as a group of cells that is one cell thick in the collapsing direction. The faces which collapse must be quadrilaterals, but those forming the upper and lower surfaces of the layer may be quadrilateral or triangular. The collapsing cell faces on the outer perimeter of the group form boundaries. Either the upper or lower surface of a layer may coincide in whole or part with a boundary, but not both surfaces simultaneously. No more than one layer may be removed at each event step. The layer must not be composed of tetrahedral cells. Trimmed (polyhedral) cells can only be collapsed if they have been formed by extruding another cell in the direction of collapse.

The reverse operation, adding a cell layer, is achieved by expanding the removed layer in the direction it was collapsed. This means that layers to be added must have been removed first. Thus, all restrictions on cell removal also apply to cell addition so that: • • • •

Only one entire layer (or partial layer) may be restored at each event step. When cells are restored, they reappear next to the neighbours they had at the time of their collapse. If any of their faces were boundaries, those boundaries are also restored. Cell layers must be restored in the reverse order in which they were removed.

The cells to be removed or added, and the time at which to do this (i.e. event step and event time) are specified in the EVENTS command module. A cell removal or addition event is executed when the current simulation time equals the time specified by the event step, within a given tolerance. Note that cell removal or addition changes only the cell connectivity within the mesh. The actual change of mesh geometry has to be specified explicitly through a moving mesh operation of the kind described in “Moving Meshes” on page 12-9. In the event of cell removal, the user has to ensure that: • •

12-14

The mesh geometry changes in a way that reflects the fact that cells have been removed. Cells remain collapsed until they are restored. This means that vertices belonging to the removed cells must move with the moving boundary for all subsequent time steps.

Version 4.02

Chapter 12

ROTATING AND MOVING MESHES Cell-layer Removal/Addition

Setting up models Cell Removal or Addition operations should always be combined with either • •

Change Grid operations in the EVENTS command module, or the user subroutine NEWXYZ.

The main steps for setting up a model of this kind are outlined below. Step 1 Generate the mesh at time t = 0. The layers to be removed can be given different cell index numbers using command CTABLE. .

7

19

20

21

6

16

17

18

5

13

14

15

Cell index

Cell number

Y (2)

4

10

11

12

3

7

8

9

2

4

5

6

1

1

2

3

X (1)

Figure 12-5

Cell layer removal illustration

Referring to the example of Figure 12-5 the relevant commands would be: CTAB,1,Fluid RP7,1 *SET,CTY,1,1 *SET,C1,1,3 *SET,C2,3,3 *DEFINE CTYPE,CTY CSET,NEWS,CRANG,C1,C2,1 CMOD,CSET *END *LOOP,1,6,1 Step 2 Issue the following commands: TIME,TRANS Version 4.02

(turn on the transient solution option) 12-15


Chapter 12

Cell-layer Removal/Addition

MVGRID,ON,EVENT,PROSTAR (turn on the moving-grid option) Step 3 •

Define an event step data file, e.g. EVFILE,INITIAL,case.evn (initialise the events data file)

•

Turn on the Change Grid operation at time t = 0 EVSTEP,1,TIME,0.0 EGRID,READ,case.cgrd

(get the description of mesh operations from file case.cgrd, in coded form) (save this information as event no. 1)

EVSAVE,1 •

Specify cell layer removal via the cell type EVSTEP,2,TIME,0.05 EDDIR,LOCAL,1,2 EDCELL,ADD,CTYPE,1 ECLIST,DEACTIVATED,ALL EVSAVE,2

•

(remove cells in direction no. 2 in the local coordinate system) (remove cells with index no. 1) (list removed cells)

Specify cell layer removal via a cell range EVSTEP,3,TIME,0.08 EDDIR,LOCAL,1,2 EDCELL,ADD,CRAN,4,6,1 EVSAVE,3

•

Specify cell layer addition, assuming the last cell layer removed had index no. 2 EVSTEP,4,TIME,0.2 EACELL,ADD,CTYPE,2 ECLIST,ACTIVATED,ALL EVSAVE,4

(add all cells with index 2) (list added cells)

Note that: 1. The event time can also be specified using global parameters. For example EVPARM

12-16

PISTON 1000. ↑ ↑ piston rotating engine speed (rpm)

0.04 ↑ crank radius

0.13 0.015 ↑ ↑ length initial of con. piston rod position

COMP 0.1015 ↑ piston location at TDC

Version 4.02

Chapter 12

ROTATING AND MOVING MESHES Cell-layer Removal/Addition

EVSTEP

1 ↑ event step

PCOMP ↑ compression stage

0.02 ↑ piston position

2. An event step can be (a) deleted, if necessary, with command EVDELETE and remaining event steps re-numbered via command EVCOMPRESS; (b) modified with command EVGET; (c) listed on the screen with command EVLIST. 3. Command EVUNDELETE restores a previously deleted event step. 4. User-specified offsets can be applied to the actual event time via command EVOFFSET. Step 4 • • •

If using the method described in Chapter 5, “Load-step based solution mode”, define the load step for the transient run. Check the validity of specified events and prepare the events data file for subsequent use via command EVPREP. Save the problem's data files using commands GEOMWRITE, PROBLEMWRITE, etc. or their equivalent GUI operations accessible from the File menu.



Step 5 Exit from pro-STAR and then run STAR from your session’s X-window, as described in Chapter 2, “Running a STAR-CD Analysis”, Step 6. Step 6 Post-process the data. For example, the commands needed to process time step no. 10 are: SUBTITLE Results at time step 10 Velocity field EVFI CONN case.evn (connect the event file) TRLOAD case.pstt (load the transient post data file) STORE ITER 10 (the appropriate events are loaded and executed automatically) GETC ALL (get the cell data)

Version 4.02

12-17


Chapter 12

Sliding Meshes

POPT VECT PLTY NORM CSET NEWS FLUID CPLOT QUIT,NOSAVE Be very careful not to save problem information to file case.mdl as the current geometry corresponds to the state of the mesh at time step no. 10. Useful points 1. You are strongly advised to identify cell layers intended for removal/addition by assigning a unique cell index to each of them. 2. Cell layers can be removed at negative event times. This is useful, for example, in reciprocating piston engine models where simulation starts with the piston at top dead centre. In such cases the previously removed cell layers can thus be added at positive event times. 3. You are advised to first run the model in ‘mesh preview’ mode in order to check whether the intended cell removal/addition and mesh movement are carried out correctly. This can be done by issuing the following command in pro-STAR: RCONSTANT, 4, 1.

4.

5.

6. 7.

(set constant number 4 to 1)

The message “MESH PREVIEW RUN” should appear both on the screen and in the run-time output (.run) file when running STAR. It is very important to ensure that the locations chosen for reference pressure and field variable monitoring (via commands PRESSURE and MONITOR, respectively) correspond to cells that will never be removed. If the simulation includes combustion modelling and the definition of ignition regions (see Chapter 8, “Setting Up Chemical Reaction Schemes”, Step 5), make sure that no cells corresponding to these regions have been removed during the time that ignition takes place. You are strongly advised to set the pressure correction under-relaxation factor to a value less than 1.0 (e.g. 0.8) before starting the analysis. For STAR-HPC runs, you need to ensure that the removed cell layers do not collapse towards the inter-processor boundaries. In another words, the removed cell layers and the inter-processor boundaries should always be perpendicular to each other. This can be achieved through manual decomposition.

Sliding Meshes Regular sliding interfaces One way of implementing sliding meshes is the regular sliding interface method. This enables the interface cells to progressively change their connectivity during the solution. The change of cell connectivity is activated through a ‘cell attachment’ operation. Cell pairs to be attached and the time of attachment (i.e. event step and 12-18

Version 4.02

Chapter 12

ROTATING AND MOVING MESHES Sliding Meshes

event time) are specified by the user in the EVENTS command module. The cell attachment event is executed when the current simulation time equals the time specified by the event step within a given tolerance. Setting up models The regular sliding interface method combines both the Cell Attachment and the Change Grid operation in the EVENTS command module. The main steps for setting up a case are outlined below. Step 1 •

Generate the mesh at time t = 0. The sliding interface is defined as two coincident boundaries, one for the stationary and one for the moving part of the mesh. Thus, two sets of coincident vertices must be defined at that location. The two coincident boundaries have to be defined as different boundary regions and declared as attachment boundaries using the RDEFINE command: (define boundary region no.1 as an attachment boundary)

RDEF,1,ATTACH

1 ↑ local coordinate system

0 ↑ alternate wall system (see “Cell Attachment and Change of Fluid Type” on page 12-22 for an explanation of this parameter)

RDEF,2,ATTACH 1,0 •

Issue the following commands: TIME,TRANS (turn on the transient solution option) MVGRID,ON,EVENT,PROSTAR (turn on the moving-grid option)

Step 2 •

Define an event step data file. EVFILE,INITIAL,case.evn (initialise the events data file)

•

Perform an initial attachment operation for the relevant boundary pairs (otherwise they will be treated as detached). EVSTEP,1,TIME,0.0

(event step 1 occurs at time t = 0.0)

followed by either EAMATCH,1,2 Version 4.02

(match regions 1 and 2) 12-19


Chapter 12

Sliding Meshes

or EATTACH,6,1 RP5,1,1 EALIST,ALL •

Turn on the Change Grid operation at time t = 0. EGRID,READ,case.cgrd EVSAVE,1

•

(attached boundaries 6 and 1) (attach the rest of the boundary pairs) (list out all attached boundary pairs)

(get the description of mesh operations from file case.cgrd, in coded form) (save this information as event no. 1)

Specify subsequent attachment operations, e.g. EVSTEP,2,TIME,0.02 EATTACH,6,2 EAGENERATE,4,1,1,1,1

EATTACH,10,1 EVSAVE 2

(attach boundaries 6 and 2) (EAGENERATE works similarly to CGENERATE, see “Command-driven facilities” on page 2-44 of the Meshing User Guide) (attached boundaries 10 and 1) (save event no. 2)

Note that: 1. The attached boundary set definitions in an event step can be (a) deleted, if necessary, with command EADELETE and remaining definitions re-numbered via command EACOMPRESS; (b) listed on the screen with command EALIST. 2. An event step can be (a) deleted, if necessary, with command EVDELETE and the remaining event steps renumbered via command EVCOMPRESS; (b) modified with command EVGET; (c) listed on the screen with command EVLIST. 3. Command EVUNDELETE restores a previously deleted event step. 4. User-specified offsets can be applied to the actual event time via command EVOFFSET. Step 3 • • •

12-20

If using the method described in Chapter 5, “Load-step based solution mode”, define the load step for the transient run. Check the validity of specified events and prepare the events data file for subsequent use via command EVPREP. Save the problem’s data files using commands GEOMWRITE, PROBLEMWRITE, etc. or their equivalent GUI operations accessible from the File menu. Version 4.02

Chapter 12

ROTATING AND MOVING MESHES Sliding Meshes



Step 4 Exit from pro-STAR and then run STAR from your session’s X-window, as described in Chapter 2, “Running a STAR-CD Analysis”, Step 6. Step 5 Post-processing the data. For example, the commands needed to process time step no. 10 are: SUBTITLE Results at time step 10 Velocity field EVFI CONN case.evn (connect the event file) TRLOAD case.pstt (load the transient post data file) STORE ITER 10 (the appropriate events are loaded and executed automatically) GETC ALL (get the cell data) POPT VECT PLTY NORM CSET NEWS FLUID CPLOT QUIT,NOSAVE Be very careful not to save problem information to file case.mdl as the current geometry corresponds to the state of the mesh at time step no. 10. Useful points 1. At time t = 0, cell pairs are detached. They become attached only when an event containing EATTACH or EAMATCH commands is executed. Once attached in this way, they remain attached until another EATTACH or EDETACH command references them, or they are deactivated. 2. When the model’s mesh is being created, it is very useful to set up a regular boundary numbering scheme at the interface, because this simplifies the specification of cell attachment. 3. At the initial stages of the analysis, the solution can be accelerated by using pure sliding only (i.e. without shearing), which in turn allows larger time steps. In terms of Figure 15-1 in Chapter 15 of the Methodology volume, this is equivalent to going from Stage 1 to Stage 4 in a single time step. If this is the case, the time step dt should be made equal to dtsl, where, for cylindrical systems dtsl = cell face angle at interface / rotating speed

Version 4.02

(12-2)

12-21


Chapter 12

Cell Attachment and Change of Fluid Type

In general, the time step dt should equal dtsl divided by an integer. If accuracy is not at a premium, one may also slide the mesh by more than one cell width (e.g. two cell widths) in a single time step. 4. In cylindrical systems, periodic results are usually reached after about seven revolutions. 5. The transient post data (.pstt) file is usually very large, so care must be taken when defining the output frequency of post-processing data. If the analysis is split into several stages, it is also advisable to give the .pstt file produced at the end of each stage a unique filename. This helps to spread the output produced amongst several files and thus ease the data management and manipulation processes. 6. It is advisable to first run the model in ‘mesh preview’ mode in order to check whether the intended cell sliding and mesh movement are carried out correctly. This can be done by issuing the following command in pro-STAR: RCONSTANT,4,1.

7. 8.

9. 10. 11.

(set constant number 4 to 1.)

The message “MESH PREVIEW RUN” should appear both on the screen and in the run-time output (.run) file when running STAR. EATTACH commands are allowed only between active cells. All boundaries belonging to a given region must couple only to boundaries belonging to a (different) unique region. For example, it is illegal for some boundaries from region 1 to couple to boundaries from region 2, while other boundaries from region 1 couple to boundaries from region 3. If one cell of an attached pair is deactivated, the other side reverts to the alternate wall region. If both cells of an attached pair are deactivated simultaneously and then reactivated, the EATTACH command must be re-issued. For STAR-HPC runs, you need to ensure that the sliding part of the mesh resides completely on one processor. This can also be achieved through manual decomposition.

Cell Attachment and Change of Fluid Type Basic concepts Cell attachment permits the following situations to be modelled: 1. The connection of unconnected neighbouring cells in different fluid domains, say on the basis of local flow conditions. This can be used, for example, to model leaf valves which pop open when the pressure difference across them exceeds a given value. 2. The complete disconnection of neighbouring cells. This situation necessitates two kinds of operation: (a) A ‘Cell Attachment/Detachment’ operation. (b) A ‘Change Fluid Type’ operation. The latter enables a fluid domain to become completely cut off from the rest of the flow field. Once cut off, the flow solution in such a domain can have its own reference pressure and temperature. A special type of boundary (‘Attachment’ type) 12-22

Version 4.02

Chapter 12

ROTATING AND MOVING MESHES Cell Attachment and Change of Fluid Type

must also be declared at the interface where cell attachment and detachment is to take place. STAR performs a cell detachment by connecting the detached cells to an appropriate wall or inlet region. Cell attachment/detachment operations are specified in the EVENTS command module. The connection/disconnection event is initiated when the current simulation time equals the time specified by the event step within a given tolerance. The same also applies to the ‘Change Fluid Type’ operation. However, when the designated time for connecting cells is reached, the operation may not necessarily be carried out immediately. Instead, the precise connection/disconnection time is determined by the flow solution. All conditions defined for a particular event are maintained in the next event unless disabled explicitly. Thus, once a boundary pair is attached, it remains attached until it is explicitly detached. Setting up models The main steps for setting up a cell attachment and change of fluid type case are outlined below. Step 1 •

Generate the mesh at time t = 0. This requires a boundary interface to be set up separating the (presently or potentially) different fluid domains. The interface is defined as two coincident boundaries made up of two sets of coincident vertices. The two boundaries must be first specified as different boundary regions and then declared as attachment boundaries (see Figure 12-6) using command RDEFINE: (define boundary region no. 1 as an attachment boundary)

RDEF,1,ATTACH

1

8

↑ local coordinate system

↑ alternate wall or inlet region

RDEF,8,inlet

(boundary region no. 8 is a dummy region)

(could also be of type wall)

The alternate wall or inlet region is specified in order to enable the code to assign appropriate (wall or inlet) properties to the attachment boundaries, if they happen to be detached. RDEF,2,ATTACH 1,8 RDEF,3,ATTACH 1,8 RDEF,4,ATTACH 1,8 • Version 4.02

Issue the following commands: 12-23


Chapter 12


TIME,TRANS (turn on the transient solution option) MVGRID,ON,EVENT,PROSTAR (turn on the moving-grid option)

IMAT = 2

2 1

3 1

IMAT = 3

4 2

4 3

7 5

8 6

1, 2, 3, 4, Boundary 5, 6, 7, 8 numbers

Cell numbers 151, 152 96, 97

1, 2, 3, 4

Y (2)

Boundary region numbers

Cell numbers IMAT = 1: cells 1-100 IMAT = 2: cells 101-150 IMAT = 3: cells 151-200

IMAT = 1

X (1)

Figure 12-6

Outline of conditional cell attachment operation

Step 2 •

Assign a material property reference no. to each fluid domain using command CTABLE. For the model shown in Figure 12-6: For domain no. 1 (IMAT = 1) CTAB

1 ↑ cell index

FLUID ↑ cell type

3 ↑ colour index

CSET,NEWS,CRAN,1,100 CTYPE,1

0 ↑ porosity reference number

1 ↑ material property reference number

1 ↑ group number

(collect together all cells with property ref. no. 1) (change the currently active cell type to 1)

CMOD,CSET For domain no. 2 (IMAT = 2) CTAB,10,FLUID,4,0,2,2 12-24

Version 4.02

Chapter 12


CSET,NEWS,CRAN,101,150

(collect together all cells with property ref. no. 2)

CTYPE,10 CMOD,CSET For domain no. 3 (IMAT = 3) CTAB,20,FLUID,5,0,3,3 CSET,NEWS,CRAN,151,200

(collect together all cells with property ref. no. 3)

CTYPE,20 CMOD,CSET •

Define the monitoring cell and pressure reference for each material type using the MONITOR and PRESSURE commands: For domain no. 1 PMAT 1 MONI,20 PRES,1.0E05,10

(define the monitoring cell) (define the reference cell and reference pressure)

STATUS For domain no. 2 PMAT 2 MONI,120 PRES,1.0E05,110 STATUS For domain no. 3 PMAT 3 MONI,170 PRES,1.0E05,180 STATUS Step 3 •

Define an event step data file using the EVFILE command (see Figure 12-6): EVFILE,INITIAL,case.evn (initialise the event data file)

•

Perform an initial Attachment and Change Fluid operation for relevant boundary pairs (otherwise they will be treated as detached and the attachment boundary type will become equivalent to a wall). For example, to connect region nos. 1 and 2: EVSTEP,1,TIME,0.0 EAMATCH,1,2

Version 4.02

(event step no. 1 occurs at time t = 0) (connect regions 1 and 2) 12-25


Chapter 12


•

Change the fluid material property reference number in region 2 to that in region 1 EFLUID,1,ADD,CRANGE,101,150 (or EFLUID,1,ADD,CTYPE,10) (or EFLUID,1,ADD,GROUP,2)

•

List the latest definitions and save the information supplied ECLIST,CFLUID,ALL EVSAVE,1

•

(list all cells of type ‘Change Fluid’) (save this information as event no. 1)

If, at time t = 1., region no. 2 is to be cut off from the rest of the flow, issue the following commands: EVSTEP,2,TIME,1. EDETACH,ADD,REGION,1 (add region no. 1 to the ‘detach’ set) (or EDETACH,ADD,BRAN,1,2) EDLIST,ALL (list all detached boundary pairs) EFLUID,2,ADD,CTYPE,10 EVSAVE,2 Note that the detached boundary set definitions in an event step can be deleted, if necessary, with command EDDELETE and remaining definitions re-numbered via command EDCOMPRESS.

Step 4 • If it is to be assumed that the valve between boundary regions 3 and 4 opens when the average pressure in region 4 is greater than that in region 3, set up a conditional event as follows: EVCND,3 EAMATCH,3,4 EFLUID,1,ADD,CTYPE,20

(attach region nos. 3 and 4) (change all cells with cell id. 20 to fluid no. 1)

EVSAVE,3 •

Enable conditional attachment in an actual event EVSTEP,4,TIME,2. ECONDITIONAL,3,ENABLE EVSAVE,4

(enable conditional event no. 3)

Step 5 •

Define all other events required. Note that: 1. An event step can be (a) deleted, if necessary, with command EVDELETE; (b) modified with command EVGET;

12-26

Version 4.02

Chapter 12


(c) listed on the screen with command EVLIST. 2. Command EVUNDELETE restores a previously deleted event step. 3. User-specified offsets can be applied to the actual event time via command EVOFFSET. Step 6 • • •

If using the method described in Chapter 5, “Load-step based solution mode”, define the load step for the transient run. Check the validity of specified events and prepare the events data file for subsequent use via command EVPREP. Save the problem’s data files using commands GEOMWRITE, PROBLEMWRITE, etc. or their equivalent GUI operations accessible from the File menu.



Step 7 Exit from pro-STAR and then run STAR from your session’s X-window, as described in Chapter 2, “Running a STAR-CD Analysis”, Step 6. Step 8 Post-processing the data. For example, the commands needed to process time step no. 10 are: SUBTITLE Results at time step 10 Velocity field EVFI CONN case.evn (connect the event file) TRLOAD case.pstt (load the transient post data file) STORE ITER 10 (the appropriate events are loaded and executed automatically) GETC ALL (get the cell data) POPT VECT PLTY NORM CSET NEWS FLUID CPLOT QUIT,NOSAVE Be very careful not to save problem information to file case.mdl as the current geometry corresponds to the state of the mesh at time step no. 10. Useful points 1. At time t = 0, cell pairs are detached. They become attached only when an event containing EATTACH or EAMATCH commands is executed. Once Version 4.02

12-27


Chapter 12

Mesh Region Exclusion

attached in this way, they remain attached until another EATTACH or EDETACH command references them, or they are deactivated. 2. When the model’s mesh is being created, it is very useful to set up a regular boundary numbering scheme at the interface as this simplifies the specification of cell attachment.

Mesh Region Exclusion Basic concepts A group of cells can be excluded from the solution domain by defining an ‘exclude’ event and issuing command EECELL. Note that: •

• •

This is possible only if the cells in the group are not connected to any other cells in the model. Thus, the group must first be detached from the rest of the model using a cell detachment event, as described in the section on “Cell Attachment and Change of Fluid Type”. Only active cells can be excluded. There are no other restrictions on the cells that may be excluded (e.g. more than one adjacent layers may be removed at a time).

An important difference with respect to cell deactivation, discussed in the section on “Cell-layer Removal/Addition”, must also be noted. The mass contained in excluded cells is removed from the solution; by contrast, the mass in the deactivated cells is ‘squeezed out’ into the neighbouring cells.

Moving Mesh Pre- and Post-processing Introduction The various mesh motions and connectivity changes caused by the execution of event-type commands can be visualised and verified using special pro-STAR facilities. These help both in setting up the events (pre-processing) and in examining the results of the analysis (post-processing). The same facilities can also be used during the actual solution run, in combination with mesh changes caused by event execution. Note that mesh changes can be classified into • •

geometry changes connectivity changes

Geometry changes should occur only as a result of the EGRID event. All other events can only cause connectivity changes. Event processing is useful at three different stages of flow modelling and serves the following requirements: 1. Pre-processing Here the emphasis is on: (a) (b) (c) (d) (e) 12-28

Testing out different event combinations. Checking out commands read in by EGRID. Making corrections as needed and re-executing the events. Working with incomplete events. Testing out parts of events, e.g. to see if cells to be attached are adjacent Version 4.02

Chapter 12

ROTATING AND MOVING MESHES Moving Mesh Pre- and Post-processing

to each other. (f) Using events to generate future events, e.g. use EGRID commands to move the mesh and then EAMATCH to define the attach pairs. 2. Solution run Here, STAR calls up pro-STAR to alter the grid in some way. 3. Post Processing By this stage, the mesh geometry applicable to any given point in time is available from the actual solution. Therefore, the goal here is to generate vertex data for various flow variables (via command CAVERAGE), display them using the correct surface and edge plotting options and create particle tracks. Some error checking capabilities are also needed to detect event errors which may have previously gone unnoticed. These detected errors are highlighted in the plots. Action commands Commands EVLOAD and EVEXECUTE belong to this category. EVLOAD is used to ‘load’ all events up to a specified point in time. There are two basic components involved in this operation: •

•

Creation of internal tables defining the current status of each cell. These tables can then be used by command CSET via keywords ACTIVE, DEACTIVE or ATTACHED. For example, CSET,NEWSET,ACTIVE creates a cell set of the currently active cells. Execution of any grid-changing commands read in by EGRID.

Note that, in general, application of EVLOAD results only in changes to the mesh geometry and not to the mesh connectivity. The various options of the EVLOAD command deal with different ways of specifying the current time. There is also a ‘reset’ option which restores the geometry to the ‘original state’, as defined below. The first time EVLOAD is called, the ‘original state’, i.e. the vertex, cell and boundary definitions of the model, are saved. Command EVLOAD,RESET restores the model to this original state. If the model is changed at this point, the next EVLOAD command will create fresh ‘original state’ files that correspond to the changes. Command EVEXECUTE should be used only after a successful EVLOAD operation. This command applies the current status, stored in the internal tables mentioned above, to the mesh. Thus: •

• •

Cells marked as ‘deactivated’ are deleted (equivalent to command CDELETE) and vertex numbers on adjacent cells are changed to reflect their new connectivity. Cells marked as having changed material type are changed to a different cell type. Vertices on the common face between two cells marked as ‘attached’ will be merged.

The end result of the above is changes to cell connectivity due to cell removal. Using option OFF with command EVEXECUTE restores the model connectivity to the ‘original state’ defined by EVLOAD. The internal status tables also retain their Version 4.02

12-29


Chapter 12

Moving Mesh Pre- and Post-processing

original setting. A succeeding EVLOAD command also implicitly performs an EVEXECUTE,OFF operation. Status setting commands Commands EVFLAG, EVCHECK and PLATTACH belong to this category. EVFLAG and EVCHECK modify the behaviour of EVLOAD. Any subsequent plotting is controlled by the PLATTACH options. Command EVFLAG can be used to selectively turn on or off different types of events loaded by EVLOAD. It contains two groups of parameters that can be set independently, one for pre-processing and the other for post-processing. The option specified with command EVCHECK (PREP or POST) determines which of the two groups is to be set. The EVLOAD components that can be selectively turned on or off are: 1. COND — executes enabled conditional events 2. UPARM — calls user subroutine UPARM 3. GRID — processes grid change commands This option is essential if EVLOAD is to be used for changing the mesh geometry when pro-STAR is called by STAR. For example, suppose the following commands are read in by EGRID: ..... EVFLAG,PRE,OFF,GRID

EVLOAD,UPTO,TIME,TIME

(if the GRID flag is not set to OFF, the EVLOAD command that follows will cause EGRID commands to be executed repeatedly and ad infinitum) (Note the use of the predefined parameter TIME)

CSET NEWS ACTIVE VSET NEWS SURFACE VSMOOTH ..... 4. NEWXYZ — calls user subroutine NEWXYZ 5. DEACTIVE — checks that deactivated cells have zero volume. If they do not, the error is reported and EVLOAD is stopped. 6. ACTIVE — checks that active cells have non-zero volume. If they do not, the error is reported and EVLOAD is stopped. 7. ATTACH — checks that cell faces to be attached have coincident vertices. If they do not, the error is reported and EVEXECUTE is stopped. Note that this particular option only applies to EVEXECUTE. 8. NEWSET — creates a set of cells which fail any tests during EVLOAD. 9. SCDEF — creates scratch files containing the initial mesh state. This option may be turned off whenever there is no need to backtrack in time, for example when EVLOAD is called from STAR. This saves CPU time and disk space, which may be considerable for large models. Finally, command PLATTACH controls the plotting of attached faces. When it is set to ON, attached faces are treated like internal faces and thus are not displayed on any surface plots. 12-30

Version 4.02

Chapter 13

OTHER PROBLEM TYPES Multi-component Mixing

Chapter 13

OTHER PROBLEM TYPES

Multi-component Mixing The theory behind flow problems of this kind and the manner of implementing it in STAR-CD is given in the Methodology volume (Chapter 16, “Multi-component Mixing”). The present chapter contains an outline of the process to be followed when setting up problems involving multiple species and includes cross-references to appropriate parts of the on-line Help system. The latter contains details of the user input required and important points to bear in mind when setting up problems of this kind. Setting up multi-component models Step 1 Go to the Thermophysical Models and Properties folder in the STAR-GUIde system and open the “Additional Scalars” sub-folder. Set up a scalar variable for each species participating in the fluid mixture. The properties of each scalar are specified in the “Molecular Properties (Scalar)” panel, in two ways: 1. By choosing option Define scalar material and then typing in values yourself. Clicking Defaults instructs pro-STAR to fill the remaining boxes with default values (those of air). 2. By choosing option Select scalar from database (see topic “Fluid Property Database”). pro-STAR then fills in all the required values using data stored in file props.dbs. It is important that definition of all material (stream) properties via panel “Molecular Properties” has already been completed before any scalar properties are defined. In multi-stream flow problems, a scalar can be present in some streams but not in others, or it can be present in more than one stream. The allocation of scalar variables to streams is entirely up to the user, subject to the following conditions: • •

Each scalar must be defined only once. Some scalar physical properties are stream-independent and must be set when the scalar is first defined. These include molecular weight, specific heat, molecular viscosity and thermal conductivity. • Other properties such as diffusivity and turbulent Schmidt number are stream-dependent and must be set on a stream-wise basis (see Step 3 below) Step 2 Once all scalars are defined, scroll through them one by one via the Scalar # scroll bar at the bottom of the panel to • • •

check all property values in the “Molecular Properties (Scalar)” panel modify a current value by overtyping in the relevant text box; the change is made permanent by clicking Apply delete an unwanted scalar by clicking Delete Scalar.

Step 3 Specify the stream-dependent (or material-dependent) scalar properties using the Version 4.02

13-1

OTHER PROBLEM TYPES

Chapter 13

Multi-component Mixing

“Binary Properties” panel. Once the settings for all scalars in a given stream are complete, click Apply and then move on to the next stream in your model. Step 4 Specify values for the initial mass fraction of each scalar in each stream using the “Initialisation” panel. Step 5 If the stream incorporates porous media sub-domains (see Chapter 6 in this volume), specify the effective mass diffusivity and turbulent Schmidt number for each additional scalar present in your model using the “Additional Scalar Properties” panel (“Porosity” sub-folder). Step 6 Specify scalar boundary conditions using the “Scalar Boundaries” panel (Define Boundary Conditions folder). Step 7 Go to the “Analysis Controls” folder and specify solution control parameters for all currently defined scalars using the “Additional Scalars” panel (Equation Behaviour sub-folder). In multi-stream problems where each stream has a different scalar composition, this panel enables you, in effect, to select which scalars exist in what stream. Step 8 If a transient analysis is to be performed, use the “Analysis Output” panel (“Transient tab”) to specify whether cell and/or wall data for selected scalars need to be printed or written to the transient post file. For transient problems defined in terms of load steps, go instead to the Advanced Transients dialog (see Chapter 5, “Load step controls”) and click one of the Scalars Select buttons. The button to click depends on whether cell or wall data are needed and whether these are to be printed or written to the transient post file. The scalars to be printed or post-processed are selected in the Transient Scalar Selection dialog shown below, by clicking the option button corresponding to the desired scalar number.

Command: 13-2

SCTRANS Version 4.02

Chapter 13

OTHER PROBLEM TYPES Aeroacoustic Analysis

Note that this process should be repeated for every load step in the transient setup. Step 9 If the stream incorporates additional sources for any of the scalars, specify the source strength and distribution using the Scalar tab in the “Source Terms” panel (sub-folder Sources). Useful points on multi-component mixing 1. The under-relaxation factors for all scalar transport equations should be set to the same value. Note that this factor has no effect for scalars calculated by an internal method or by user coding. 2. For thermal problems, the scalar under-relaxation factors should equal that for the energy equation. For combusting or reacting flows, the recommended range is 0.3 to 0.7. 3. For efficient utilisation of computer memory, it is recommended that scalar variable numbers are continuous and start at 1. 4. For problems involving large changes in temperature, it is recommended that the specific heat of both background fluid and active species is defined as a polynomial function of temperature (see reference [1]). For scalars, this can be done in the “Polynomial Function Definition (Viscosity and Conductivity)” dialog that opens from the “Molecular Properties (Scalar)” panel. A polynomial variation for molecular viscosity and thermal conductivity can be specified in the same way. An ideal-gas variation for the density is also recommended, if necessary with a compressible setting. 5. pro-STAR allows new scalar species to be added to its built-in property database (see topic “Fluid Property Database” in the on-line Help system). 6. Details of existing scalar definitions can be saved to a file of form case.scl for use in other problems. To do this, issue command CDSCALAR from pro-STAR’s I/O window. Note that the scalar data are written in the form of appropriate pro-STAR commands (SC, SCPROPERTIES, SCCONTROL, etc.). Thus, it is possible to read them back into a model by executing an IFILE command (see “File manipulation” on page 17-9). 7. STAR uses default wall functions for calculating heat and mass transfer at wall boundaries. Users can supply alternative expressions for heat and mass transfer coefficients in subroutine MODSWF, activated via the “Miscellaneous Controls” STAR-GUIde panel.

Aeroacoustic Analysis The theory behind aeroacoustic analysis and the manner of its implementation in STAR-CD is given in the Methodology volume (Chapter 16, “Aeroacoustic Analysis”). The present section contains an outline of the process to be followed when setting up a problem of this type. Also included are cross- references to appropriate parts of the on-line Help system, containing details of the user input required. Setting up aeroacoustic models Step 1 Switch on the aeroacoustic modelling facility using STAR GUIde’s “Select Version 4.02

13-3

OTHER PROBLEM TYPES

Chapter 13

Aeroacoustic Analysis

Analysis Features” panel: • •

Select On from the Aeroacoustic Analysis menu If a transient analysis mode has already been selected, a pop-up panel will appear, warning you that the model must be run in steady-state mode. Click Yes to confirm your choice and proceed with the analysis. Note that the displayed option in the Time Domain menu will automatically change to Steady State. • Click Apply. Note that an additional folder called Aeroacoustic Analysis will now appear in the NavCenter tree. Step 2 Open the “Aeroacoustic Analysis” panel. By default, the Aeroacoustic Equation Sources switch is turned On. The default control parameters required for the numerical solution algorithms are also set and are explained by the on-line Help text. If you wish to make any changes, enter the required values in the panel and then click Apply. Step 3 Perform the usual model setup in the Thermophysical Models and Properties folder: In particular, make sure that: •

A density option appropriate to incompressible flow is selected in the “Molecular Properties” panel • A two-equation, k-ε type turbulence model has been selected in the “Turbulence Models” panel Step 4 Specify initial conditions, boundary conditions and control parameters and then run STAR as normal, making sure that the analysis has converged. The aeroacoustic results will be automatically stored in the solution (.ccm) file as an extra scalar variable called AALS (Aeroacoustic Lilley Source). If the maximum number of iterations is reached without convergence, it is important to restart the analysis and run it to convergence. Step 5 Use the facilities of the Post-Processing folder to load and display the distribution of the AALS variable, using only cell-based or vertex-based values Useful points on aeroacoustic analyses 1. If you require an initial solution without the overheads of calculating aeroacoustic source terms at the last iteration, simply turn the Aeroacoustic Equation Sources switch Off, click Apply, and then perform the analysis as usual. You will then need to restart the analysis, turn the switch On and perform one iteration to obtain the aeroacoustic results. 2. Note that STAR-CD returns the logarithmic values of the aeroacoustic sources. If you want to display the actual values, you will first need to calculate the antilogarithm of the stored scalar using the facilities of the Post Register Operations dialog (see Chapter 13, “The OPERATE utility” in the Post-Processing User Guide).

13-4

Version 4.02

Chapter 13

OTHER PROBLEM TYPES Liquid Films

Liquid Films The theory behind the liquid film model and details of its implementation in STAR-CD is given in Chapter 16, “Liquid Films” of the Methodology volume. This section contains an outline of the steps to be followed when setting up a liquid film simulation. Setting up liquid film models The liquid film model can only be used in transient cases. Simulations employing this feature typically involve droplet deposition on wall boundaries, formation of liquid films and film interaction with the surrounding fluid and walls. The basic steps for setting up such a model are as follows: Step 1 Open the “Select Analysis Features” panel in STAR GUIde and turn On the Liquid Films option. A pop-up panel may appear, warning you that the model must be run in transient mode. In such a case, click Yes to confirm your choice. The Time Domain menu setting will then change to Transient. Click Apply. The Liquid Films folder will appear in the NavCenter tree, containing the necessary panels for liquid film analysis. Step 2 If necessary, allow for the presence of droplets in your model by selecting option Lagrangian Multi-Phase from the Multi-Phase Treatment menu and clicking Apply. The Lagrangian Multi-Phase folder will then appear in the NavCenter tree, containing panels needed for specifying droplet parameters (see Chapter 9, “Setting Up Lagrangian Multi-Phase Models” in the CCM User Guide). This option should be selected if either • •

droplets are injected into the solution domain and their behaviour needs to be modelled as part of the analysis, and/or droplets are generated by the film itself through a stripping process.

Step 3 The Liquid Films folder will contain a set of four panels called Film Controls, Film Physical Models and Properties, Film Initialization and Film Boundaries. Note that film property specifications under the second panel of the above set must be supplied even if there are no films initially present in the problem. •

The “Film Controls” panel sets up the basic film modelling parameters. The panel also includes a Liquid Film Creation facility that enables you to specify which (wall or baffle) boundary regions cannot support liquid films.

•

The “Film Physical Models and Properties” panel activates the liquid film model for specified film materials and sets up a property table for each of them. Note that: (a) There is a one-to-one correspondence between film materials defined in the “Film Models” tab and fluid domain materials defined in the “Molecular Properties” panel. Films created for and corresponding to (gaseous) materials in different domains are topologically separate and

Version 4.02

13-5

OTHER PROBLEM TYPES

Chapter 13

Liquid Films

their liquid contents do not mix with each other. (b) In the “Film Properties” tab, you can use the Evaporates to Scalar entry to specify which liquid film component evaporates to/condenses from which gas component. To determine which droplet component becomes which liquid film component when a droplet hits a wall, the following rules are used: i) If a liquid film component and a droplet component evaporate to the same scalar, these components are assumed to exchange mass. ii) If the Evaporates to Scalar setting for a liquid film component is NONE, then this component name is compared against each droplet component name (for single-component droplets, the droplet name is taken as the component name). If the component names match, the matched components are assumed to exchange mass. •

In multi-component liquid film simulations: (a) The specified single value of binary diffusivity is assumed to apply to all components in the film mixture (b) For problems involving evaporation from the film surface, the partial pressure of each component on the gas side of the interface should be calculated using subroutine LQFPRO (see the “Multi Component” on-line Help topic)

•

The “Film Initialization” specifies film initial conditions for each boundary region that can support films. If no initialization is specified, the film thickness on that region is assumed to be zero.

•

The “Film Boundaries” panel sets up film boundary conditions. Each boundary condition is applied to the edges shared by a film and a non-film region. The currently available boundary condition types are Outlet and Inlet. In the latter case, if no boundary conditions are specified for a given variable, the cell value is used as an inlet value (i.e. a Neumann condition applies).

Step 4 Specify initial conditions, boundary conditions and solution control parameters for the domain material (normally gas) surrounding the film and then run the STAR solver as normal. Due to internal parameter settings and having to work with possibly very small numbers such as film thickness, it is recommended that the solver be run in double precision. Step 5 Analysis results pertaining to films are treated by pro-STAR as wall data. Such data items appear in the scroll lists of panel “Analysis Output”, in both the “Post tab” tab and the “Transient tab” tab, so that you may select what is to be included in the .ccm and .pstt files, respectively. pro-STAR assigns names such as LFTHK (film thickness) and LFT (film temperature) to film variables. A complete list can be obtained by issuing the PLIST command from the I/O window. Assuming that the contour plot mode is already selected, a typical pro-STAR macro to plot a scalar film variable is: 13-6

Version 4.02

Chapter 13

OTHER PROBLEM TYPES Liquid Films

trlo,, store last clrw getb lfthk cset news shell cset subs name wall wplot In conjugate heat transfer cases, the following commands may be useful for selecting only liquid film cells on the fluid side of the interface: cset newset fluid cset add atsh cset subs name wall To load film velocity components (LFU, LFV, LFW) as a vector, use getb lu,lv,lw Film stripping This process can be modelled in two ways: 1. Via user subroutine FDBRK. If active, the subroutine will be called at all wall faces containing films, at a point just before the first droplet tracking stage in a new time step. The user code must provide all necessary information regarding the new (stripped) droplets leaving the film, including initial injection velocity and global position coordinates. 2. Via an internal stripping model, currently available as a beta feature (see Chapter 13 of the Supplementary Notes volume). If droplets are generated solely by the stripping process, it is still necessary to define droplet properties in advance, as for normal injected droplets. The new (stripped) droplets must have a type associated with them, which has previously been defined in pro-STAR. Obviously, droplet properties should be consistent with those of the parent film.

Version 4.02

13-7

Chapter 14

USER PROGRAMMING Introduction

Chapter 14

USER PROGRAMMING

Introduction This chapter describes how the user can modify or supplement some of the standard features and operations of STAR, such as physical properties, boundary conditions, additional sources of momentum, energy, etc. via user-supplied FORTRAN subroutines. The latter are collectively referred to as UFILE routines. The full set of currently available user programming inputs comprises: 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32.

Boundary conditions Density (equation of state) Molecular viscosity (including non-Newtonian flow) Specific heat Temperature to enthalpy conversion and vice versa Thermal conductivity Molecular diffusivity for chemical species Properties of distributed resistance Thermal and mass diffusion within distributed resistance sub-domains Effective viscosity and turbulence length scale Turbulence model parameters (including two-layer models) Turbulence characteristics within distributed resistance sub-domains Local injection or removal of fluid Momentum, enthalpy and turbulence sources Solar and gaseous radiation properties Free surface and cavitation models and properties Heat, mass and momentum transfer in two-phase Lagrangian flow Droplet initial conditions and physical properties Droplet behaviour near walls Inter-droplet collision modelling Eulerian multi-phase drag, turbulence and heat transfer Chemical reaction rates and chemical species mass fractions Chemical species and thermal NOx sources Parameters for sliding mesh and rotating reference frame problems Moving mesh coordinates Cell layer removal or attachment Initial conditions Formation and behaviour of liquid films on walls and baffles Wall functions for momentum, heat and mass transfer Time-step size for transient problems Special post-processing Variation of blending factor for higher-order discretisation schemes

Subroutine Usage To use UFILE routines you must execute the following steps: Step 1 Create a subdirectory called ufile under your present working directory as Version 4.02

14-1

USER PROGRAMMING

Chapter 14

Subroutine Usage

follows: • • •

Choose File > System Command from the menu bar to display the System Command dialog Type ufiles in the command text box Click Apply and then Close

Step 2 Select the User option in the appropriate STAR GUIde panel or pro-STAR command, depending on the special feature that needs to be modelled, as discussed in “Description of UFILE Routines” on page 14-5. Step 3 Before a user routine can be used, it must be copied into its own individual file within the ufile directory created earlier. If you are doing this from scratch, it is convenient to start by copying a skeleton (dummy) version of the relevant subroutine into ufile. •

•

14-2

If you want to do this immediately, click Define user coding in your current panel. A file of the right name containing the right dummy subroutine will be created automatically. If you want to inspect the dummy subroutine listing before proceeding further, go to the main pro-STAR window and select Utility > User Subroutines from the menu bar. This activates the User Subroutines dialog shown below. The dialog box is made up of two sub-windows. The lower one lists all subroutine names, their description and the pro-STAR command that activates them. Selecting any line with the mouse displays the default (dummy) code for that subroutine in the upper part of the box. The relative size of the two sub-windows can be adjusted by dragging the control ‘sash’ (the small square on the right-hand side) up and down.

Version 4.02

Chapter 14

USER PROGRAMMING Subroutine Usage

Command:

USUBROUTINE

The required subroutine(s) may be copied into the ufile directory in one of the following ways: 1. Automatically — click the Write Auto button. This copies all subroutines already selected implicitly via the User option in the various STAR GUIde panels (or via the corresponding pro-STAR commands). Such subroutines are also marked in the above list with an asterisk. Note that if more selections are made after the above dialog box has been opened, it is necessary to update the display of selected routines by clicking the Update List button. 2. Explicitly — click the Write File button. This copies the subroutine that is currently on view. In Unix systems, the subroutine file names are of the form Usubname.f. If a file of the same name already exists in the ufile subdirectory, a new file will be created called Usubname.f.new. Note that generating a subroutine file in this way is necessary only if • • • Version 4.02

the subroutine is to be set up for the first time, or an existing subroutine is to be replaced, or you are updating user code from an earlier version of STAR-CD. 14-3

USER PROGRAMMING

Chapter 14

Subroutine Usage

Step 4 Edit the existing or newly-created subroutine file as required, for example by using pro-STAR’s built-in file editor (see the section on “File manipulation” on page 17-9). This is necessary in order to either • •

utilise some or all of the existing example coding (by removing the comment character, C, from the beginning of the line), or add other coding, as appropriate.

Step 5 The version of a subroutine that is to be used in the current run should always be located in a file called Usubname.f within the ufile subdirectory. Older files bearing the same name should either be overwritten or renamed. Once the above process is complete, the required user routines are automatically passed on to the STAR-CD system in source form. They are then compiled and linked to the main program modules (see Chapter 17, “pro-STAR environment variables”). Note that STAR will issue a warning message if it does not find any of the required subroutines but will carry on with the run all the same. Useful points As a general rule, user routines should be written with due care. You should ensure that results produced by user code are reasonable and physically meaningful, by implementing suitable checks and by printing appropriate diagnostic messages whenever necessary. Default user routines for all modelling functions listed in the “Introduction” are supplied, containing sample coding. It should be noted that: 1. Most routines are called for every cell, boundary, or droplet (as appropriate for the routine and model in hand), so a penalty is paid in terms of execution time when they are active. However, the increase in CPU time may be minimised through efficient programming, while keeping the source coding as brief and simple as possible. 2. Each routine has appropriate input data, described in a nomenclature text stored in file nom.inc in the ufile directory. 3. Each routine includes a file called comdb.inc, designed to ensure that the routine uses the same precision as STAR itself. This is done by exploiting the IMPLICIT typing construct present in FORTRAN. According to this, a variable is given a type based upon its initial letter, those beginning with the letters A through H and O through Z being REAL variables, while those beginning with I through N are INTEGER variables. Thus, TIME, ANGLE and SPEED are real but NUMI, IVAL and JUNK are integer. The IMPLICIT typing above can be overruled by an explicit declaration of type, e.g. REAL ITIME makes ITIME real and INTEGER ZVAL will make ZVAL an integer. It is also possible to change the scope of the IMPLICIT typing. This is in fact what comdb.inc does: (a) When STAR is used in single-precision runs, the file contains a single line C

14-4

IMPLICIT DOUBLE PRECISION (A-H,O-Z)

Version 4.02

Chapter 14

USER PROGRAMMING Description of UFILE Routines

which is just a comment, thus preserving the standard implicit typing of real and integer variables. (b) When STAR is used in double-precision runs, the file reads IMPLICIT DOUBLE PRECISION (A-H,O-Z) This means that the IMPLICT typing has been overruled to use double-precision real variables. The implication for users is that to make sure a routine works correctly, variables should be named according to the IMPLICT typing shown above. That way the routine will be compiled with the correct precision. Typical input data for a subroutine includes the following: • • • • • • • •

Cell number Global Cartesian or user-defined local coordinates of the cell centroid Cell table numbers as defined in pro-STAR Material numbers Porous media sub-domain numbers Iteration number Time Nodal values of the field variables

For more information on input data for the UFILE routines, see the nomenclature file (nom.inc). The variables in the argument list are never passed uninitialised: they always have a sensible value, which is usually the value from the previous iteration/time step, if applicable, or more generally the “default” value from the pro-STAR panel. A brief description of each subroutine and how it is activated from pro-STAR is given in the next section.

Description of UFILE Routines Boundary condition subroutines The first ten of the subroutines listed below (all those with names starting with BCD) are activated from the Options menu in the Define Boundary Regions panel, or by command RDEFINE. They specify spatial variations of the boundary conditions at various boundary types. In order to use them, the boundaries comprising the region are first defined in the usual way, including the local coordinate system for the velocity components, the rotational speed of the coordinate frame and any default boundary values that become input values for the subroutines. The coordinates passed to the subroutine are defined in the local coordinate system of the boundary and u, v, w are the corresponding velocities. The latter will be in a rotating frame if this was originally specified. The transformation to the global Cartesian coordinate system is done by STAR. BCDEFI

Version 4.02

Specifies distributions for all dependent variables that vary spatially over an inlet boundary.

14-5

USER PROGRAMMING

Chapter 14

Description of UFILE Routines

BCDEFO

Can be used to specify variations in flow split or mass outflow at outlet boundaries (e.g. in a transient run).

BCDEFP

Specifies boundary conditions at pressure boundaries, i.e. pressure, turbulence intensity, length scale, temperature and species mass fractions.

BCDNRP

Specifies boundary conditions at non-reflective pressure boundaries

BCDEFS

Specifies boundary conditions at stagnation boundaries

BCDNRS

Specifies boundary conditions at non-reflective stagnation boundaries

BCDEFW

Specifies variations in wall boundary conditions, including moving wall velocities in local coordinates and in a rotating reference frame. In addition, wall temperature, chemical species mass fraction and heat and mass fluxes, can all be varied over the specified region.

BCDEFF

Specifies non-uniform boundary conditions at free-stream transmissive boundaries, e.g. velocity components, pressure and temperature.

BCDEFT

Specifies boundary conditions at transient wave transmissive boundaries, e.g. velocity components, pressure and temperature.

BCDEFR

Specifies boundary conditions at Riemann invariant boundaries, e.g. velocity components, pressure and temperature.

ROUGHW

Activated from the Roughness menu in the Define Boundary Regions panel for walls and baffles, or by command RDEFINE. It specifies a user-supplied wall roughness model, in problems where wall functions are used for modelling flow near the wall. STAR will default to the smooth-wall behaviour should you activate this subroutine but provide no code for it.

Material property subroutines CONDUC

14-6

Activated from the Conductivity menu in the Molecular Properties (Liquids and Gases) panel or Material Properties (Solids) panel, or by command CONDUCTIVITY. It specifies the thermal conductivity within a material in heat transfer problems. The thermal conductivity can vary both spatially and with temperature.

Version 4.02

Chapter 14


CONVET

Activated from the Specific Heat menu in the Molecular Properties (Liquids and Gases) panel or by command SPECIFICHEAT. The activation works in an exclusive manner, i.e. choosing this option excludes use of subroutines CONVTE, COTEET and SPECHT. It supplies the variation of temperature T with enthalpy h and any other scalar variable, i.e. T ( h , m 1 , m 2 , … ) , in any way chosen by the user (e.g. analytically or by means of a table). The returned values are valid over a specified temperature range. If the relationship involves other scalar variables, it is necessary to supply values for the partial derivatives ∂T ⁄ ∂h and ∂T ⁄ ∂m k . STAR also requires the inverse relationship, h ( T , m 1 , m 2 , … ) , for internal calculation purposes and inverts T automatically, using an efficient iterative technique.

CONVTE

Activated from the Specific Heat menu in the Molecular Properties (Liquids and Gases) panel or by command SPECIFICHEAT. The activation works in an exclusive manner, i.e. choosing this option excludes use of subroutines CONVET, COTEET and SPECHT. It supplies the variation of enthalpy h with temperature T and any other scalar variable, i.e. h ( T , m 1 , m 2 , … ) , in any way chosen by the user (e.g. analytically or by means of a table). The range of validity of the relationship should be specified in terms of a corresponding range in the values of T. If enthalpy is dependent on a scalar variable, it is also necessary to supply the relevant partial derivatives ∂h ⁄ ∂m k . STAR needs the inverse relationship, T ( h , m 1 , m 2 , … ) , for internal calculation purposes and inverts h automatically using an efficient iterative technique. It is helpful (but not essential) to assist the iteration process by supplying ∂h ⁄ ∂T .

COTEET

Activated from the Specific Heat menu in the Molecular Properties (Liquids and Gases) panel or by command SPECIFICHEAT. The activation works in an exclusive manner, i.e. choosing this option excludes use of subroutines CONVET, CONVTE and SPECHT. It supplies two relationships: (a) The variation of enthalpy h with temperature T and any other scalar variable, i.e. h ( T , m 1 , m 2 , … ) , and (b) the variation of temperature T with enthalpy h and any other scalar variable, i.e. T ( h , m 1 , m 2 , … ) . These should be valid over a given temperature range. Obviously, the two relationships must be consistent. If additional scalar variables are involved, it is also necessary to supply the relevant partial derivatives ∂h ⁄ ∂m k . The COTEET option should be used if the user wants to bypass STAR’s internal calculation procedure for the inverse temperature/enthalpy relationship (see the CONVET, CONVTE description above) in favour of a supplied relationship.

Version 4.02

14-7

USER PROGRAMMING

Chapter 14


14-8

DENSIT

Activated from the Density menu in the Molecular Properties (Liquids and Gases) panel or by command DENSITY. It supplies equations of state for density calculations that are not included in the standard options. For compressible flow cases where density is a function of pressure, the routine must also specify the partial derivative ∂ρ ⁄ ∂ p and return it in parameter DENDP.

DIFFUS

Activated from the Material Mass Diffusivity menu in the Binary Properties (Additional Scalars) panel or by command DIFFUSIVITY. It supplies the molecular diffusivity of the background material in multi-component mixing problems.

PORCON

Activated from a menu in the Thermal Properties (Porosity) panel or by command POREFF. It supplies functions for the calculation of effective thermal conductivity and turbulent Prandtl number within a distributed resistance sub-domain.

PORDIF

Activated from a menu in the Additional Scalar Properties (Porosity) panel or by command SCPOROUS. It supplies functions for the calculation of effective mass diffusivity and turbulent Schmidt number within a distributed resistance sub-domain.

PORKEP

Activated from a menu in the Turbulence Properties (Porosity) panel or by command PORTURBULENCE. It specifies non-uniform distributions of turbulence intensity and dissipation length scale within a distributed resistance sub-domain.

POROS1

Activated from the Resistance Coefficients menu in the Resistance and Porosity Factor panel or by command POROSITY. It defines spatially varying coefficients α and β within a distributed resistance sub-domain. The user can also specify them in terms of a local coordinate system.

POROS2

Activated from the Resistance Coefficients menu in the Resistance and Porosity Factor panel or by command POROSITY. It defines the resistance components ( k 1 , k 2 , k 3 ) directly instead of via the resistance coefficients α and β. This facility is a useful alternative way of specifying a non-linear variation of porous resistance with velocity. For this purpose, the global Cartesian velocity components are supplied to the subroutine.

Version 4.02

Chapter 14


SPECHT

Activated from the Specific Heat menu in the Molecular Properties (Liquids and Gases) panel or Material Properties (Solids) panel, or by command SPECIFICHEAT. The activation works in an exclusive manner, i.e. choosing this facility excludes use of subroutines CONVTE, CONVET and COTEET. The subroutine supplies the variation of fluid or solid mean specific heat with temperature and other quantities, at constant pressure. It is particularly useful in modelling combusting or reacting flows exhibiting substantial variation in the value of this property. STAR calculates the temperature T from the iterative expression T

(n)

h = --------------------(n – 1) (cp)

(14-1)

where n is the iteration number and c p is the mean specific heat. THDIFF

Specifies a user-supplied method of calculating the thermal diffusion coefficient for chemical species scalars (see Chapter 5, “Subroutine THDIFF Set-up” in the Supplementary Notes volume)

VISMOL

This subroutine is activated from the Molecular Viscosity menu in the Molecular Properties (Liquids and Gases) panel or by command LVISCOSITY. It can specify an arbitrary distribution of molecular viscosity, but its principal use is for supplying functions that describe non-Newtonian viscous behaviour.

Turbulence modelling subroutines

Version 4.02

LSCALE

Activated automatically when the k-l model is selected via menu option k-l in panel Turbulence Models (Turbulence tab). It can also be activated by command TURBULENCE. The subroutine supplies the spatial variation of dissipation length scale (l) required by the k-l model.

TWLUSR

Activated from the Two-Layer Model menu in the Turbulence Models panel (Near-Wall Treatment tab) or by command TLMODEL. It defines the user’s own formulation of turbulent behaviour in problems using a two-layer model.

VISTUR

This subroutine is activated from panel Turbulence Models (Turbulence tab) or by command TURBULENCE. The subroutine specifies the turbulent viscosity distribution for a turbulent flow calculation.

14-9

USER PROGRAMMING

Chapter 14


Source subroutines

14-10

FLUINJ

Activated from the Define Source menu in the Source Terms panel (Mass tab). Alternatively, use command RSOURCE. The subroutine initiates fluid injection or removal at specified cells and at a prescribed rate (in units of kg/s/m3). In the case of injection, the properties of the injected fluid, i.e. velocity components, turbulence parameters, temperature, etc. must also be prescribed. This is not required when fluid is removed.

SORENT

Activated from the Define Source menu in the Source Terms panel (Enthalpy tab) or by command RSOURCE. It specifies additional enthalpy sources or sinks due, for example, to electric or magnetic fields, chemical or nuclear reaction and thermal radiation. It can also fix the temperature value within a cell by making S1P=GREAT* T fix and S2P=GREAT, where T fix is the desired fixed temperature value and GREAT is a large number used internally by pro-STAR.

SORKEP

Activated from the Define Source menu in the Source Terms panel (Turbulence tab) or by command RSOURCE. It allows the user to redefine the source term components for the k and ε equations, e.g. to account for special effects due to streamline curvature, magnetic fields, etc. The subroutine can also be used to fix the value of k. Note that the quantities S1P and S2P in the example code are the ‘standard’ source and sink terms given in the Methodology volume. Thus the user, in modifying or supplementing the standard expressions, effectively replaces the built-in source terms.

SORMOM

Activated from the Define Source menu in the Source Terms panel (Momentum tab) or by command RSOURCE. It enables the modelling of additional momentum source terms, for example due to magnetic or electric fields. The source terms must be specified per unit volume and linearised as S1P-S2P* φ P , where φ P is the value of the velocity component in question at node P (see the Methodology volume for details). The two components S1P and S2P must be separately specified for the U, V and W momentum equations. The cells in which to insert these sources can be selected by their indices IP, global Cartesian coordinates XP, YP, ZP or the cell table number ICTID.

Version 4.02

Chapter 14


SORSCA

Specifies additional source terms for the scalar variable equations and is activated from one of the following locations: (a) The Define Source menu in the Source Terms panel (Scalar tab) or by command SCSOURCE. The source terms might consist of, for example, the chemical kinetics and rate expressions of a combustion process. (b) The Model Selection menu in panel Cavitation Model or by command CAVITATION. In this case the source terms are used to specify a special cavitation model. The mass fraction value at selected cells can also be fixed via the source terms, in the same manner as that described above for enthalpy.

Radiation modelling subroutines RADPRO

Activated from the Radiative Properties menu in panel Thermal Models (Liquids and Gases) when radiation with participating media is turned on. May also be activated from the Radiative Properties (Solids) panel if solid-fluid heat transfer is turned on. Alternatively, use command RADPROPERTIES. It specifies non-uniform distributions of absorptivity and scattering coefficients within the medium filling the space between radiating boundaries.

RADWAL

Specifies a user-supplied method of calculating radiative properties for solid walls (see Chapter 6, “Surface Properties” in the Supplementary Notes volume)

USOLAR

Activated from the Define Parameters menu in the Thermal Options panel (Solar Radiation section) or by command SOLAR. In transient problems, it enables specification of solar angle and intensity at every time step of the analysis.

Free surface / cavitation subroutines CAVNUC

Version 4.02

This subroutine is required only in cavitation problems using the bubble two-phase model. It is activated from the Parameters for BTF Model menu in panel Cavitation Model or by command CAVNUCLEI. It specifies the number of bubble nuclei per cubic metre and a functional relationship between equilibrium radius and cell pressure.

14-11

USER PROGRAMMING

Chapter 14


CAVPRO

This subroutine is needed in cavitation or free surface problems requiring variable properties. It is activated from one of the following locations: (a) The Saturation Pressure menu in panel Cavitation Model or by command CAVPROPERTY. It then specifies the speed of sound in the current material (for both the liquid and vapour phases) and the saturation vapour pressure. (b) The Saturation Property Variation menu in panel Mass Transfer (Free Surface folder) or by command VAPORIZATION. It then specifies the vaporisation properties of the current material (saturation temperature and vapour pressure plus latent heat of vaporisation).

COMDEN

Calculates species density and its derivative with respect to pressure and temperature for compressible free-surface flows (see Chapter 1 of the Supplementary Notes volume)

FSEVAP

Activated from the Vaporization Rate menu in the Mass Transfer panel (Free Surface folder). Alternatively, use command VAPORIZATION. It calculates the vaporization rate in problems involving mass transfer by evaporation across a free surface.

FSTEN

Activated from the Additional Properties menu in the Heavy Fluid Molecular Properties panel (Free Surface or Cavitation folders). Alternatively, use command STENSION. It calculates values for surface tension coefficient and contact angle in free surface and cavitation problems.

Lagrangian multi-phase subroutines

14-12

COLLDT

Activated from the Collision Model menu in panel Droplet Physical Models and Properties (tab Global Physical Models) or by command DCOLLISION. It specifies the method of detecting inter-droplet collisions in transient Lagrangian flow problems.

COLLND

Activated from the Collision Model menu in panel Droplet Physical Models and Properties (tab Global Physical Models) or by command DCOLLISION. It specifies the method of calculating the droplet number density used for collision modelling in transient Lagrangian flow problems.

DRAVRG

Activated from the Droplet Averaging menu in the Droplet Controls panel or by command DRAVERAGE. It supplies information about average droplet properties calculated while tracking a droplet parcel through the solution domain.

Version 4.02

Chapter 14


Version 4.02

DRHEAT

Activated from the Heat Transfer menu in panel Droplet Physical Models and Properties (tab Droplet Physical Models) or by command DRHEAT. It enables the user to define the heat transfer process between droplets and the surrounding carrier fluid in two-phase Lagrangian flow problems.

DRMAST

Activated from the Mass Transfer menu in panel Droplet Physical Models and Properties (tab Droplet Physical Models) or by command DRMASS. It enables the user to define the mass transfer process between droplets and the surrounding carrier fluid in two-phase Lagrangian flow problems. This subroutine can also be used for specifying mass transfer between a droplet component and multiple scalars in the surrounding medium. This is done by first selecting the component in the scroll list of the Droplet Properties tab and then typing the keyword User in the Evaporates to Scalar box. Alternatively, use command DRCMPONENT.

DROBRK

Specifies a user-supplied droplet break-up model (see Chapter 10 of the Supplementary Notes volume)

DROICO

Activated from the Droplet User Subroutine (Lagrangian Multi-Phase) or by command DRUSER. The subroutine enables the user to specify droplet initial conditions for two-phase, Lagrangian flow problems. In transient problems, the subroutine sets the initial conditions for any calculation time step at which parcels are released.

DROMOM

Activated from the Momentum Transfer menu in panel Droplet Physical Models and Properties (tab Droplet Physical Models) or by command DRMOMENTUM. It enables the user to calculate momentum transfer between droplets and the surrounding carrier fluid in two-phase Lagrangian flow problems.

DROPRO

Enables the user to specify any physical property appearing in panel Droplet Physical Models and Properties (tab Droplet Properties). It is activated by selecting the Subroutine Usage button next to any of the properties displayed on the tab, or by command DRPROPERTIES.

DROWBC

Activated from the Droplet-Wall Interaction menu in panel Droplet Physical Models and Properties (tab Droplet Physical Models) or by command DRWALL. It enables the user to calculate momentum, heat, and mass exchange between droplets and wall boundaries.

14-13

USER PROGRAMMING

Chapter 14


Liquid film subroutines FDBRK

Specifies the method of calculating liquid film stripping by the carrier fluid. It is activated by selecting User in the Stripping and Re-entrainment menu of the Film Controls panel or by command LFSTRIP.

LQFBCD

Specifies boundary conditions at liquid film inlets. It is activated by selecting User in the Film Boundaries panel or by command LQFBC.

LQFINI

Specifies initial conditions for all liquid film variables at a given boundary region. It is activated by selecting User in the Options menu of the Film Initialization panel or by command LQFINITIAL.

LQFPRO

Specifies liquid film physical properties or liquid film component partial pressures. It is activated by selecting the Subroutine Usage button next to any of the properties displayed on the Film Properties tab (panel Film Physical Models and Properties), or by command LQFPROPERTY.

LQFSOR

Modifies the source terms of the mass, momentum and enthalpy equations for liquid films. It is activated by selecting User in the User Defined Source Term menu of the Film Controls panel or by command LFQSOR.

Eulerian multi-phase subroutines

14-14

UEDRAG

This subroutine is used in Eulerian multi-phase problems to calculate the total drag force, per unit volume of the computational cell. It is activated from the main menu in the Drag Forces panel (Eulerian Multi-Phase folder) or by command EDRAG.

UETURB

This subroutine is employed in Eulerian multi-phase problems to calculate the response coefficient C t . The latter is used to derive the dispersed phase turbulence characteristics from those of the continuous phase. It is activated from the Ct Model menu in the Turbulence Models panel (Multiphase Options tab) or by command ETURB. Note that the drag force per unit volume referred to above is supplied as an input variable since it is often a parameter in C t formulations.

Version 4.02

Chapter 14


UEHEAT

This subroutine is employed in Eulerian multi-phase problems to calculate the Nusselt number. The latter is then used in the calculation of the mean interface heat transfer coefficient, which in turn is used to compute the interphase heat transfer when solving for energy for either phase. The subroutine is activated from the Interphase Heat Transfer panel (Eulerian Multi-Phase folder) or by command EHTRANSFER.

Chemical reaction subroutines COALC

Specifies a user-supplied char combustion model in coal combustion cases (see Chapter 3, “User Coding” in the Supplementary Notes volume)

COALV

Specifies a user-supplied volatile evolution model in coal combustion cases (see Chapter 3, “User Coding” in the Supplementary Notes volume)

FULPRO

Specifies user-defined fuel physical properties and chemical reaction parameters for use with the Shell ignition and knock models. It can be activated in two ways: (a) From the Ignition Reaction Based On menu in panel Ignition (folder Chemical Reactions). Alternatively, type command IGNMODEL. (b) From the Knock Reaction Based On menu in panel Knock (folder Chemical Reactions). Alternatively, type command KNOCK.

Version 4.02

NOXUSR

Activated by the Thermal NOx, Prompt NOx, or Fuel NOx menus in panel Emissions (Chemical Reactions folder), or by command NOX. It contains user coding for the calculation of thermal, prompt or fuel NOx sources.

PARUSR

Specifies a user-supplied particle component evolution model in coal combustion cases (see Chapter 3, “User Coding” in the Supplementary Notes volume)

RATUSR

Specifies a user-supplied method of calculating turbulence effects in complex chemistry models (see Chapter 4, “RATUSR User Subroutine” in the Supplementary Notes volume)

REACFN

Activated from the Rate Model menu in the Reaction System (Chemical Reactions) panel when option Combined/User is chosen as the current reaction model. Alternatively, type command RRATE. It specifies a user-supplied reaction rate for chemical reactions of any type.

14-15

USER PROGRAMMING

Chapter 14


REACUL

Specifies a user-supplied reaction rate for the Coupled Complex Chemistry model. It is activated from panel Reaction System, having previously selected option User from the Reaction Rate Calculated by menu in panel Scheme Definition (folder Chemical Reactions). Alternatively, type command CRMODEL.

SCALFN

In some circumstances, chemical species mass fractions can be calculated from user-prescribed algebraic relationships, e.g. stoichiometric relationships, rather than from finite-volume transport equations. These algebraic relationships can be specified in this routine, activated from the Solution Method menu in panel Additional Scalars (Solution Controls > Equation Behavior sub-folder). Alternatively, use command SCPROPERTIES.

Rotating reference frame subroutines UOMEGA

Calculates values of angular velocity (omega) for problems involving rotating reference frames. It is activated by the User Option menu in the Rotating Reference Frames panel or by command SPIN.

UPOSTM

Generates post-processing data at coupled boundaries. It is used in problems with multiple rotating frames of reference that are solved explicitly. The subroutine is called automatically in the Rotating Reference Frames panel if option Multiple Frames - Explicit is selected from the Reference Frame Treatment menu. Alternatively, use command MFRAME.

Moving mesh subroutines

14-16

NEWXYZ

Activated by selecting Modules > Transient from the main pro-STAR menu to open the Advanced Transients dialog, and then selecting On in the Moving Grid Option menu. Alternatively, use command MVGRID. The subroutine specifies the cell vertex coordinates at a new time. The old time level coordinates are available in the VCORN array and must be overwritten with new coordinates. The sample coding supplied describes a moving mesh that is linearly expanding and contracting between a reciprocating piston and a fixed cylinder head; the piston is driven by a rotating crank mechanism.

UASI

Specifies the time-varying offsets used in matching arbitrary sliding interface (ASI) boundaries. It is called automatically if a model employing sliding events is defined using command EVSLIDE.

Version 4.02

Chapter 14


UBINIT

Specifies initial conditions for cells that are re-incorporated into the solution domain via an INCLUDE event. It is called automatically if command EICOND is issued in a model employing such cells.

UPARM

Generates parameters required for moving meshes. It is called automatically if a moving mesh model is defined using commands in the EVENTS module; i.e. if command MVGRID,ON,EVENT,PROSTAR is issued.

Miscellaneous flow characterisation subroutines

Version 4.02

INITFI

Activated from the Values menu in the Initialization panel (Liquids and Gases or Solids folders). Alternatively, use command INITIAL. It initialises flow field variables to user-specified values. These values override any constant values also appearing in those panels. During an initial field restart, the subroutine can also be used to selectively reset some of the variable values in the field. Note that the subroutine returns velocities in a local coordinate system. STAR transforms them to a stationary global Cartesian system. Velocities in this system will differ from the velocities produced by the subroutine because of this transformation and, when that feature is active, the transformation from a rotating reference frame.

MODSWF

Activated by a button labelled Heat and Mass Transfer in the Miscellaneous Controls (Other Controls) panel, or by command HCOEFF. It modifies or supplies new wall functions for heat and mass transfer. This is useful, for example, in problems involving strong natural convection where the standard formulae for the transfer coefficients might be inaccurate. One such example is included in the sample coding. Mean temperatures and mass fractions for all fluid materials are made available through the parameter list.

PORHT2

Specifies user-supplied coefficients for a quasi-linear relationship between porous solid and fluid temperatures in problems involving conjugate heat transfer in porous media (see Chapter 16, “Inter-phase heat transfer term” in the Supplementary Notes volume).

14-17

USER PROGRAMMING

Chapter 14


VARPRT

Enables specification of either a variable Prandtl number for enthalpy or a variable Schmidt number. It can be activated in two ways: (a) From the Prandtl(Enth). menu in panel Turbulence Models (Turbulence tab). Alternatively, type command COKE. (b) Via the Schmidt Number pop-up menu in the Binary Properties (Additional Scalars) panel, or by typing command SCPROPERTIES. In this case, the subroutine should supply special functions for calculating the turbulent Schmidt number of chemical species in multi-component mixing problems.

Solution control subroutines DTSTEP

Enables the user to specify a variable time step for transient, single-transient or pseudo-transient simulations. It can be activated in three ways: (a) For single-transient cases, select option User in the Time Step Method menu of the Set Run Time Controls panel (Analysis Preparation/Running folder). Alternatively, use command DELTIME (b) For pseudo-transient cases, select option User in the Time Step Option menu of the Set Run Time Controls panel (Analysis Preparation/Running folder). Alternatively, use command TIME. (c) For transient cases, open the Advanced Transients dialog, select the appropriate load step, and then click the User Flag button in front of the time step (Delta Time) box. Alternatively, use command LSTEP. The subroutine can be used, for example, in fire and smoke movement simulations that involve a large, concentrated heat source. The time step can be adjusted in terms of the number of PISO correctors and maximum Courant number. Note that STAR does not alter the number of time steps in a load step, so your code must ensure that the time step lengths are such that the length of the load step is correct.

14-18

Version 4.02

Chapter 14

USER PROGRAMMING Sample Listing

POSDAT

Activated by the User subroutine button in the Analysis Output (Output Controls) panel or by command PRFIELD. It performs special post-processing operations. For example: (a) Variable values at several monitoring locations can be written to user-designated output files for subsequent processing. (b) A bulk averaging scheme can be prescribed for selected flow variables and printed at specified intervals. (c) Calculation of lift and drag coefficients. This subroutine may be called both at the beginning and at the end of every time step or iteration. The place from which it is called is distinguished by the value of parameter LEVEL (=1 — beginning, =2 — end)

VARBLN

Activated by the Blending Method pop-up menus in the Primary Variables panel (Differencing Schemes tab). It can be used to vary the blending factor for higher-order discretisation schemes over the computational domain. Alternatively, use command DSCHEME.

Sample Listing The listing for subroutine CONDUC is given below as an example of the default source code available in STAR-CD. Users wishing to inspect the contents of any other subroutine should start a pro-STAR session and then activate the User Subroutines dialog, as explained in “Subroutine Usage” on page 14-1. C************************************************************************* SUBROUTINE CONDUC(CON,CKNX,CKNY,CKNZ) C Conductivity C************************************************************************* C--------------------------------------------------------------------------* C STAR-CD VERSION 4.00.000 INCLUDE ’comdb.inc’ C COMMON/USR001/INTFLG(100) C INCLUDE ’usrdat.inc’ DIMENSION SCALAR(50) EQUIVALENCE( UDAT12(001), ICTID ) EQUIVALENCE( UDAT11(001), CP ) EQUIVALENCE( UDAT11(002), DEN ) EQUIVALENCE( UDAT11(003), ED ) EQUIVALENCE( UDAT11(006), P ) EQUIVALENCE( UDAT11(007), T ) EQUIVALENCE( UDAT11(008), TE ) EQUIVALENCE( UDAT11(009), SCALAR(01) ) EQUIVALENCE( UDAT11(059), U ) EQUIVALENCE( UDAT11(060), V ) EQUIVALENCE( UDAT11(061), W ) EQUIVALENCE( UDAT11(062), VISM ) EQUIVALENCE( UDAT11(063), VIST ) Version 4.02

14-19

USER PROGRAMMING

Chapter 14

New Coding Practices EQUIVALENCE( UDAT11(067), X ) EQUIVALENCE( UDAT11(068), Y ) EQUIVALENCE( UDAT11(069), Z ) C C----------------------------------------------------------------------C C This subroutine enables the user to specify thermal conductivity. C C CON - Isotropic conductivity or MAX(CKNX,CKNY,CKNZ) if the C conductivity is anisotropic C CKNX - Anisotropic conductivity in x-direction C CKNY - Anisotropic conductivity in y-direction C CKNZ - Anisotropic conductivity in z-direction C C C STAR calls this subroutine for cells and boundaries. C C ** Parameters to be returned to STAR: CON,CKNX,CKNY,CKNZ C C----------------------------------------------------------------------C C Sample coding: To specify thermal conductivity for a group of C cells with cell table numbers 2 and 11 as a function C of temperature C C IF (ICTID.EQ.2.OR.ICTID.EQ.11) CON=4.3+0.001*T C-------------------------------------------------------------------------C RETURN END C

New Coding Practices Most standard STAR-CD V3.2X user coding will work without any modification. The table below explains the differences between V3.2X and V4.00 user routines. Note that STAR-CD V3.2X / V3.1X common blocks / variable names should not be used.

14-20

User routine

Difference

CAVPRO

Freedom to modify AL, AV, TSAT and HVAP has been removed

DENSIT

For the Free Surface model, freedom to modify the heavy/light density has been removed. IFLUTYP and DENDT are no longer present

Version 4.02

Chapter 14

USER PROGRAMMING New Coding Practices

DROWBC

a/ The DRCOMP common block needs to be included to the source, e.g. COMMON /DRCOMP/ NDRCOM_max b/ The DEMUCO array needs to be dimensioned as DEMUCO(NDRCOM_max,*) For more information, please consult the default source code for DROWBC supplied with the STAR-CD installation.

POSDAT

Data structure is different

SPECHT

For the Free Surface model, only the mixture Cp needs to be specified. IFLUTYP is no longer present

VISMOL

For the Free Surface model, only the mixture viscosity needs to be specified. IFLUTYP is no longer present

Users are advised to consult subroutine POSDAT to gain familiarity with STAR-CD’s new face-based data structure. Commonly used V3.2X data items that are now accessed differently in V4.00 are listed below.

V3.2X data item KEY

See material/and loop in posdat.f. doma(nd)%mattyp.eq.FLUID test tells you whether you have a fluid or a solid.

LQ, LCU, LCY, LCO, LSI, LX

lfc

S, SB, SBSI

sv

WF, WFCU, WFCY, WFSI

w

VOLF, VOLCU, VOLCY, VOLSI

svol

AC, COCU, COCY, COSI, COCO, ACB

af

F, FB, FBSI

fl

KEYSUB T(1,IP) T(scalar index +1, IP)

Version 4.02

Equivalent V4.00 data item

doma(nd)%sd(nsd)%irot t(ip) c(IP,scalar index), see posdat.f

14-21

USER PROGRAMMING

Chapter 14

User Coding in parallel runs

User Coding in parallel runs If user coding is present in a parallel run, it is possible that some of the required operations need access to flow field values that are distributed throughout the various computational domains. In such cases, it is necessary to collect such values prior to manipulation and to do this, the supplied coding needs to use special message passing routines. The example shown below is an extract from user subroutine NEWXYZ and it employs a parallel function called IGSUM to find the global number of active cell layers in an engine simulation problem. NLIVE=0 ICELL1=15904 ICELLEND=62209 NCOF=1029 DO I=ICELL1,ICELLEND,NCOF CALL LIVCLL(I,ISTAT) IF (ISTAT.EQ.1) NLIVE=NLIVE+1 ENDDO c. NHPC > 1 if parallel run IF(NHPC.GT.1) NLIVE=IGSUM(NLIVE)

A synopsis of the available message passing-routines is given in Appendix E. These routines should only be called if required; they are not necessary for sequential runs. To aid diagnostics, four variables are provided via file usrdat.inc to user subroutines: IHPC

—

this is the local process number (1 ≤ IHPC ≤ NHPC) IHPC = 1 for a sequential analysis IHPC = 1 for the ‘master’ process in STAR HPC IHPC > 1 for the ‘slave’ process in STAR HPC

NHPC

—

Number of processes (NHPC = 1 for a sequential analysis)

NHHPC

—

Number of fluid ‘halo’ cells on the local process

NTHHPC —

14-22

Total number of ‘halo’ cells on the local process (fluid plus solid)

Version 4.02

Chapter 15

PROGRAM OUTPUT Introduction

Chapter 15

PROGRAM OUTPUT

Introduction Run-time screen output from STAR provides a summary of the input specification for the problem being solved and also allows monitoring of the calculation progress. It is therefore important for users to understand this information and examine it regularly to ensure that • •

the problem has been correctly set up; the calculations are proceeding satisfactorily.

The amount of information displayed is largely up to the user, apart from a core of information that is always produced. The various checks and outputs which are specially activated from pro-STAR’s STAR GUIde environment (Output Controls folder) are described below, along with the permanent output.

Permanent Output The core-level screen information from STAR can be divided into two sections: 1. An echo of the input data provided by the user 2. A display of analysis results and information on the progress of STAR calculations Input-data summary As can be seen in Table 15-1 on page 15-4, the input summary begins with the STAR-CD version number and the date and time of the run. This is followed by a table of essential model data for checking that all important user-defined inputs are correct. All listed data reflect the values stored in the problem (.prob) file. The table is divided into distinct sections, as follows: General Data This section provides general information on the problem at hand, including: • • • •

The case name Number of cells The model’s overall physical dimensions The run precision (single or double)

This section of the table also summarises: • • • • • •

The character of the flow (i.e. steady or transient) The starting iteration number for the calculations The frequency of solution data and screen output The solution algorithm and linear equation solver selected The residual tolerance used for convergence tests The maximum number of iterations or time steps specified

A sample output can be seen in Section A of Table 15-1.

Version 4.02

15-1

PROGRAM OUTPUT

Chapter 15

Permanent Output

Fluid Properties This section comprises one or more tables containing the properties of all fluid domains (streams) included in the model. For each domain, the information supplied includes: • • • • • • •

The variables calculated, including the turbulence model selected and the appropriate characteristic length The physical properties specified, such as density, viscosity, specific heat, and conductivity The pressure/temperature reference locations for the domain The reference pressure and temperature (when appropriate) Any fixed boundary fluxes included in the model The specified initial field values The specified boundary conditions

An example of the output for a multi-domain case appears in Table 15-1, Section B. This shows data for two fluid domains with different physical properties. Solid Properties The fluid domain tables are followed, in the case of solid-fluid heat transfer problems, by tables of properties for solid domains such as density, specific heat and conductivity. This can be seen in Table 15-1, Section C. The reference temperature, initial field values and boundary conditions are also included here. Additional Features In this part of the table, information is provided on any additional features that are active in the model, such as: • • •

Radiation Free surface Run-specific system settings

The sample output of Table 15-1, Section D, indicates use of memory-based scratch files and platform-specific solver optimization. User FORTRAN Coding This section of the table only appears when user-defined FORTRAN coding is active during the calculations. The sample case presented in Table 15-1 does not use this option. Solution Parameters This last section of the table deals with the settings for the control parameters used by the numerical algorithm, such as • • • • • •

relaxation factors, type of differencing scheme used, the corresponding blending factors, residual normalising factors for each fluid and solid domain, solver tolerances, sweep limits.

For transient PISO runs the printout of relaxation factors is suppressed as irrelevant, except for the pressure correction relaxation factor. A typical printout of the above 15-2

Version 4.02

Chapter 15

PROGRAM OUTPUT Printout of Field Values

parameters can be seen in Table 15-1, Section E. Run-time output The run-time output that provides information on the progress of the calculations at each iteration or time step can be seen in Table 15-2 on page 15-6 and is arranged in two sections: 1. The left-hand section contains the global absolute residual histories for each group of transport equation solved (momentum, mass, turbulence, etc.). 2. The right-hand section contains values of the corresponding dependent variables (velocity magnitude, pressure, turbulent viscosity, etc.) at a pre-defined monitoring location in domain no. 1. In steady-state runs, satisfactory progress of the calculations should show • •

a steady reduction in the global absolute residuals from iteration to iteration; stabilisation of the values of flow field variables at the monitoring location.

However, residuals do not always decrease from iteration to iteration and, in some cases, oscillations can be observed. These can be ignored as long as the overall residual levels are reduced over a reasonable number of iterations. Information on total CPU and elapsed times is also given. This output appears on the screen during an interactive session and is also saved in the run-time output (.run) file. Any warning messages generated during the course of the calculations are stored in the run-time optional output (.info) file and should be inspected by the user separately. The .run file also contains a reminder to the effect that warning messages have been produced.

Printout of Field Values The printout of field values for the solution variables is optional and, if present, follows the analysis history output. The output quantity and frequency is left up to the user and may be set using various options in the “Analysis Output” STAR GUIde panel — see the “Print Cell Range” section for steady-state problems. There is also a similar “Print Cell Range” section in the transient-problem version of this panel (“Post tab”). An example of such a printout can be seen in Table 15-3 on page 15-7.

Optional Output All additional outputs are optional and, if requested, will appear in the .info file. Output of additional data is activated by various options in the “Monitor Numeric Behaviour” STAR GUIde panel. Table 15-4 on page 15-8 shows typical information appearing in this file (in this case up to and including data for iteration no. 1). Note that velocity component data at the flow field’s extrema are given in the local coordinate system. This is in contrast to the data described in Chapter 17, “Data repository file (.ccm)”, which are always stored in the global Cartesian coordinate system.

Version 4.02

15-3

PROGRAM OUTPUT

Chapter 15

Example Output

Example Output Table 15-1: |-------------------------------------------| | STAR-CD VERSION 4.00.000 | | THERMOFLUIDS ANALYSIS CODE | | Operating System: Linux | | Stardate: 6-DEC-2005 Startime: 14:32:39 | |-------------------------------------------| |-----------------------------------------------------------| | STAR Copyright (C) 1988-2005, Computational Dynamics Ltd. | | Proprietary data --- Unauthorized use, distribution, | | or duplication is prohibited. All rights reserved. | |-----------------------------------------------------------|

|-------------------------------------------------------------------------------------------| | ---------------------------- PROBLEM SPECIFICATION SUMMARY ---------------------------- | |-------------------------------------------------------------------------------------------| | CASE TITLE .................. => | | NUMBER OF CELLS ............. => 192 | | MESH DIMENSIONS XMIN XMAX YMIN YMAX ZMIN ZMAX | | (IN METRES) ............ => 0.0E+00 6.0E-01 0.0E+00 8.0E-01 0.0E+00 5.0E-02 | | MESH QUALITY ................ => | | Expansion factor .......... => Aver = 1.00, Max = 1.00 (CVs: 43, 44) | | Non-orthogonality (deg).... => Aver = 0.00, Max = 0.00 (CVs: 0, 0) | | RUN PRECISION ............... => Single | | STEADY ANALYSIS ..............=> START FROM ITERATION = 0 | | INITIALISATION .............. => WILL NOT BE EMPLOYED | | DATA DUMP (FILE.ccm) ........ => EVERY 10 ITERATIONS | | SOLUTION PROCEDURE .......... => SIMPLE | | RESIDUAL TOLERANCE .......... => 1.00E-03 | | MAX. NO. OF ITERATIONS ...... => 100 | | RESTART DATA ................ => WILL BE SAVED ON out.ccm | | SURFACE DATA ................ => WILL NOT BE SAVED | | CONVERGENCE DATA ............ => WILL BE PRINTED ON FILE.info | | FIELD DATA .................. => WILL BE PRINTED | | LIN. ALG. EQU. SOLVER ....... => Conjugate gradient with Incompl. Choleski precond. | |-------------------------------------------------------------------------------------------| |-> DOMAIN 1: FLUID ------------------------------------------------------------------| |-------------------------------------------------------------------------------------------| | SOLVE ....................... => U, V, P, TE, ED, | | FLUID FLOW .................. => TURBULENT INCOMPRESSIBLE | | TURBULENCE MODEL ............ => HIGH RE K-EPS MODEL | | CONSTANTS ................. => C_mu=0.09, C_1=1.44, C_2=1.92, C_3=1.44, C_4=-0.33 | | => cappa=0.419, Pr_k=1.00, Pr_eps=1.219, Pr=0.90 | | REFERENCE PRESSURE .......... => PREF = 1.000E+05 Pa | | DENSITY ..................... => IDEAL GAS: MOLW = 2.891E+01 | | MOLECULAR VISCOSITY ......... => CONSTANT MU = 1.810E-05 Pas | | | | INITIAL FIELD VALUES ........ => u v w p | | => 0.0E+00 0.0E+00 0.0E+00 0.0E+00 | | => Tur.In. Len.Sc. | | => 2.6E-02 1.0E-01 | | BOUNDARY CONDITIONS ......... => | | Reg. 0 Wall: U = 0.000E+00 V = 0.000E+00 W = 0.000E+00 Om = 0.000E+00 in C.Sys. 1 | | Elog = 9.000E+00 | | Reg. 1 Inlet: U = 0.000E+00 V = 5.000E+01 W = 0.000E+00 Om = 0.000E+00 in C.Sys. 1 | | TI = 5.000E-02 TLS = 5.000E-03 | | Reg. 2 Constant piezomet. pressure: P = 0.000E+00 | | TI = 5.000E-02 TLS = 5.000E-03 | | Reg. 5 Symmetry plane | |-------------------------------------------------------------------------------------------| |-> DOMAIN 2: FLUID ------------------------------------------------------------------| |-------------------------------------------------------------------------------------------| | SOLVE ....................... => U, V, P, | | FLUID FLOW .................. => LAMINAR INCOMPRESSIBLE | | TURBULENCE MODEL ............ => | | PRESSURE REF. CELL .......... => 145 | | REFERENCE PRESSURE .......... => PREF = 1.000E+05 Pa | | DENSITY ..................... => CONSTANT RHO = 1.000E+03 kg/m3 |

15-4

Version 4.02

A

B

B

Chapter 15

PROGRAM OUTPUT Example Output | MOLECULAR VISCOSITY ......... => CONSTANT MU = 1.000E-03 Pas | | | | INITIAL FIELD VALUES ........ => u v w p | | => 0.0E+00 0.0E+00 0.0E+00 0.0E+00 | | BOUNDARY CONDITIONS ......... => | | Reg. 0 Wall: U = 0.000E+00 V = 0.000E+00 W = 0.000E+00 Om = 0.000E+00 in C.Sys. 1 | | Elog = 9.000E+00 | | Reg. 3 Inlet: U = 0.000E+00 V = 1.000E-01 W = 0.000E+00 Om = 0.000E+00 in C.Sys. 1 | | Reg. 4 Outlet: Flow split = 1.000E+00 | | Reg. 5 Symmetry plane | |-------------------------------------------------------------------------------------------| |-> DOMAIN 3: SOLID ------------------------------------------------------------------| |-------------------------------------------------------------------------------------------| | SOLVE ....................... => T | | REFERENCE TEMPERATURE ....... => TREF = 2.730E+02 K | | DENSITY ..................... => CONSTANT rho = 9.000E+03 kg/m3 | | SPECIFIC HEAT ............... => CONSTANT c = 3.800E+02 J/kgK | | CONDUCTIVITY ................ => CONSTANT k = 3.800E+02 W/mK | | INITIAL FIELD VALUES ........ => T | | => 2.9E+02 | | RELAX. FACT. IN SOLID ....... => URF = 1.00 | | BOUNDARY CONDITIONS ......... => | | Reg. 0 Wall: U = 0.000E+00 V = 0.000E+00 W = 0.000E+00 Om = 0.000E+00 in C.Sys. 1 | | Reg. 5 Symmetry plane | |-------------------------------------------------------------------------------------------| |-> ADDITIONAL FEATURES USED --------------------------------------------------------------| |-------------------------------------------------------------------------------------------| | PLUG AND PLAY | | RAMFILES OPTION ENABLED | | TURBO OPTION ENABLED | |-------------------------------------------------------------------------------------------| |-> SOLUTION CONTROL PARAMETERS | |-------------------------------------------------------------------------------------------| | VARIABLE | Mome Mass Turb ---| |-------------------------------------------------------------------------------------------| | RELA. FAC. | 7.000E-01 2.000E-01 7.000E-01 ---| | DIFF. SCH. | UD CD UD ---| | DSCH. FAC. | 0.000E+00 1.000E+00 0.000E+00 ---| | SOLV. TOL. | 1.000E-01 5.000E-02 1.000E-01 ---| | SWEEP LIM. | 100 1000 100 ---| |-------------------------------------------------------------------------------------------|

Version 4.02

C

D

E

15-5

PROGRAM OUTPUT

Chapter 15

Example Output

Table 15-2: Iter. I--------------- GLOBAL ABSOLUTE RESIDUAL ------------------I I-------- FIELD VALUES AT MONITORING No Mome Mass Turb ----Vel Pres TurVis -1 1.00E+00 1.00E+00 9.98E-01 ----5.66E+01 6.20E-01 1.60E-03 -2 6.78E-02 1.36E-01 3.01E-01 ----5.46E+01 4.04E+00 8.05E-03 -3 2.36E-02 3.95E-02 1.56E-01 ----5.38E+01-2.15E+00 1.48E-02 -4 1.26E-02 4.50E-02 9.22E-02 ----5.32E+01-4.45E+00 1.94E-02 -5 8.15E-03 3.10E-02 5.23E-02 ----5.25E+01-3.93E+00 2.19E-02 -6 6.02E-03 2.11E-02 2.85E-02 ----5.18E+01-2.62E+00 2.32E-02 -7 4.54E-03 1.64E-02 1.48E-02 ----5.11E+01-1.32E+00 2.38E-02 -8 3.41E-03 1.33E-02 7.31E-03 ----5.06E+01-5.78E-01 2.40E-02 -9 2.56E-03 1.13E-02 3.30E-03 ----5.01E+01-4.11E-01 2.41E-02 -10 1.90E-03 9.15E-03 1.79E-03 ----4.98E+01-1.98E-01 2.40E-02 -11 1.39E-03 7.66E-03 1.12E-03 ----4.96E+01 2.35E-01 2.40E-02 -12 1.00E-03 6.56E-03 7.23E-04 ----4.94E+01 4.78E-01 2.39E-02 -13 7.22E-04 5.37E-03 5.25E-04 ----4.94E+01 6.69E-01 2.39E-02 -14 5.25E-04 4.33E-03 3.95E-04 ----4.93E+01 9.90E-01 2.38E-02 -15 3.83E-04 3.48E-03 2.69E-04 ----4.93E+01 1.08E+00 2.38E-02 -16 2.81E-04 2.86E-03 1.78E-04 ----4.93E+01 1.29E+00 2.38E-02 -17 2.09E-04 2.39E-03 1.06E-04 ----4.93E+01 1.34E+00 2.38E-02 -18 1.56E-04 2.01E-03 6.37E-05 ----4.92E+01 1.38E+00 2.38E-02 -19 1.18E-04 1.73E-03 3.72E-05 ----4.92E+01 1.41E+00 2.38E-02 -20 9.05E-05 1.46E-03 2.25E-05 ----4.92E+01 1.44E+00 2.38E-02 -21 7.07E-05 1.22E-03 1.28E-05 ----4.92E+01 1.45E+00 2.38E-02 -22 5.52E-05 1.01E-03 8.80E-06 ----4.92E+01 1.46E+00 2.38E-02 -23 4.36E-05 8.36E-04 5.53E-06 ----4.92E+01 1.50E+00 2.38E-02 -&&&& ---------------------------------------------------------------------------------------------------------------------------

POINT -------------------------

63 ----------I -------------------------------------------------

************************************************** * THERE ARE WARNINGS IN FILE out.info **************************************************

*** CALCULATIONS TERMINATED - CONVERGENCE CRITERION SATISFIED

15-6

Version 4.02

Chapter 15

PROGRAM OUTPUT Example Output

Table 15-3: I-------------------------------------------- FIELD VALUES AT ITERATION NO CELL NO 1 11 21 31 41 51 61 71 81 91 101 111 121 131 141 151 161 171 181 191

U VEL -3.697E-03 -1.021E-02 -2.619E-02 -2.700E-02 -4.814E-02 -5.854E-02 -5.772E-02 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 3.276E-05 2.712E-05 2.454E-05 6.541E-06 1.097E-05 6.675E-06 -1.134E-05

V VEL 5.000E+01 5.018E+01 5.007E+01 5.026E+01 5.015E+01 4.989E+01 4.935E+01 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 9.994E-02 9.966E-02 1.001E-01 1.000E-01 1.001E-01 1.000E-01 1.001E-01

PRESS 1.635E+01 6.546E+00 1.337E+01 1.697E+00 9.038E+00 1.610E+01 3.980E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 -1.737E-03 -1.239E-02 -6.395E-03 -9.026E-04 -9.892E-03 -4.241E-03 -9.797E-03

TUR EN 8.578E+00 4.465E+00 6.329E+00 3.718E+00 5.222E+00 1.116E+01 1.223E+01 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00

DISSI 8.004E+02 2.369E+02 4.550E+02 1.673E+02 2.980E+02 5.849E+02 6.708E+02 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00

VISCO 9.818E-03 8.987E-03 9.403E-03 8.825E-03 9.775E-03 2.275E-02 2.381E-02 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00

I--------------------------------------------| Wall data at iteration No Region No: 0 Cell No Y-PLUS 49 2.858E+03 50 2.939E+03 51 2.999E+03 52 3.043E+03 53 3.075E+03 54 3.098E+03 55 3.114E+03 56 3.125E+03 57 3.132E+03 58 3.136E+03 59 3.139E+03 60 3.140E+03 61 3.140E+03 62 3.139E+03 63 3.137E+03 64 3.135E+03 129 0.000E+00 130 0.000E+00 131 0.000E+00 132 0.000E+00 133 0.000E+00 134 0.000E+00 135 0.000E+00 136 0.000E+00 137 0.000E+00 138 0.000E+00 139 0.000E+00 140 0.000E+00 141 0.000E+00 142 0.000E+00 143 0.000E+00 144 0.000E+00 65 0.000E+00 81 0.000E+00 97 0.000E+00 113 0.000E+00 80 0.000E+00 96 0.000E+00 112 0.000E+00 128 0.000E+00

NORM DIST 2.500E-02 2.500E-02 2.500E-02 2.500E-02 2.500E-02 2.500E-02 2.500E-02 2.500E-02 2.500E-02 2.500E-02 2.500E-02 2.500E-02 2.500E-02 2.500E-02 2.500E-02 2.500E-02 2.500E-02 2.500E-02 2.500E-02 2.500E-02 2.500E-02 2.500E-02 2.500E-02 2.500E-02 2.500E-02 2.500E-02 2.500E-02 2.500E-02 2.500E-02 2.500E-02 2.500E-02 2.500E-02 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00

END OF EXECUTION - STAR CPU time is 0.36

Version 4.02

WALL TEMP 2.930E+02 2.930E+02 2.930E+02 2.930E+02 2.930E+02 2.930E+02 2.930E+02 2.930E+02 2.930E+02 2.930E+02 2.930E+02 2.930E+02 2.930E+02 2.930E+02 2.930E+02 2.930E+02 2.930E+02 2.930E+02 2.930E+02 2.930E+02 2.930E+02 2.930E+02 2.930E+02 2.930E+02 2.930E+02 2.930E+02 2.930E+02 2.930E+02 2.930E+02 2.930E+02 2.930E+02 2.930E+02 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00

Elapsed time is

HTRAN 3.047E+01 3.124E+01 3.181E+01 3.223E+01 3.253E+01 3.274E+01 3.289E+01 3.300E+01 3.306E+01 3.311E+01 3.313E+01 3.314E+01 3.314E+01 3.313E+01 3.311E+01 3.310E+01 1.172E+04 1.172E+04 1.172E+04 1.172E+04 1.172E+04 1.172E+04 1.172E+04 1.172E+04 1.172E+04 1.172E+04 1.172E+04 1.172E+04 1.172E+04 1.172E+04 1.172E+04 1.172E+04 1.172E+04 1.172E+04 1.172E+04 1.172E+04 1.172E+04 1.172E+04 1.172E+04 1.172E+04

23--------------------------------------------I DENSI 1.187E+00 1.187E+00 1.187E+00 1.187E+00 1.187E+00 1.187E+00 1.187E+00 9.000E+03 9.000E+03 9.000E+03 9.000E+03 9.000E+03 9.000E+03 1.000E+03 1.000E+03 1.000E+03 1.000E+03 1.000E+03 1.000E+03 1.000E+03

23 |--------------------------------------------I

HFLUX 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00

0.92

15-7

PROGRAM OUTPUT

Chapter 15

Example Output

Table 15-4: |-------------------------------------------| | STAR-CD VERSION 4.00.000 | | THERMOFLUIDS ANALYSIS CODE | | Operating System: Linux | | Stardate: 6-DEC-2005 Startime: 14:32:39 | |-------------------------------------------|

CASE NAME : out

*** Warning: Molecular weight of material in domain set to 28.96.

2 is

*** Reduction of solution matrix bandwidth enabled *** ______________________________MEANING OF PRINTOUT QUANTITIES________________________________________ ---NOMENCLATURE (mass balance) ------FVIN - total flow in through inlet boundaries, kg/s FVOUT - total flow out through outflow boundaries, kg/s FPIN - total flow in through pressure,stagnation,free-stream and transient-wave boundaries, kg/s FPOUT - total flow out through pressure,stagnation,free-stream and transient-wave boundaries, kg/s FCYIN - total flow in through partial cyclic boundaries, kg/s FCYOT - total flow out through partial cyclic boundaries, kg/s FLOUT - total flow out through outlet boundaries, kg/s SDRDT - mass accummulation by density change in time, kg/s SDVDT - mass accummulation by volume change in time, kg/s FLINJ - mass injection, kg/s MSDRO - mass transfer from dispersed phase (droplets) to continuous phase, kg/s FDIFF - mass balance kg/s SUM - sum of mass sources RESP - sum of absolute mass sources RES0 - starting residual in the solver ITERATION NUMBER = 1 -------------------------*** Warning: Residuals in eq. UMOM have reached round-off error limit in iteration 1 and trying to reduce them further can result in the solver divergence. Because of this further iterating is stopped. ______________________________BALANCE DATA________________________________________________ DOMAINWISE MASS BALANCE (kg/s) MAT. NO. PHASE NO. FDIFF TOTAL_FLOW_IN TOTAL_FLOW_OUT MSDRO (FVIN) (FPIN ) (FLOUT) (FVOUT) (FPOUT) 1 1 1.2811E-03 5.9337E-01 5.9209E-01 0.0000E+00 5.9337E-01 0.0000E+00 0.0000E+00 0.0000E+00 5.9209E-01 2 1 0.0000E+00 1.0000E+00 1.0000E+00 0.0000E+00 1.0000E+00 0.0000E+00 1.0000E+00 0.0000E+00 0.0000E+00 ------------ BOUNDARY REGIONWISE -----------REGION NO. TYPE FLOW-IN(kg/s) FLOW-OUT(kg/s) 1 INLET 5.9337E-01 0.0000E+00 2 PRESSURE 0.0000E+00 5.9209E-01 3 INLET 1.0000E+00 0.0000E+00 4 OUTLET 0.0000E+00 1.0000E+00

______________________________FIELD DATA_________________________________________________ *** FOR FLUID STREAM *** 1 Field Extrema: Umax Vmax 4.9233E+00 6.2337E+01 Umin Vmin -2.1391E+00 3.5014E+01

Wmax 0.0000E+00 Wmin 0.0000E+00

VMAGmax 6.2349E+01 VMAGmin 3.5014E+01

Pmax 5.5077E+00 Pmin 2.2318E-01

Field Volume-Averages: Pvav RHOvav 2.8710E+00 1.1867E+00

Tvav 2.9300E+02

TKEvav 1.3542E+00

EPSvav 2.2631E+01

TKEmax 6.5615E+00 TKEmin 3.9212E-02

EPSmax 2.6307E+02 EPSmin 1.3236E-01

Tmax 2.9300E+02 Tmin 2.9300E+02

RHOmax 1.1867E+00 RHOmin 1.1867E+00

Field Mass-Averages:

15-8

Version 4.02

Chapter 15

PROGRAM OUTPUT Example Output

Pmav 2.8710E+00

RHOmav 1.1867E+00

Tmav 2.9300E+02

TKEmav 1.3542E+00

Field Totals: Mass Volume 9.4939E-03 8.0000E-03

TKE 1.2857E-02

EPS 2.1485E-01

*** FOR FLUID STREAM ***

EPSmav 2.2631E+01

2

Field Extrema: Umax Vmax 1.4449E-02 1.5644E-01 Umin Vmin -7.3840E-03 5.7720E-02

Wmax 0.0000E+00 Wmin 0.0000E+00

VMAGmax Pmax 1.5644E-01 0.0000E+00 VMAGmin Pmin 5.8799E-02 -2.7495E-02

Field Volume-Averages: Pvav RHOvav -1.4465E-02 1.0000E+03

Tvav 2.9300E+02

TKEvav 0.0000E+00

EPSvav 0.0000E+00

Field Mass-Averages: Pmav RHOmav -1.4465E-02 1.0000E+03

Tmav 2.9300E+02

TKEmav 0.0000E+00

EPSmav 0.0000E+00

Field Totals: Mass Volume 8.0000E+00 8.0000E-03

TKE 0.0000E+00

EPS 0.0000E+00

*** FOR SOLID domain *** 3 Temperature data: Tmax = 2.9300E+02; Tmin

= 2.9300E+02; Tvav

TKEmax 0.0000E+00 TKEmin 0.0000E+00

EPSmax 0.0000E+00 EPSmin 0.0000E+00

Tmax 2.9300E+02 Tmin 2.9300E+02

RHOmax 1.0000E+03 RHOmin 1.0000E+03

= 2.9300E+02

Field Solver information: NSU = 1 NSV = 2 NSW = 0 NSP = 14 NSTE = 1 NSED = 1 NST = 0 CPU time is 0.13 Elapsed time is 0.33 __________________________________________________________________________________________

Version 4.02

15-9

Chapter 16

PRO-STAR CUSTOMISATION Set-up Files

Chapter 16

pro-STAR CUSTOMISATION pro-STAR provides four means by which users can customise the way they work with the program: • • • •

Set-up files Panels Macros Function keys

All are geared towards making problem data input faster and more flexible and can be used in combination with each other. The choice of which ones to use is largely a matter of user preference and the requirements of the model being built.

Set-up Files These files are read automatically as part of the pro-STAR start-up process and are used in creating a suitable pro-STAR environment for the problem in hand. The files have standard names, given below, and are located in a directory chosen by the user. On Unix systems, the path to this directory is stored in an environment variable (STARUSR) specified outside pro-STAR using the appropriate Unix environment setup command (see Chapter 17, “pro-STAR environment variables”). The available set-up files are as follows: 1. PROINIT — contains pro-STAR commands that are read and executed as the first action in the current session. This provides a convenient way of setting up (initialising) pro-STAR in a standard way (regarding, for example, plot type, viewing angle, etc.) every time a session begins. Some pro-STAR commands are in fact best used from within the PROINIT file. For example: (a) Command OPANEL — typically used to open a set of tools (standard pro-STAR GUI dialogs or user-defined panels) that the user wants on screen at the start of every new session. (b) Command SETFEATURE — reports or changes the byte ordering format of binary files to suit machines such as the Compaq Alpha range. This facility replaces settings previously made through environmental variables. 2. PRODEFS — this file is created automatically if the *ABBREVIATE command is used during the session. *ABBREVIATE enables one or more frequently used commands and their parameters to be joined together and executed in sequence, simply by associating them with an abbreviation name. The command group comes into action every time an existing ‘abbreviation’ is typed in the I/O window. File PRODEFS stores all current abbreviation definitions and, once created, may be used in all subsequent pro-STAR sessions. The file itself may be edited with any suitable text editor to add/modify/delete any particular abbreviation, as needed. 3. .Prostar.Defaults — a hidden system file containing definitions of function keys (see “Function Keys” on page 16-9), panel size and location (see “Panel definition files” on page 16-5) and ‘favourite’ panels (see “Panel Version 4.02

16-1

PRO-STAR CUSTOMISATION

Chapter 16

Panels

navigation system” on page 2-40). If the set-up file directory is not defined through STARUSR, pro-STAR creates default set-up files automatically in your current working directory.

Panels Panels are user-definable tools capable of simplifying the use of pro-STAR operations that are either not available in the existing GUI menus and dialog boxes or require additional functionality. Panels are often employed to facilitate the use of Macros, which are groups of commands that are saved in a separate file (see “Macros” on page 16-6). Macros can be assigned to Panel buttons so that a large number of commands can be executed simply by clicking such a button. Panel creation Panels can be created or modified by choosing Panels > Define Panel from the main menu bar to display the Define Panel dialog box shown below.

New panels are created by entering a name in the text box of the Define Panel dialog box and then clicking on the New action button. This results in the panel name being added to the list above the text box. Once this is done, the panel itself can be opened by • • •

double-clicking on its name in the list, or selecting the name in the list and then clicking on the Open action button, or clicking on Panels in the main menu bar and selecting the panel name from the drop-down list.

Any of the above actions will display a panel such as the one shown below.

16-2

Version 4.02

Chapter 16

PRO-STAR CUSTOMISATION Panels

Once the new panel has been opened, the user can specify its layout and define its buttons and menu items. The Panel Layout dialog box can be opened by selecting File > Layout from the panel’s menu bar.

The above dialog box allows definition of the number and layout of the panel buttons (a maximum of 100). Users may also specify menus for panels by selecting File > Menus from the panel’s menu bar. This opens the Define User Menus dialog box, shown below, where one can define up to six menus, their names and the pro-STAR commands that will be executed upon selecting a particular menu item. By default, a single menu called User 1 is defined containing a single menu item called Replot which executes command REPLOT.

Version 4.02

16-3


Chapter 16

Panels

Panel button names and definitions are assigned by first selecting a button, and then entering a new name or definition into the appropriate text box. A button’s definition is the pro-STAR command(s) that will be executed when the button is pushed. The following three examples illustrate the way in which frequently repeated operations may be simplified by assigning them to panel buttons: Example 1 Select a number of cells with the screen cursor and then refine them by a factor of 2 in all directions. Assign to option button CCREF.

Example 2 Select a range of fluid cells by drawing a polygon around them, change them to solid cells and then plot the mesh. Assign to option button CZMOD.

16-4

Version 4.02

Chapter 16

PRO-STAR CUSTOMISATION Panels

Example 3 Display vertex coordinates in local coordinate system 2 by pointing at the required vertex with the cursor. Assign to option button VCOOR2.

To select a button without executing the corresponding button definition, move the mouse pointer to the button and press (but do not release!) the mouse button. Next, move the mouse pointer clear of the button and then release the mouse button. This sequence will set the newly selected panel button as the active button, but will not execute the button function. Note that selecting File > Reload from the panel’s menu bar will cancel out any changes made to the panel definition since it was last saved. Panel definition files A panel’s button and menu settings as well as its size and location are saved in a panel definition file when File > Save is selected from the panel’s menu bar. This file is created using the panel name specified by the user in the Define Panel dialog box and the suffix .PNL. The file location depends on its name. If the name entered was prefixed with the letter L or G (note that a space must be typed after each letter), the file will be placed in directory PANEL_LOCAL or PANEL_GLOBAL, otherwise it will be put in your current working directory. On Unix systems, the local and global directory names are stored in environment variables that can be set outside pro-STAR using the appropriate Unix environment setup command (see Chapter 17, “pro-STAR environment variables”). The environment variables can also be set within pro-STAR by selecting Panels > Environment from the main menu bar. This displays the Set Environment dialog box shown below, which allows entry of local and global directory names in the corresponding text boxes. Version 4.02

16-5


Chapter 16

Macros

Command:

SETENV

Note that a list of available panels can be viewed by opening the Define Panel dialog box. Panels found in your current working directory are shown in the list with a ‘.’ before the panel name. Any panel definitions found in the directories specified by the PANEL_LOCAL and PANEL_GLOBAL variables are shown in the list with an L or G prefix before the panel name, respectively. Once added to the list, a panel can be opened in a number of ways, as described in “Panel creation” on page 16-2. Note that panels can also be opened from the pro-STAR input/output window by typing OPANEL, PANEL but this command is more typically issued from within the PROINIT set-up file (see “Set-up Files” on page 16-1). In addition to the panel definition file, a panel’s size and location are also saved in a hidden system file called .Prostar.Defaults (see “Set-up Files” on page 16-1). Definitions stored there have priority over the size and location information stored in the panel definition file. This enables you to override such information if the panels are located in a directory for which you do not have write permission. Panel manipulation The Define Panel dialog box provides additional facilities for manipulating panels, as follows: •

• • •

The Re-Scan button recreates the list of available panels. Those that were removed from the list will re-appear, while those created via the New button but never saved will disappear. The Copy button creates new panels by copying an existing panel definition file to another file whose name must be typed in the text box. The Rename button changes the name of a panel definition file to another name typed in the text box. The Delete button allows you to remove panels from the list but does not delete the corresponding definition files. The latter can only be deleted outside pro-STAR by using the appropriate operating system command.

Macros A macro is a set of user-defined commands that can be executed at any stage of the pro-STAR session. The constituent commands must be stored in a special file, identified by a ‘.MAC’ extension and included within 16-6

Version 4.02

Chapter 16

PRO-STAR CUSTOMISATION Macros

• • •

the current working directory, or a pre-defined local macro directory, or a pre-defined global macro directory.

As with panel directories, the local and global directory names are stored in environment variables MACRO_LOCAL and MACRO_GLOBAL that can be set outside pro-STAR using the appropriate Unix environment setup command (see Chapter 17, “pro-STAR environment variables”). The environment variables can also be set within pro-STAR by selecting Panels > Environment from the main menu bar. This displays the Set Environment dialog box which allows entry of local and global macro directory names in the corresponding text boxes. Macros can be created, renamed, copied, and deleted in the Define Macro dialog box in the same way that panels are in the Define Panel dialog box. The Define Macro box, shown below, is opened by choosing Panels > Define Macro from the main menu bar. The name of a new macro must be typed in the text box. An existing macro can be selected and displayed, by double-clicking its name in the macro list. Several macros can be displayed simultaneously in multiple windows, by highlighting them in the list with the mouse and then clicking the Open button. pro-STAR looks for macro files in three places. Macros found in the user’s current working directory are shown in the list with a ‘.’ in front of the macro name. Those found in the directories specified by the MACRO_LOCAL and MACRO_GLOBAL environment variables are shown with an L or G prefix before the macro name, respectively.

Clicking the Open or New button in the Define Macro box opens a macro editor to display the macro file(s) that has been selected in the macro list (or a blank sheet for new macros), as shown below. The user can then type in the required pro-STAR commands or amend existing ones. Command PROMPT, which displays messages in the area underneath the plotting window (see “Main window” on page 2-15) is particularly useful inside a macro as it can prompt the user to, say, supply required data or to click an appropriate menu item with the mouse.

Version 4.02

16-7


Chapter 16

Macros

The macro editor facilities are arranged under three menus in the editor’s menu bar: 1. File (a) (b) (c) (d) (e)

Open — open another macro Save — save the current changes Save As — save the current changes to a different macro file Clear All — clear the editor window Quit — terminate the editing session

2. Edit (a) Find — find a character string typed in the dialog box shown below:

(b) (c) (d) (e)

16-8

Mark Selection — mark the selected characters for subsequent searches Find Selection — find the selected characters in the macro body Find Again — repeatedly find the selected characters Replace — find a character string and replace it with another string. Both strings are typed in the dialog box shown below:

Version 4.02

Chapter 16

PRO-STAR CUSTOMISATION Function Keys

3. Execute (a) Execute Macro — execute the whole macro (b) Execute Selection — execute only the highlighted lines in the editor window As with panels, the Define Macro dialog box provides additional facilities for manipulating macros, as follows: • •

• • •

The Execute button executes the selected macro. The Re-Scan button recreates the list of available macros. Those that were removed from the list will re-appear, while those created via the New action button but never saved will disappear. The Copy button creates new macros by copying an existing macro file to another file whose name must be typed in the text box. The Rename button changes the name of a macro file to another name typed in the text box. The Delete button allows users to remove macros from the list of available macros but does not delete the corresponding files. The latter can only be deleted outside pro-STAR by using the appropriate operating system command.

Note that panel buttons are often used to execute macros, by setting the button definition to issue command *macro,exec This assignment can be made as follows: • • • • • •

Open the Define Macro dialog box and highlight a macro in the list. Open the Define Panel dialog box, select a panel from the list and display it by double-clicking it. Click on a free button in the panel. Select Assign from the panel’s Macro menu. This assigns the macro name to the button and generates the appropriate *MACRO command. If necessary, select Edit from the panel’s Macro menu to open the macro text editor discussed above and type in any further changes Save all changes by selecting Save from the File menus of both macro and panel editors before closing their corresponding dialog boxes

Function Keys Users can program the keyboard function keys (F2 - F12) to execute pro-STAR Version 4.02

16-9


Chapter 16

Function Keys

commands or macros. This is done by choosing Utility > Function Keys from the menu bar to display the Edit Function Keys dialog box shown below.

Any valid pro-STAR command (or set of commands if a $ character is used to separate them) can be mapped to individual function keys by typing it in the appropriate text box. Command parameters such as ‘VX’ or ‘CX’ may be used and will be interpreted in the normal way. Command strings are limited to 80 characters in length. In addition to standard pro-STAR commands, the function keys can also be used to repeat the last executed command and to open dialog boxes. Thus: • •

16-10

Command repeat will literally repeat the last command executed, including parameters such as ‘VX’ or ‘CX’. Command string open dialog1,dialog2,... will open the dialog boxes or tools specified. Available items are:

Version 4.02

Chapter 16

PRO-STAR CUSTOMISATION Function Keys

Name ANIM BLIS BLLI BLOC CELL CHEC CHEM CLIS COLO CONT COUP CSYS DROP FORE GENE GRAP GRDI GRLO GRRE POST PROP SPLI SPLL STAR TRAN VERT VLIS

Description Animation Module Boundary List Block List Block Tool Cell Tool Check Tool Chemical Module Cell List Colour Tool Control Module (unsupported panel) Couple Tool Coordinate Systems panel Droplets panel Convert Foreign Formats panel Convert Generic panel Graph Tool Graph Module Load Graph Registers panel Graph Registers panel Post Register Data List Property Module (unsupported panel) Spline Tool Spline List Convert STAR panel Transient Module Vertex Tool Vertex List

The default function key definitions are: F5 – repeat F6 – replot F7 – cplot F8 – zoom,off $replot Note that the F1 key is reserved for displaying context-sensitive, on-line Help information on pro-STAR commands (see “Getting On-line Help” on page 2-35) Any changes to the function key definitions are saved in a file called .Prostar.Defaults (see “Set-up Files” on page 16-1) at the location specified by environment variable STARUSR (or in your current directory, see page 16-1). This file can be modified either through the Function Keys dialog box within pro-STAR or outside it via any suitable editor. Users may find it useful to keep a single .Prostar.Defaults file in the STARUSR location so that the particular setup that they define is available for any pro-STAR session.

Version 4.02

16-11

Chapter 17

OTHER STAR-CD FEATURES AND CONTROLS Introduction

Chapter 17

OTHER STAR-CD FEATURES AND CONTROLS

Introduction This chapter describes some of the less commonly used features and controls in STAR-CD and covers the following topics: • • • •

File organisation, naming conventions and general utilisation Special pro-STAR and STAR features and settings The StarWatch utility Hard copy production

File Handling Naming conventions At every session, pro-STAR creates a set of files whose names are based on a user-supplied model name or case name. Each file name is of the form case.xxxx, where xxxx is a three- or four-character filename extension. Thus, if the case in question is called test, then all its associated files will be called test.ccm, test.mdl, etc. and will be used for the appropriate input/output operation during the model building and numerical solution processes. You should always supply a case name at the beginning of a pro-STAR session (see “pro-STAR Initialisation” on page 2-12). A case name may be overridden at any time during a pro-STAR session by choosing File > Case Name from the menu bar. This displays the Change Case Name dialog shown below:

Command:

CASENAME

Supply a new case name (up to 70 characters long) in the text box provided. This changes the default file name but does not affect any files that are already open. It also determines which files will be used during subsequent file operations. Note that the names of the input and output restart (.ccm) files will be reset by this operation. Commonly used files A few key files are always read and/or written to by pro-STAR, whereas the majority are opened and accessed only in response to a command or a GUI operation. These key files are described below:

Version 4.02

17-1


Chapter 17

File Handling

Echo file (.echo) Used exclusively by pro-STAR and is always opened. It holds a copy of every command typed by the user or, for GUI operations, their command equivalents, as generated automatically by pro-STAR during the session. The file can be: • • •

Reviewed Used for recovery purposes (see “Error messages” on page 2-19) Copied to a temporary file which can be subsequently edited to make changes to the recorded commands (see item 12 on page 17-10). Once the editing process is complete, the modified command file can be replayed into pro-STAR using the editor’s file execution facilities (or by typing command IFILE).

Model file (.mdl) Used exclusively by pro-STAR. Choosing option File > Save Model from the menu bar instructs pro-STAR to write a full description of your model to this file, using the specified case name as the file name. It is advisable to save data regularly during a session so as to minimise the chance of losing large amounts of information due to user error or system failure. Note that every time you save the model file, its previous version (i.e. the model you started out with before making any changes) is also automatically stored as a backup, in a file of form case.bak If you need to save the .mdl file under a name other than the case name, choose option File > Save As from the menu bar. This displays the Save As dialog, shown below, which allows the name to be typed exactly as required. Alternatively, an existing file may be selected by utilising pro-STAR’s built-in file browser facilities (see page 17-9).

Command:

SAVE

Option File > Resume Model performs the reverse operation, i.e. it instructs pro-STAR to read a model description from an existing .mdl file. If you need to resume from a .mdl file that does not have the same name as the case name, choose option File > Resume From from the menu bar. This displays the Resume From dialog shown below, which allows the name to be typed exactly as required. Alternatively, the file may be selected by clicking the browser button provided and utilising pro-STAR’s built-in file browser facilities (see page 17-9).

17-2

Version 4.02

Chapter 17

OTHER STAR-CD FEATURES AND CONTROLS File Handling

Command:

RESUME

Problem setup (command) file (.inp) Although the model file is normally saved in binary form, there may be occasions when you need to write the model data in text (coded) form. Examples of such instances are: •

• •

To allow you to quickly produce a set of coded pro-STAR input files that will re-create the case as defined in the model file. This is especially useful if you want to set up several runs with parametric changes and then submit the job in batch. To enable you to find out which commands would activate certain features present in your model. To facilitate testing of models that were created with a previous version of pro-STAR.

To write model data in text form, choose File > Save As Coded from the menu bar to display the CDSave dialog shown below:

Command:

CDSAVE

The dialog uses default file names with extensions .inp, .cel, .vrt, and .bnd for four files that will contain problem set-up, cell, vertex and boundary information, respectively. Alternative names for any of these files may be entered in the boxes provided. For moving mesh cases, event definitions (see “Moving Meshes” on page 12-9) can also be written to file .evnc. For cases containing droplets, an additional droplet data file (.drpc) is created. Once the files have been copied to a suitable directory, the model may be re-activated by choosing File > Read File from the menu bar. This will display the Version 4.02

17-3


Chapter 17

File Handling

Input Coded File dialog shown below:

Command:

IFILE

Check that the file shown in the File Name window is correct and then click Apply. All data in files .cel, .vrt, .bnd, .drpc (if present) and, for moving mesh cases, .evnc will be read in automatically. Transient history file (.trns) This is used exclusively by pro-STAR for transient problems specified by means of load steps (see “Load-step based solution mode” on page 5-6) and contains all additional information (changes to boundary conditions, number and length of time steps, etc.) needed for such problems. You must make this file available to your current session before changing or adding data concerning the analysis. This is done via the Advanced Transients dialog (see “Load step controls” on page 5-10), or by typing command TRFILE. The file is normally written in binary form but a facility also exists for writing it instead in text (coded) format and to a file with extension .trnc. This is done by selecting Modules > Transient from the menu bar to display the Advanced Transients dialog, specifying the file name in the box provided at the bottom of the dialog and then clicking Apply. Alternatively, use command CDTRANS. If an existing file needs to be used, pro-STAR’s built-in file browser can help locate it. Plot file (.plot) This is used exclusively by pro-STAR and is always open to receive neutral plot information, i.e. machine-independent representations of a set of plots. The file may be written in either binary or text (coded) format. CD-adapco supply source code for several decoding programs that drive hard-copy devices in a variety of formats (e.g. Postscript), or screen output devices (e.g. X-window workstations). These programs can also serve as templates for constructing plot drivers for other, unsupported devices. To make use of the neutral plot facility: • •

Specify the plot file name (if other than case.plot) and type (if not CODED) using command NFILE. Switch the plot output from the terminal or workstation to the plot file by choosing item Plot > Plot To File from the menu bar (or use command TERMINAL in the form TERMINAL,,FILE

• 17-4

Perform the plotting operations required, as normal. Graphical output is now Version 4.02

Chapter 17


•

diverted to the file instead of being displayed on the screen. To restore normal operation, choose option Plot To Screen from the Plot menu.

Details of data representation in the neutral plot file can be found in the Post-Processing User Guide, Appendix B. Data repository file (.ccm) This file has a special format that facilitates different sets of information to be stored in it. It is written and read by both pro-STAR and STAR in the following ways: 1. pro-STAR saves all cell topology and model geometry information in the file once mesh building is complete. Cells are defined as a collection of faces, i.e. a general polyhedral cell definition is used regardless of the actual cell shape. The file must be rewritten whenever (a) the mesh geometry is modified (b) boundaries are added, subtracted, or assigned to different boundary regions (c) the cell type definitions are changed The file is created by selecting File > Write Geometry File from the menu bar to display the Geometry File Write dialog shown below:

Command:

GEOMWRITE

The input required is: (a) File Name — enter a name in the text box provided or click the adjacent button to select an existing file using pro-STAR’s built-in file browser (see page 17-9) (b) File Type — select the file format according to the solver (CCM, CEDRE, BAE) for which the geometry file is intended (c) Geometry Scale Factor — an optional scale factor applied to all dimensions of the problem’s geometry (d) Write Backup File — define the action to be taken if the specified repository filename already exists in your current directory: i) Backup — the existing file is renamed as casename.ccm.bak Version 4.02

17-5


Chapter 17

File Handling

(or casename.ccm.bak1 if casename.ccm.bak already exists, or casename.ccm.bak2 if casename.ccm.bak1 already exists, etc.). ii) No Backup — the geometry information is overwritten 2. STAR’s action is one of the following: (a) For initial runs, STAR reads the geometry information contained in the file and then appends the solution results at the end of the analysis. (b) For restart runs, STAR reads in addition the results of a previous (partially converged, interrupted or transient) analysis before starting the new solution. The new analysis results then overwrite the previous ones at the end of the run. (c) If solution residual values are required as part of the analysis, STAR also stores them in this file. 3. Apart from storing problem geometry data, pro-STAR will also (a) read the file for post-processing purposes, i.e. to make contour, vector or graph plots of any variable calculated by STAR. (b) write mapped solution data when an existing mesh is refined (see Chapter 5, “Solution Control with Mesh Changes” in this volume) Each set of data stored in the repository file is called a “state”. The table below lists the names of the states and the data stored in them. The available states in a given file may be displayed using command CCLIST. State name

Type of data

default

Problem geometry for sequential runs Problem geometry for parallel runs Solution (Restart) data Mapped Solution (Restart) data Field residuals

geom_par Restart_1 smap Residue_n n=1,2,3,...

STAR-CD 3.2X equivalent file .geom .geom (decomposed) .pst .smap .rpo

Problem data file (.prob) This is written by pro-STAR and read by STAR. It contains information on what kind of analysis is to be performed and what data are to be printed or saved for post processing. It also contains all material property values, solution control settings, boundary conditions and initial conditions. It is written independently of the geometry file and should be rewritten every time any of the above model parameters is modified. The file is created by selecting File > Write Problem File from the menu bar to display the dialog shown below. Note that a filename other than the default 17-6

Version 4.02

Chapter 17


(casename.prob) may be entered if necessary in the text box provided.

Command:

PROBLEMWRITE

Transient post data file (.pstt) This is written by STAR and contains selected transient analysis data at pre-defined points in time (see “Output controls” on page 5-12). It is used by a subsequent pro-STAR post-processing run to make contour or vector plots based on the selected data. Note that the file holds only part of the available information on the model, so it cannot be used for restarting the analysis; that function can be performed only by using the data repository (.ccm) file. File relationships The use and relationship between files in the STAR-CD environment is illustrated by Figure 17-1. Appendix B in this document contains a complete list of all files that can be written or read by either pro-STAR or STAR. The same information may also be displayed on-line in the Help dialog (choose Help > pro-STAR Help from the menu bar, select Misc. from the Module pop-up menu, and then highlighting item FILE). For the great majority of problems, however, only the files shown below are ever needed.

Version 4.02

17-7


Chapter 17

File Handling

case.trns Transient history data

pro-STAR

case.prob Boundary conds. Solution params.

case.mdl Model data case.echo Command echo case.plot Neutral plot

case.ccm Geometry

STAR

case.pstt Transient output data

case.ccm Solution data

Figure 17-1

STAR-CD file use

In addition to solution and transient post data files, pro-STAR provides a utility for converting solution monitoring and droplet track data files to coded (text) format and vice versa. This is useful, for example, in manipulating and displaying the data outside the pro-STAR environment or for checking the validity of the file contents. The utility allows conversions between a variety of formats and is accessed by selecting Tools > Convert > Post from the menu bar. This activates the Post Convert dialog shown below:

Commands:

SMCONVERT

PTCONVERT

You may then •

• 17-8

select option Solution Monitoring or Particle/Droplet Track depending on the file type you wish to convert. The first option deals with residual or solution monitoring data conversion (see Chapter 5, “Output controls”), the second with droplet track data conversion (see “Trajectory displays” on page 9-8) or particle track data conversion (see “The Particle Track File” on page 7-6 of the Post-Processing User Guide) enter the name of the file containing the data to be converted (Input File with Version 4.02

Chapter 17


• • •

extension .rsi or .trk) or select it using pro-STAR’s built-in file browser (see page 17-9) choose the file type (normally Binary) from the available options in the adjacent pop-up menu enter the name of the file that will store the converted data (Output File with extension .rsic or .trkc) choose the file type (normally Coded) from the available options in the adjacent pop-up menu

The above operation may also be performed in reverse, i.e. converting the text file back to binary format, using the same dialog but with Input now being Coded and Output being Binary, plus a reversal of the file name extensions. In the course of a session pro-STAR also opens several scratch files. These are opened automatically and deleted at the end of the session. Their use is normally transparent except when their size exceeds the amount of free space on your disk. While some scratch space is used for hidden-line plotting, the largest amount is needed while the geometry (.ccm) file is being written. The space used varies linearly with the number of vertices present and the maximum number of cells connected to any single vertex. File manipulation The file-manipulation related capabilities of pro-STAR are as follows: 1. Finding files — If you are not sure of the exact location or name of an existing file, use pro-STAR’s file browser facility. This is activated by clicking the browser button

included in numerous GUI dialogs. The button displays the File Selection dialog shown below:

The scroll lists and filters included in the above dialog allow easy navigation through various levels of sub-directories until the required file is located. Version 4.02

17-9


Chapter 17

File Handling

2. Switching program input from a terminal (or standard input) to any disk file (of form case.inp) containing pro-STAR commands, and vice versa. This can be done at any time during a session, using either the pro-STAR editor’s Execute menu options (see item 12 below) or by typing command IFILE. In the latter case, input switches back to the terminal automatically at the end of the specified file. The command also supports a ‘nesting’ capability, i.e. the new input stream can itself contain IFILE commands that will direct input to yet another source file and so on. 3. Switching output from a terminal screen (or standard output) to a disk file (of form case.out), and vice versa. This can be done at any time during a session using command OFILE. Using parameter NONE with this command turns the output off completely. The facility enables you to save lists of various pro-STAR items, for use in other programs or for later review. 4. Writing the geometry file (see “Data repository file (.ccm)” above) by choosing File > Write Geometry File from the menu bar. 5. Writing the problem data file (see “Problem data file (.prob)” above) by choosing File > Write Problem File from the menu bar. 6. Restoring a previously created model from a saved model file (see “Model file (.mdl)” above) by choosing File > Resume From... from the menu bar. When used for the first time in a pro-STAR session, RESUME will also automatically read and execute commands stored in a special file called PROINIT (see “Set-up Files” on page 16-1). This provides a very convenient way of setting up pro-STAR in a standard way (regarding, for example, plot type, viewing angle, etc.) every time a session starts. 7. Saving the current model description in binary format to file .mdl, as described above, by choosing File > Save Model from the menu bar. 8. Saving the model description in text (coded) format, as described above, by choosing File > Save As Coded from the menu bar. 9. Repositioning a previously used file (including a pro-STAR macro file) to its starting point by typing command REWIND. 10. Closing a previously used file by typing command CLOSE. The command may also close all currently open files. 11. Printing a summary of all currently open files by typing command FSTAT. 12. File editing via pro-STAR’s built-in editor — This is activated by choosing File > Edit File from the menu bar to display the panel shown below. Files that may be conveniently manipulated using this editor are: (a) Command files — these allow execution of a set of pre-recorded pro-STAR commands. As noted in the section on “Commonly used files” on page 17-1, a common source for them are echo files from previous pro-STAR sessions. To avoid problems, however, an echo file should be copied and renamed before using it as a command file. (b) User subroutine files — these contain special user-supplied FORTRAN code and are discussed in detail in Chapter 14.

17-10

Version 4.02

Chapter 17


The available facilities are arranged under three menus in the editor’s own menu bar, as follows: File (a) Open — open a specified file. This activates the File Selection dialog shown on page 17-9, enabling the required file to be located. (b) Save — save the current changes. (c) Save As — save the current changes to a different file. The dialog box above re-appears to aid specification of the destination file location. (d) Clear All — clear the editor window. (e) Quit — terminate the editing session. Edit (a) Find — find a character string, typed in the dialog box shown below.

(b) (c) (d) (e)

Version 4.02

Mark Selection — mark the selected characters for subsequent searches. Find Selection — find the selected characters in the file body. Find Again — repeatedly find the selected characters. Replace — find a character string and replace it with another string. Both strings are typed in the dialog box shown below.

17-11


Chapter 17

Special pro-STAR Features

Execute (Command files only) (a) Execute All — execute all commands in the file. This is equivalent to typing command IFILE in pro-STAR’s Input window. (b) Execute Selection — execute only the highlighted lines in the editor window. In addition, the usual keyboard- or mouse-driven cut, copy and paste functionality is also available with the editor window.

Special pro-STAR Features pro-STAR environment variables pro-STAR uses the values of several environment variables. Some specify the path to various system directories while others control the operation of the system. You should ensure that these values are correctly set before using STAR-CD. The syntax for setting environment variables depends on the shell program you are using (if in doubt type the command echo $SHELL). The current list of such variables is as follows: MACRO_LOCAL and MACRO_GLOBAL Paths to the local and global pro-STAR macro directories, respectively (see “Macros” on page 16-6) PANEL_LOCAL and PANEL_GLOBAL Paths to the local and global pro-STAR panel directories, respectively (see “Panel definition files” on page 16-5) STARBROWSER (not needed for Windows ports) Path to the user’s choice of Internet browser (Netscape or IE) that will be launched from the pro-STAR Help menu (see page 2-37). The user’s search path must be amended to include the directory defined by this variable. The default is to run Mozilla from your current working directory. STARFONT0 / starfont0 Font name and size to use for plot title, plot legend, graph title and main axes label (see the description of command TSCALE in the Commands volume) STARFONT1 / starfont1 Font name and size to use for the contour and vector scales (see the description of command TSCALE in the Commands volume) STARFONT2 / starfont2 Font name and size to use for the secondary contour and vector scales (for droplets and particle ribbons; see the description of command TSCALE in the Commands volume) STARFONT3 / starfont3 Font name and size to use for entity numbers (NUMBER command), x and y tick labels on graphs and local coordinate system axes (see the description of command 17-12

Version 4.02

Chapter 17

OTHER STAR-CD FEATURES AND CONTROLS Special pro-STAR Features

TSCALE in the Commands volume) Note: Variables STARFONT 0-3 described above apply only to X-window plotting. They have no effect on OpenGL based plotting as the fonts system there is entirely different. STAR_TCL_SCRIPT Path to the location of file STARTkGUI.tcl, containing a user-supplied Tcl/Tk script (see “The Users Tool” on page 2-35) STARUSR Path to pro-STAR files PRODEFS (abbreviations), PROINIT (initial set up) and .Prostar.Defaults (see “Set-up Files” on page 16-1) Resizing pro-STAR pro-STAR is a dynamic-memory executable code and requires a file called param.prp to be present in your current working directory. The file contains a list of parameters that determine the data size of the executable on start-up. If this file is missing, incomplete, or out-of-date, pro-STAR will automatically write a new local param.prp based on the values in the model (.mdl) file being read, and also on any values that could be read from an existing param.prp. This happens the first time pro-STAR is run using the prostar script described in Chapter 2, “Running a STAR-CD Analysis”. It is almost always necessary to resize the pro-STAR executable to cater for special problems (such as moving mesh problems) or to accommodate cases with a larger number of cells, vertices, etc. (or a smaller number, if you are having problems with available memory in your machine). In any of the above situations, file param.prp should be modified but this should never be done using a text editor. Rather, a new version of the file containing parameters of the right magnitude must be created in one of the following ways: 1. By running the prosize script. This is accessed by typing prosize The script first asks whether you want to modify some of the parameters in the current file or create a brand new param.prp. You may also exit here without modifying or creating any files. If continuing, prosize asks: Is your mesh primarily hex or tet? (Answer H or T) (The T option should be chosen for wholly or predominantly tetrahedral meshes; H is appropriate for all others, including meshes containing trimmed cells) After this, the script prompts you to specify the values of the parameters to be stored in param.prp. A carriage return instructs the script to use the indicated default value, while entering -1 will terminate the script and use the remaining defaults to write param.prp. The most usual variation from the default values is in the maximum number of cells (MAXCEL) and vertices (MAXVRT). Otherwise, the default values suggested by prosize should be sufficient for most cases. Version 4.02

17-13


Chapter 17

Special pro-STAR Features

2. By issuing command MEMORY from within the pro-STAR session. This command can be used only to increase the parameter sizes. If during the session it is found that the value of any sizing parameter(s) is insufficient, a warning message will appear in the I/O window. pro-STAR will sometimes be able to adjust the parameter value(s) automatically and then continue. However, in most cases you will be prompted to enter an appropriate new value for the indicated parameter(s) using MEMORY, after which you may continue as normal. Either way, the parameter values are changed internally without changing the param.prp file. To use the new values in future pro-STAR sessions, you will need to save them explicitly via a MEMORY,WRITE instruction. This will rename the existing param.prp file as param.bak and write the new parameters into a new param.prp file. Note that, after running pro-STAR with a given model, it is possible to clear all model parameters (i.e. delete all cells, vertices, boundaries, etc.) but leave the current memory size intact. This is done using command WIPEOUT and is useful if you want to abandon the current model and start a new one from scratch without exiting from pro-STAR. Furthermore, option MEMORY of this command will also reset the pro-STAR executable back to the size given in the param.prp file. Special pro-STAR executables On occasion, you may need to use a user-defined subroutine file, user1.f. This option refers to subroutines that work in conjunction with pro-STAR, not STAR, and is not supported in Windows ports at present. In such a case, the required special pro-STAR executable may be created using script prolinkl. This is accessed by typing prolinkl The script looks for a file named user1.f in the current directory. That file will be compiled into object code (user1.o) and converted into a dynamically-loaded shared object (.so or .sl or .dll depending on the operating system). The directory with the shared object must be added to the shared object library path (usually LD_LIBRARY_PATH) in order to be found and used by any subsequent pro-STAR runs. prolinkl will advise the user on how to create this path for the given operating system. Use of temporary files by pro-STAR Choosing the location of temporary files You can control the location of most pro-STAR temporary files for POSIXcompliant computers. You should ensure temporary files reside where there is sufficient capacity and where they can be accessed quickly. In practice, this means on a fast hard disk on the same computer as that doing the calculations (rather than on a remote disk accessed through a local area network). Note that the usual location for Unix temporary files, a directory called /tmp, often has insufficient capacity for pro-STAR’s temporary files. You select the location of temporary files by setting an environment variable, named TMPDIR, to the path name of the directory where pro-STAR should write the temporary files. 17-14

Version 4.02

Chapter 17

OTHER STAR-CD FEATURES AND CONTROLS The StarWatch Utility

Deleting temporary files Take care not to delete pro-STAR’s temporary files during a calculation; it will crash if you do. pro-STAR may leave temporary files behind if it crashes or you halt its execution. For POSIX-compliant systems, the operating system automatically deletes most temporary files if pro-STAR halts or crashes. For other systems, you might have to manually delete abandoned temporary files after a crash or halt.

The StarWatch Utility This is a free-standing utility that enables you to monitor the progress of a selected STAR job running anywhere in your computer network. The monitoring is done from a special window opened by StarWatch, as shown below.

Specific advantages of StarWatch are: • • • •

You can monitor progress of a number of separate STAR jobs The jobs may be running on any machine in your network, including your own You may select the variables whose solution progress you wish to monitor You may adjust the display characteristics (e.g. scaling) of the monitored variables

Running StarWatch By default, StarWatch uses ports 6200 to 6206 to establish communication between the STAR executable, the StarWatch daemon (a communication program) and StarWatch, the display program that runs on your screen. If ports 6200 — 6206 are acceptable, then no further setup is required. If they are not, perhaps because they Version 4.02

17-15


Chapter 17

The StarWatch Utility

conflict with other programs using those ports, they can be set to any ports that the user (or more likely) system administrator wants to use. The only proviso is that if the ports are changed on one system, the same change must be made for all systems for which StarWatch communication is required. If the defaults are not acceptable, then an administrator must edit the /etc/services file and add the following lines: star-chartd star-chart1 star-chart2 star-chart3 star-chart4 star-chart5 star-chart6

6200/tcp 6201/tcp 6202/tcp 6203/tcp 6204/tcp 6205/tcp 6206/tcp

# # # # # # #

Star/Stripchart client/server daemon Local Stripchart 1 Local Stripchart 2 Local Stripchart 3 Local Stripchart 4 Local Stripchart 5 Local Stripchart 6

where port numbers 6200 — 6206 can be replaced by any set of port numbers. Step 1 If using the STAR GUIde environment to run a numerical analysis interactively, StarWatch will start automatically as soon as STAR itself begins execution and will open a monitoring window like the one shown above (see Chapter 2, “Running a STAR-CD Analysis”, Step 6). If you are not using STAR GUIde, or if you want to monitor the progress of another currently active job, you may open the StarWatch window explicitly by following the steps below: • • •

Open a new window on your computer or go to an existing one Type starwatch, then send this application to the background also. The StarWatch application panel should appear on your screen. Start your STAR job in the same window using the -watch option. Note that when running a parallel job, the -watch option must precede any other options used.

Note that you may also start STAR first and then StarWatch. Step 2 Go to the StarWatch panel and select option Host from the Connect menu. Choose the name of the machine running your job in the Select Host dialog shown below and click OK.

Note that: • • •

17-16

Only STAR jobs owned by you and only those that have registered with the StarWatch daemon can be selected Registration usually takes place roughly at the end of the first iteration If STAR cannot find the daemon, it will keep trying for a small amount of time and then continue without trying further contact. Version 4.02

Chapter 17


Now choose the PID of the STAR job you wish to monitor from the list displayed in the Select STAR Job dialog and click OK.

StarWatch should now start displaying the monitored flow variables against iteration number or time step. Choosing the monitored values The following choices are available: 1. Material (stream) number In multi-stream applications, select the stream you wish to monitor using the Material Number slider control. 2. Field or residual values Select the type of variable to be displayed by clicking the toggle button at the bottom of the Legend section. The button label changes from Plot Field Values to Plot Residual Values and vice versa, depending on your choice. The labelling and scale of the adjacent graph also changes accordingly. 3. Monitored variable Choose the flow variables to be monitored, in terms of either field or residual values, by clicking the option buttons next to the variable names. The latter appear in the Legend section under the Property column and comprise the three velocity components, turbulence kinetic energy and dissipation rate, pressure and temperature. The colour used to display each variable is shown next to the name. It is also possible to monitor changes in scalar variables, if present in your model, by selecting View > Selected Data > Scalar Variables from the menu bar. The contents of the Legend section and the graph labelling will change accordingly. The method of selecting scalars is the same as for the main (global) variables. Note that since only seven quantities can be monitored, option View > Select Scalars lets you decide which scalars you want to look at; by opening a secondary (Select Scalars) dialog in which the required scalars and the order in which they appear in the StarWatch display may be determined. Controlling STAR At the beginning of a numerical analysis, STAR reads all files prepared for it by pro-STAR. Many of the parameters set in pro-STAR can be viewed and altered dynamically while the solution is in progress by selecting Settings > STAR Control Variables from the StarWatch menu bar. This brings up the Star Control Variables dialog shown below: Version 4.02

17-17


Chapter 17


The dialog’s purpose is to allow the user to interactively change the values of several STAR solution and output control parameters. These are grouped into six tabs according to function, as shown above, and all act in the same way. The meaning of the available parameters is listed in the table below: Parameter

Meaning

General Settings DT Time step size MAXCOR Maximum number of correctors for the PISO algorithm RESOC Residual tolerance for the PISO algorithm SORMAX Overall convergence criterion IJKMON Monitoring cell number for fluid domains File Output ECHO =.T. Control information will be written to file .info BOECHO =.T. Boundary data will be written to file .info ITEST =.T. Write all conservation balance information to file .info IRESI =.T. Write all solver convergence information to file .info NDUMP Frequency of writing data to file .ccm NFSAVE Backup frequency (frequency of saving file .pst_iternum) NCRPR Number of cell Courant numbers (starting from the largest) to be printed out NFRRE Iteration frequency for dumping residuals to file .ccm 17-18

Version 4.02

Chapter 17


Parameter

Meaning

Under-Relaxation Factors FPCR Under-relaxation factor for pressure correction (PISO) FUVW Under-relaxation factors for velocities FP Under-relaxation factor for pressure FTE Under-relaxation factors for k and ε FT Under-relaxation factor for temperature FTVS Under-relaxation factor for turbulent viscosity FDEN Under-relaxation factor for density FLVS Under-relaxation factor for laminar viscosity FCON Under-relaxation factor for heat conductivity FR Under-relaxation factor for radiation Blending Factors GGUVW Blending factor for velocities GGKE Blending factor for k and ε GGT Blending factor for temperature GGDEN Blending factor for density GGSCA Blending factor for scalars Residual Tolerances SORU Solver residual for U velocity SORV Solver residual for V velocity SORW Solver residual for W velocity SORP Solver residual for pressure SORK Solver residual for k SORE Solver residual for ε SORT Solver residual for temperature Number of Sweeps NSWPU Total number of solver sweeps for U in one run NSWPV Total number of solver sweeps for V in one run NSWPW Total number of solver sweeps for W in one run NSWPP Total number of solver sweeps for P in one run NSWPK Total number of solver sweeps for k in one run NSWPE Total number of solver sweeps for ε in one run NSWPT Total number of solver sweeps for T in one run Solution control can then be exercised as follows: 1. During execution, monitor the behaviour of normalised residual sums for each variable being solved for, by looking at the displayed values at the end of each Version 4.02

17-19


Chapter 17


iteration or time step. In addition, look at the flow variable values at the monitoring location, as specified in the “Monitoring and Reference Data” STAR GUIde panel. 2. While monitoring this display, you may decide to alter the course of the calculations by altering a model parameter, e.g. by (a) re-specifying an under-relaxation factor in order to speed up solution convergence (b) increasing the value of parameter SORMAX to stop the run at an earlier stage The values currently in use are shown on the dialog. If you want to change one or more of them, enter the new value in the appropriate box(es) and click Apply. This change is treated as pending. You can now either Cancel the change and then make others, or click Send to confirm it. 3. In the latter case, the parameter(s) will change inside STAR from the beginning of the next iteration (or time step) following the Send operation and a marker will be placed on the graph indicating the point at which something was changed. Note also the following points: • •

•

•

The colour of marker matches the colour of the tab in which the alteration was made and STAR itself will print a message indicating the change If you make multiple changes, you can highlight any one line and use the dialog’s Edit menu to copy/paste that line into other boxes and then edit any of the numbers. If you do not copy a line in, the code assumes that you are making changes to the last line. StarWatch also keeps a control history file called casename.ctrl.hist recording the changes made during a run. If you re-run a job without removing the control history file, STAR will make the same changes to the job that you made during the original run (so you can duplicate and repeat your changes to, say, under-relaxation factors). You do not have to have StarWatch running for the above changes to take place at various iterations. STAR will read the casename.ctrl file and make the changes to the run at the appropriate iteration. If you do not want the run changed the same way, delete casename.ctrl.hist before re-running a job.

Manipulating the StarWatch display The monitored variables chosen in the previous section are continuously displayed in the StarWatch panel as the calculation progresses, in two ways: •

•

As numerical values in the Iteration / Time Step Data section. The maximum and minimum values reached so far and the change since the previous iteration are also shown. As a graph of variable value versus iteration number/time step.

The detailed appearance of this graph may be adjusted as follows:

17-20

Version 4.02

Chapter 17

OTHER STAR-CD FEATURES AND CONTROLS Hard Copy Production

1. Horizontal scale Use the H: slider to achieve a reasonable scale, depending on the number of iterations 2. Vertical scale Use the V: slider to achieve a reasonable scale, depending on the variable being monitored. Note that this scale changes automatically as you switch from residual to field values. 3. Horizontal range Use the Iteration Number / Time Step slider to move the graph window to the required iteration range, after the job has finished executing. 4. Vertical range Use the vertical slider to move the graph window to the desired variable value range. Whether you need to do this or not depends on the vertical scale chosen. 5. Display size Select View > Partial View from the menu bar to reduce the extent of the StarWatch display, which now only shows the graph and associated legend. Selecting View > Full View restores the original display. Monitoring another job If you have several STAR jobs running simultaneously and you want to switch your monitoring to a different job, follow the procedure below: Step 1 Select Connect > Disconnect from the menu bar to terminate monitoring of the current job. Step 2 Select Connect > Host, enter the name of the machine running the job you wish to monitor in the Select Host dialog and click OK. Step 3 Choose the job’s PID from the list displayed in the Select STAR Job dialog and click OK. StarWatch should now start displaying the monitored variables for the new job. Alternatively, you may simply open another window and load another StarWatch panel, as described in “Running StarWatch”. Note that the number of panels that may be open simultaneously will depend on the setting specified in file /etc/services.

Hard Copy Production Neutral plot file production and use To obtain hard copy of a screen plot, switch the graphical output temporarily to the neutral plot file (see “Plot file (.plot)” on page 17-4). Once the required plot is on-screen, type TERMINAL,,FILE,RAST (switches to the neutral plot file in raster, i.e. colour-fill, mode) Version 4.02

17-21


Chapter 17

Hard Copy Production

or TERMINAL,,FILE,VECT (switches to the neutral plot file in vector, i.e. line-contour mode) followed by REPLOT (sends the picture to this file) TERMINAL,, (switch output back to the screen) The above process can be repeated as often as is necessary to write all required plot data to file case.plot. It is recommended that colour plots destined for a black-and-white printer should be converted to the grey-scale shading scheme (see “Colour settings” on page 4-10) before sending them to the neutral plot file. This can be done either by selecting the Post - Gray option in the Color Tool or by typing command CLRTABLE,GRAY To produce the hard copy, process the pictures stored in the neutral plot file outside the pro-STAR environment using one of the supplied programs in the PLOT suite. The latter are special graphics post-processors that either • •

generate files suitable for plotting on a given type of hard-copy device, or display the contents of the neutral plot file on your screen (see Appendix B in the Post-Processing User Guide for more details).

The PLOT programs available on your particular installation are normally accessed by opening a window and typing plot This produces a response of the form: Please enter the required plot driver: Available drivers are: ai fr gif hp ps pst su x xm [xm] where ai — Adobe Illustrator file output fr — Adobe Freehand file output gif — GIF file output hp — HP Graphics Language file output ps — PostScript file output pst — utility for adding an extra title to an existing PostScript file su — utility for reducing the size of an existing neutral plot file by removing hidden polygons x — X-windows terminal display xm — X Motif graphics display 17-22

Version 4.02

Chapter 17

OTHER STAR-CD FEATURES AND CONTROLS Hard Copy Production

Type the desired option and then follow the instructions on your screen, supplying additional information as required. Note that options such as xm are suitable for screen displays while options such as ps are for hard copy production. Note also that extended mode features such as translucency, layers, and smooth-shaded contour plots cannot be represented in the neutral plot format. To produce high-resolution hard copies in extended mode, use the high-resolution screen capture technique described in Chapter 2, “Screen capture”. Scene file production and use STAR-CD scene files provide a convenient way to store a fully post-processed model in a format that can be subsequently viewed with the lightweight and quick STAR-View viewer program. A STAR-CD scene file (extension .scn) stores the current state of the extended-mode graphics window, including the current plot and any labels, legends, and other screen information. However, unlike conventional hard copies produced using pro-STAR’s neutral plot facilities, STAR-CD scene files store a full 3-D representation of the current model so the view can be rotated, translated, and zoomed interactively in the STAR-View program. To produce such a file, first generate the desired plot in extended (OpenGL) mode (see Chapter 2, “Plotting Functions”). This can include any effects available in extended mode, including multiple layers, translucency, and smooth-shaded contours. Once the desired plot is achieved, select Utility > Write STAR-CD Scene File from the main pro-STAR menu. Select or type the desired scene file name into the File Selection dialog box which appears and press OK to write the file. Alternatively, pro-STAR command SCENE can be used to record the file. Once this file is written, simply run the STAR-View program by typing starview filename.scn in an X-window, where filename.scn is the file name containing the desired scene. Once the latter is loaded, the view in the model can be manipulated via the mouse in exactly the same way as in pro-STAR.

Version 4.02

17-23

APPENDICES CCM USER GUIDE

STAR-CD VERSION 4.02

CONFIDENTIAL — FOR AUTHORISED USERS ONLY

© 2006 CD-adapco

Appendix A

pro-STAR CONVENTIONS Command Input Conventions

Appendix A pro-STAR CONVENTIONS Command Input Conventions 1. A single command line may not be longer than 320 characters 2. Input is mostly case-insensitive; both capital and small letters are accepted (arguments such as file names, titles and screen labels are case-sensitive) 3. Command names may be abbreviated by the first four letters (with one exception: *ENDIF). Argument keywords may also be abbreviated by the first four letters (with one exception: parameter arguments for the MEMORY command) 4. Fields in a command string must be separated by a comma or by any number of spaces. 5. Multiple commands may be stacked on a single line, separated by a dollar sign ($). 6. Any command string with an exclamation mark (!) in column 1 is interpreted as a comment (and therefore not executed). 7. Double plus signs (++) at the end of a line indicates that the next line is a continuation of the current line. Individual arguments are not continued on a new line; the new line will begin a new argument. Any number of lines may be continued in this manner to form a single command line; however, the total number of characters in a command line formed in this manner may still not exceed 320 characters. 8. Any command may be entered from any module 9. In NOVICE mode (see command EXPERT), the program will prompt for arguments needed to execute the command. Command ABORT may be used at this prompt to abort the current command without performing any action. 10. Basic arithmetic is allowed on all command lines. Each operator must be separated by blanks or a comma from the numbers or parameters on either side. For example, the following command VLIST 10 * 10, A + 7 1000 / B is interpreted as VLIST 100 to (A+7) by (1000/B), where A and B are numeric parameters defined by the *ASK, *SET or *GET commands. All terms are evaluated strictly from left to right. 11. The keyword ‘ALL’ may be used in lieu of any vertex, cell, boundary, etc. range to denote that all items are to be used for the range. (Examples: CLIST,ALL and CTMOD,ALL,,,FLUID) 12. The appropriate item set keyword may be used in lieu of most item ranges to denote that all items in the set are to be used for the range. (Examples: CPDEL,CPSET and VLIS,VSET,,,1) Version 4.02

A-1

pro-STAR CONVENTIONS

Appendix A

Command Input Conventions

Keyword

Item Set

VSET

Current vertex set

CSET

Current cell set

BSET

Current boundary set

SPLSET

Current spline set

BLKSET

Current block set

CPSET

Current couple set

DSET

Current droplet set

13. The following keywords may be used in lieu of many item ranges to display the crosshair cursor in the plot window so the user may select a set to be used as the range. (Example: CLIST,CCRS). Keyword

Select

VCRS

Vertex set

CCRS

Cell set

BCRS

Boundary set

SCRS

Spline set

BLKCRS

Block set

DCRS

Droplet set

14. The following keywords may be used in lieu of entity numbers. (Example: V,MXV,1.0,2.0,3.0) Keyword

Interpreted As

MXV

Highest numbered vertex + 1

MXC

Highest numbered cell + 1

MXB

Highest numbered boundary + 1

MXS

Highest numbered spline + 1

MXK

Highest numbered block + 1

ICUR

Currently active coordinate system

15. Certain keywords (which may also be used in lieu of entity numbers) will cause pro-STAR to display the crosshair cursor in the plot window and expect A-2

Version 4.02

Appendix A

pro-STAR CONVENTIONS Help Text / Prompt Conventions

the user to select an item, as specified by the following description: (Example: STLIST,SXT) Keyword

Select

Interpreted As

BLKX

Block

Block number

BX

Boundary

Boundary number

BXP

Boundary

Boundary patch number

BXR

Boundary

Boundary region number

CX

Cell

Cell number

CXC

Cell

Cell colour index

CXG

Cell

Cell group number

CXM

Cell

Cell material number

CXP

Cell

Cell porous number

CXS

Cell

Cell spin index

CXT

Cell

Cell type number

DRX

Droplet

Droplet number

DRXT

Droplet

Droplet type number

SX

Spline

Spline number

SXC

Spline

Spline colour index

SXG

Spline

Spline group number

SXT

Spline

Spline type number

VX

Vertex

Vertex number

Help Text / Prompt Conventions 1. Words between slashes (e.g. /ANY/ALL/) represent legal alternatives for the field. 2. Numbers in parentheses represent defaults for the immediately preceding variable. 3. Variables beginning with ‘NV’ refer to vertices Variables beginning with ‘NC’ refer to cells Variables beginning with ‘NB’ refer to boundaries Variables beginning with ‘NSPL’ refer to splines Variables beginning with ‘NBLK’ refer to blocks Variables beginning with ‘NCP’ refer to couples Variables beginning with ‘NDR’ refer to droplets Version 4.02

A-3

pro-STAR CONVENTIONS

Appendix A

Control and Function Key Conventions

Control and Function Key Conventions 1. The following short-cuts using the Ctrl key are available: Control Key

Command

Ctrl-a

CSET,ALL

Ctrl-e

ZOOM,OFF $REPLOT

Ctrl-h

Query for help

Ctrl-o

ZOOM,OFF $REPLOT

Ctrl-q

QUIT

Ctrl-r

REPLOT

Ctrl-s

SAVE,,

Ctrl-w

Zoom out (by a factor of 2)

Ctrl-z

Zoom in (by a factor of 2)

2. Function key short-cuts can be defined or changed using the Function Keys option in the Utility menu. The default function key short-cuts are: Function Key

Default Command

F5

Repeat last command

F6

REPLOT

F7

CPLOT

F8

ZOOM,OFF $REPLOT

File Name Conventions The default name for any file read or written by the program is casename.ext, where casename is defined by the user and ext is the file name extension. If you enclose the file name in quotes, the extension default will be overridden and the exact name within the quotes will be used.

A-4

Version 4.02

Appendix B

FILE TYPES AND THEIR USAGE

Appendix B FILE TYPES AND THEIR USAGE

File Extension .ani .anim .bak .bnd .btr .ccd .ccm .cel .cel .cgns .cgrd .chm .cpfz .cpl .ctrl .dat .div .domain .drp .drpc .ecd .ecd2 .echo .elem .erd .erd2 .evn .evnc .fac Version 4.02

Usage Default input/output for recording animation commands Default save file for animation options Backup (i.e. previous version) of the current pro-STAR model file (binary) Default input/output for boundary definitions STAr file used for storing beam tracking data STAR file used for storing coal combustion data STAR-CD data repository (restart) file (binary/direct access) Default input/output for cell definitions Default output for surface cell definitions Default input/output for CGNS data files Default input file containing grid change commands Default output file for chemical scheme definitions (coded) Default temporary storage of ‘frozen’ vertex data used with the SAVE and MAP options of command CPFREEZE Default input/output for coupled cell definitions Editable file for interactive solution control Tecplot™ post data output file Post data file created when the solution diverges ICEM CFD™ post data output file Default output for droplet definitions (written with command PROBLEMWRITE) Default input/output for droplet data (coded) File for storing engineering data for cell monitoring File for storing dispersed-phase engineering data for cell monitoring Echo of all input typed by the user Default input/output for ANSYS™ element definitions File for storing engineering data for boundary region monitoring File for storing dispersed-phase engineering data for boundary region monitoring Default transient event save file (binary/direct access) Default input/output for ASCII event data files File containing cell face definitions B-1


File Extension .fvbnd .g3d .gen .grf .grf .ics .info .inp .inp .lfb .loop .mdl .mdl .msh .nas .neu .node .out .pat .pdft .pgr .plot .prob .proc .pstt .refi .reu .rsi .rsic .run .scl B-2

Appendix B

Usage Default input for GRID3D boundary data files Default input for GRID3D cell and vertex data files Default output for GENERIC data Default graph register data save file Default graph register ‘GET’ file Default combustion data file (binary), used in advanced IC engine models Run-time optional output file Any file containing pro-STAR commands Default input/output for miscellaneous problem data definitions File containing group and colour information for particles Default save file of current loop information Default pro-STAR model file (binary) Default input for SMAP-type data TGRID™ data output file Default input/output for NASTRAN™ data files Gambit™ data output file Default input/output for ANSYS™ node definitions Default output file Default input/output for PATRAN™ data files Look-up table file created when using PPDF chemical reaction models File containing participating media radiation data (binary) Neutral plot file Default output for STAR-CD problem data file (coded) File containing cell-to-processor mapping information used in STAR-HPC runs Default transient solution file (binary/direct access) Refinement data file used by the adaptive refinement commands (CMREFINE / CMUNREFINE) Residual history file for phase no. 2 (used in Eulerian two-phase problems) Default residual history file (binary/direct access) Default input/output of residual histories for BINARY-CODED-BINARY file conversions Standard run-time output file Default output for scalar variable definitions (coded) Version 4.02

Appendix B


File Extension .set .spd .spl .srf .stl .tabl .tbl .trk .trkc .trnc .trns .unv .uns .usr .vfs .vrml .vrt .vrt

Usage Default output for set definitions (written with the SETWRITE command) File for storing engine data (coded) Default input/output for spline definitions Default output for plotting-surface database (used to skip surface creation step in future plots) Default input for STL data files Default input file for droplet spray tables Default file for storing general table data Default input/output for particle/droplet tracks Default input/output of particle/droplet track data for BINARY-CODED-BINARY file conversions Default input for transient load data (coded) Default transient history save file (binary/direct access Default input/output for IDEAS™ (SDRC) universal file Fieldview™ data output file Default input/output for ASCII post data STAR file used for storing view factors Virtual reality data output file Default input/output for vertex definitions Default output for surface vertex definitions

The format for vertex definitions is: (file case.vrt) Vertex number, X, Y, Z (global coordinates) (I9, 3(1X,G21.14)) The format for boundary definitions is: (file case.bnd) Boundary number, cell number, face number, region number, patch number, region type (characters) (5(I9,1X),A) The format for spline definitions is: (file case.spl) Spline number, number of vertices, spline type (3I9) Up to 100 vertex numbers defining the spline (8I9) The format for couple definitions is: (file case.cpl) Couple number, number of cells (I8, 1X, I5) Up to MAXNCP cell number/face number combinations 7(I9,I2) The format for ASCII input to be used as post-processing data is: (file case.usr) Vertex and/or cell number (as appropriate), scalar value (I9, 6X, 6G16.9).

Version 4.02

B-3

Appendix C

PROGRAM UNITS

Appendix C PROGRAM UNITS

Version 4.02

Property

Units (SI)

Units (English)

AREA CONDUCTIVITY DENSITY DIFFUSIVITY DYNAMIC VISCOSITY FORCE HEAT FLUX HEAT OF FORMATION HEAT OF VAPOURIZATION LENGTH MASS MASS FLOW RATE MOLECULAR WEIGHT PRESSURE SPECIFIC HEAT SPEED OF SOUND SURFACE TENSION COEFFICIENT TEMPERATURE TIME TURBULENCE KINETIC ENERGY k TURBULENCE DISSIPATION RATE ε VELOCITY VOLUME VOLUMETRIC EXPANSION COEFF.

m2 W/mK kg/m3 m2/s Pa × s N W/m2 J/kg J/kg m kg kg/s kg/kmol Pa (N/m2) J/(kg × K) m/s N/m K (° Kelvin) s m2/s2 m2/s3 m/s m3 1/K

ft2 Btu/(hr × ft× F) lbm/ft3 ft2/s psi × s lb Btu/(hr × ft2) Btu/lbm Btu/lbm ft lbm lbm/hour lbm/kmol psi Btu/(lbm × F) ft/s lb/ft R (° Rankine) s ft2/s2 ft2/s3 ft/s ft3 1/R

C-1

Appendix D

pro-STAR X-RESOURCES

Appendix D pro-STAR X-RESOURCES The Motif version of pro-STAR utilises standard X resources for defining the layout and look of its windows. While default values for these resources are built into the program, you can override the defaults in two different ways: 1. The easiest method is to put resource definitions in your .Xdefaults file. This file is read by the Motif window manager when you log in or restart the window manager. Any changes made to this file do not take effect until either you log in again or you issue an xrdb command to re-read the X resource data base. Typically, you will issue the command as follows: xrdb -merge .Xdefaults

include the full path to the .Xdefaults file if you are not in your home directory

2. Any file can be used to set X resources. The only significance of the .Xdefaults file is that it is read automatically on start-up. You could, for example, create a file called PROSTAR.resources and put the resource definitions in that file. In this case, you would have to issue the command: xrdb -merge PROSTAR.resources before running pro-STAR in order to activate those definitions The following describes some useful resource definition commands:

Version 4.02

Prostar*background:

The default background colour for all pro-STAR applications

Prostar*foreground:

The default foreground colour for all pro-STAR applications

Prostar.geometry:

The size and position of the pro-STAR graphics window

Prostar.defaultFontList:

The font used for the pro-STAR graphics window menus

Prostar.OutputWindow.geometry:

The size and position of the pro-STAR output window

Prostar*cmdForm1Widget.height:

The height of the output history portion of the pro-STAR output window

Prostar*cmdForm2Widget.height:

The height of the input portion of the pro-STAR output text window

Prostar*Prostar_Output_Text.fontList:

The font used in the pro-STAR output window

D-1


Prostar*Prostar_Output_Text.foreground:

Appendix D

The foreground colour used in the pro-STAR output window

Prostar*Prostar_Output_Text.background: The background colour used in the pro-STAR output window Prostar*panel_name_B1.background:

The background colour of button 1 in the user panel named panel_name. Buttons in panels are numbered starting from zero and are incremented by 1 from top to bottom and from left to right. Any panel button can be defined using the proper panel name and button number.

Prostar*panel_name_B1.foreground:

The foreground colour of button 1 in the user panel named panel_name.

Prostar*panel_name_B1.fontList:

The font used for button 1 in the user panel named panel_name.

Prostar*macro_editor_text.fontList:

The font used for the text section of the macro edit dialog

Prostar*macro_editor_text.foreground:

The text foreground colour used in the macro edit dialog

Prostar*macro_editor_text.background:

The text background colour used in the macro edit dialog

Prostar*GUIde_INDEXCARD.background: The default background colour for all index cards (tabs) inside a STAR GUIde panel Prostar*GUIde_TABS.background:

The default background colour for all sub-index cards (sub-tabs) inside a STAR GUIde panel

X colour names are usually (but not always) defined in the file: /usr/lib/X11/rgb.txt Geometry definitions are in the form of W × H + X + Y where W is the width (in pixels), H the height, X the distance (in pixels) from the left of the screen to the left side of the window, and Y the distance from the top of the screen to the top of the window. Heights are also defined in pixels. Available font list names can usually be found by issuing the command: xlsfonts D-2

Version 4.02

Appendix D


The following shows a sample of resource definitions that could be used with pro-STAR: Prostar*background: Prostar*foreground:

paleturquoise3 black

Prostar.geometry: 800x800+480+0 Prostar.defaultFontList: -adobe-helvetica-bold-r-normal--14-140-75-75-p-82iso8859-1 Prostar.OutputWindow.geometry: 1000x870+0+0 Prostar*cmdForm1Widget.height: 700 Prostar*cmdForm2Widget.height: 70 Prostar*Prostar_Output_Text.fontList: -adobe-courier-bold-r-normal--18-180-75-75-m-110iso8859-1 Prostar*Prostar_Output_Text.foreground: blue Prostar*Prostar_Output_Text.background: gray85 Prostar*new_panel_B1.background: Red Prostar*new_panel_B1.fontList: -adobe-courier-medium-r-normal--12-120-75-75-m-70iso8859-1 Prostar*new_panel_B2.background: Green Prostar*new_panel_B2.fontList: -b&h-lucida-medium-r-normal-sans-24-*-*-*-*-*iso8859-1 Prostar*macro_editor_text.fontList: -adobe-courier-bold-r-normal--18-180-75-75-m-110iso8859-1 Prostar*macro_editor_text.foreground: blue Prostar*macro_editor_text.background: skyblue To customise the opening locations of Tools, Lists, etc. in pro-STAR If you run the XMotif version of pro-STAR, it is possible to arrange for tools to open in repeatable locations. This is especially useful if you have a number of favourite tools that you open each time and can make pro-STAR open them every time via the PROINIT file. There are two steps in doing this. The first is finding out where you want the tool to be. To this end, run pro-STAR and then place (and optionally resize) the tool to get the desired effect. Follow this by issuing the xwininfo command from an X-window to get the necessary numbers. For example: ibm3<68>xwininfo

Version 4.02

D-3


Appendix D

xwininfo: Please select the window about which you would like information by clicking the mouse in that window. xwininfo: Window id: 0x54007c2 "Check Tool" Absolute upper-left X: 587 Absolute upper-left Y: 374 Relative upper-left X: 0 Relative upper-left Y: 0 Width: 630 Height: 590 Depth: 8 Visual Class: PseudoColor Border width: 0 Class: InputOutput Colormap: 0x3d (installed) Bit Gravity State: ForgetGravity Window Gravity State: NorthWestGravity Backing Store State: NotUseful Save Under State: no Map State: IsViewable Override Redirect State: no Corners: +587+374 -63+374 -63-60 +587-60 -geometry 630x590-55-52 This gives us two pieces of information, the name and the location. The name is enclosed in quotes in the first line of output, for this case it is Check Tool. The location is given in the last line, -geometry 630x590-55-52. This gives the width and height as well as the location. The second step is to feed this information to pro-STAR via Xresources. The usual way is to edit file .Xdefaults in your home directory. In this case, add the following line: Prostar*CheckTool*Geometry: 630x590-55-52 This line is made up as follows: Prostar*NAME*Geometry: GEOMETRY where: NAME is the name of the window stripped of all spaces; capitalisation must be kept. GEOMETRY is the location of the window as found from the previous command. Once this line has been added to the file, pro-STAR should respond correctly. On some systems, restarting pro-STAR will suffice. Others may require you to log out and log in again or issue some variant of the xrdb command. The above has been tested and works so far on SGI and IBM machines. Other D-4

Version 4.02

Appendix D


machines may work with minor variations. A suitable PROINIT file will be: opanel tool$check Make sure that the PROINIT file is in your current directory or that it is pointed to by the STARUSR environment variable.

Version 4.02

D-5

Appendix E

USER INTERFACE TO MESSAGE PASSING ROUTINES

Appendix E USER INTERFACE TO MESSAGE PASSING ROUTINES Some user coding might need access to message passing routines when used in a parallel run. This appendix lists the parallel message passing calls that may be used within the user coding. IGSUM — Global Integer Summation Synopsis INTEGER FUNCTION IGSUM (LOCSUM) Parameters INTEGER LOCSUM — local value Returns integer sum of LOCSUM GSUM — Global Floating Point Summation Synopsis REAL1 FUNCTION GSUM (LOCSUM) Parameters REAL1 LOCSUM — local value Returns floating point sum of LOCSUM DGSUM — Global Double Precision Summation Synopsis DOUBLE PRECISION FUNCTION DGSUM (LOCSUM) Parameters DOUBLE PRECISION LOCSUM — local value Returns double precision sum of LOCSUM LGLOR — Global OR operation Synopsis SUBROUTINE LGLOR (LOC,GLO) Parameters LOGICAL LOC — local value (input parameter) LOGICAL GLO — global value (output parameter) LGLAND — Global AND operation Synopsis SUBROUTINE LGLAND (LOC,GLO) Parameters LOGICAL LOC — local value (input parameter) LOGICAL GLO — global value (output parameter)

1. Type REAL becomes DOUBLE PRECISION in double precision runs. Version 4.02

E-1

USER INTERFACE TO MESSAGE PASSING ROUTINES

Appendix E

GMAX — Global MAX operation Synopsis REAL1 FUNCTION GMAX (LMAX) Parameters REAL1 LMAX — local value Returns global MAX of LMAX GMIN — Global MIN operation Synopsis REAL1 FUNCTION GMIN (LMIN) Parameters REAL1 LMIN — local value Returns global MIN of LMIN IGMAX — Global MAX operation Synopsis INTEGER IGMAX (ILMAX) Parameters INTEGER FUNCTION IGMAX — local value Returns global MAX of ILMAX IGMIN — Global MIN operation Synopsis INTEGER FUNCTION IGMIN (ILMIN) Parameters INTEGER IGMIN — local value Returns global MIN of ILMIN

1. Type REAL becomes DOUBLE PRECISION in double precision runs. E-2

Version 4.02

Appendix F

STAR RUN OPTIONS Usage

Appendix F STAR RUN OPTIONS Usage star [options] [node1 [node2 [node3 [...]]]]

Options -version -abort -batch

-case=casename -chktime=minutes -chkdir=directory -chkpnt

-collect -dp

-devtool="program"

-fork -g

-kill

Version 4.02

Show STAR version information, which includes patch number. Send SIGABRT to stop STAR after the current iteration or time step. Generate script for running batch job. Useful if run is to be submitted via a batch-queuing system like IBM Loadleveler, LSF, OpenPBS, PBSPro, Sun Grid Engine or Torque. This requires STAR-NET 3.0.3 or later to be installed. Select the case name manually. This option is not needed in general. Enable STAR controlled check-pointing at a regular interval in minutes for fault tolerance. The default is off. Select directory for storing the check-pointed data. The default is to use a ‘CHECK’ sub-directory. Perform manual check-pointing of STAR results now. This option may be useful for visualising fields while STAR is still executing in parallel, since it will merge the case’s results. Collect and save data from previous crashed run only. Make STAR-CD run in double precision arithmetic. Current default is single precision, with the exception of combustion problems which use either STAR/KINetics or the Complex Chemistry model, in which case STAR-CD will execute in double precision. Attach a development tool like a debugger to a STAR-CD run. The use of this option is advised on sequential runs only. For parallel runs only LAM MPI and MPICH are fully supported with Totalview. Enable the use of fork() for starting local external moving mesh codes and NFS-based communications. Compile ufile source code, so that the user may employ a debugger to perform a step-by-step analysis of the coding in the user subroutines. See also option "-devtool". Send SIGKILL to terminate STAR immediately.

F-1

STAR RUN OPTIONS

Appendix F

Options -nolookahead -noramfiles -norecalc -norestart -nosave -noskip

-noufile

-restart -save="filelist"

Disable look ahead for socket-based external moving mesh code communications. Disable memory based scratch files. Disable the recalculation of radiation view factors. Disable restart (if selected in the problem file) by resetting the restart flag. Disable saving of results by using an empty save list. The "-save=" option can be used to make a new save list. Forces geometry decomposition (if applicable), events preparation (if applicable), user coding compilation and copying of input files (if applicable) before STAR-CD starts to execute. Ignore user coding in the "ufile" directory, i.e. the run’s results will not be influenced by the actual user coding. Continue the run from an existing restart file by resetting the restart flag in the problem file. Specify additional output files for treatment as results. On a parallel run, these files will be merged into a single file. Ideally, these files should be formatted into two columns: the first column containing an index numeral that can be ordered (i.e., pro-STAR cell number), and the second column containing the physical quantity of interest. Files that should not be merged should be left out from this option. Wildcards “*” and “?” are accepted.

Example: -save="file1.dat file2.dat" or -save="file1.dat" -save="file2.dat" -set variable="value" Set environmental variable to a value, especially on a parallel run, where the variable will be set on all processes.

-timer

-toolchest -ufile

F-2

Example: star -set MYVAR="on" Enable printing of detailed timing data. Use this option to extract execution time information from the run. Please note that the use of this option entails a performance penalty. Build new STAR toolchest from plug-in tools. Compile user coding and build new plugable object only. Useful to verify if user coding compiles, i.e., if it contains any syntax mistakes.

Version 4.02

Appendix F

STAR RUN OPTIONS Parallel Options -uflags="flags"

-ulib="librarylist"

-watch

Select additional flags for compiling user coding. This option gives the user added flexibility in using other compiler options that may not be listed in installation scripts. Specify precompiled user coding libraries and/or some additional dynamic shared objects required by user coding. Enable connection to the StarWatch daemon. The daemon itself and the StarWatch GUI still need to be run separately

Parallel Options -copy="filelist"

Specify additional input files for copying to domains on a parallel run.

Example: -copy="file1.dat file2.dat" or -copy="file1.dat" -copy="file2.dat" -decomp Run geometry decomposer only. Useful to check the outcome of the decomposition if it has to satisfy certain criteria. -decomphost=hostlist Selects host(s) for running the decomposer (i.e. host1:host2:…). In particular, for the Parmetis decomposition option more than two cpu’s (whether or not on the same machine) should be used. The number of cpu’s to decompose the mesh can be smaller than the number of requested partitions. Example: -decomphost="host1,2 host2" 5 In the above, STAR will decompose the mesh in 5 parts using 2 cpu’s on machine "host1" and 1 cpu on machine "host2".

Version 4.02

F-3

STAR RUN OPTIONS

Appendix F

Parallel Options -decompmeth=method

Select one of the decomposition methods listed below. The abbreviations shown in parentheses can be used instead. Their individual meanings are: optimised (o): The decomposition will be read from file .proc, composed of two columns: first column contains cell numbers, second column contains the process number to which the cells are going to be assigned. automatic (a): The decomposition will uniformly divide the number of cells between the intended number of processes, based purely on pro-STAR cell numbering. manual (m): The decomposition is done according to cell types, as they have been defined in pro-STAR sets (s): The decomposition is read from a .set file, as it has been defined in pro-STAR. metis (x): The mesh will be partitioned with Metis, a built-in graph-handling library. By default, each material domain will be decomposed in turn. ometis (y): Same as above, but with a lower memory footprint and higher execution time. geometric (g): The entire mesh (i.e., the mesh is treated as if it was just one single material domain) is decomposed in a single (Cartesian) direction in which the model is largest. parmetis (p): Parallel version of the Metis family of algorithms. Parmetis executes the domain decomposition step in parallel and requires less memory than the Metis algorithm. Parmetis calculates decompositions of similar quality to sequential Metis. By default, each material domain will be decomposed in turn. This option is to be used in conjunction with option ‘-decomphost’, above. The default is ‘metis’ decomposition, except when the model contains events, in which case the default becomes ‘sets’. Example: -decompmeth=g

F-4

Version 4.02

Appendix F

STAR RUN OPTIONS Parallel Options -decompflags="flags"

Special options for the domain decomposition step: vcom: Compress vertex indices on each geometry file; if compressed, the vertices on each geometry file will be numbered from 1 to the local maximum number; if not compressed, the vertices will retain their original numbering from the un-decomposed mesh. The vertex numbering may be important for the mesh motion operation (e.g. the vertex movement may be specified relative to a fixed vertex). The default action is to compress vertices, except for moving mesh and liquid film cases where the default is not to compress. novc: Disable vertex compression. outproc: If chosen, this option will trigger the creation of a cell assignment file in the case’s directory; this file (.proc), can be loaded into pro-STAR for the user to visualise the decomposition (with command RDPROC) or it can be used to repeat the same decomposition with, for example, a different version of STAR in conjunction with the ‘-decompmeth=o’ decomposition option. outsets: this option will trigger the creation of a sets file in the case’s directory. In this file (.sets), each set will contain the cells that belong to a certain subdomain; this file can be manipulated from within pro-STAR in the usual manner or it can be used to repeat the same decomposition with, for example, a different version of STAR in conjunction with the ‘-decompmeth=s’ decomposition option.

-distribute

-loadbalance

-mergehost=hostlist -mpi=auto

Version 4.02

Example: -decompflags=”outproc” Select distributed data parallel runs using local scratch disks, as set up at the time when STAR-CD was initialised. Please see your Systems Administrator for details. Select load balancing taking into account the relative speeds of the hosts, as set up at the time when STAR-CD was initialised. Please see your Systems Administrator for details. Selects host for merging results (i.e. host1:host2:…). Automatic selection of the MPI implementation using the vendor order shown below. This is the default behaviour which can be changed by supplying one of the flags below: F-5

STAR RUN OPTIONS

Appendix F

Resource Allocation -mpi=os -mpi=gm -mpi=hp -mpi=intel -mpi=ra -mpi=scampi -mpi=score -mpi=sgi -mpi=lam -mpi=mpich -mppflags="flags"

-mpphosts

-nocollect

-nocopy

-nodecomp

-noshmem -scratch=directory

Select Operating System Vendor’s MPI Select MPICH-GM (Myricom GM MPI) Select HP MPI Select Intel MPI Select RA-MPICH (Rapid Array MPI) Select ScaMPI (Scali MPI) Select SCore MPI Select SGI Itanium MPI Select LAM MPI Select MPICH (ANL/MSU MPI) Select additional flags for message passing protocol. Use this option to supply additional flags as expected by the MPI implementation. In general, the user should not need to use it. Select non-default network for message passing protocol using alternative host names, as set up at the time when STAR-CD was initialised. Please see your Systems Administrator for details. Disable data collection at the end of a distributed data parallel run. This also disables saving of results. It is possible to restart using the data already distributed to the local scratch disks. Please note that any updates to these files must be performed manually and the data can be manually collected using the "-collect" option at the end of the runs. Disable copying of input files by using an empty copy list. The "-copy=" option can be used to make a new copy list. Do not decompose the computation mesh on a parallel run and use the last decomposition instead. The user should not need to use this option in general. Disable shared memory communications for parallel runs on a single node. Select the scratch directory path to use on all nodes for distributed data runs. This over rides HPC_SCRATCH settings and must be unique for each running case.

Resource Allocation The user does not select sequential or parallel STAR runs directly. Instead this is automatically determined from the resources assigned by the user or the resource manager. If STAR options are required they need to be specified before the nodes list.

F-6

Version 4.02

Appendix F

STAR RUN OPTIONS Default Options -mvmeshhost=host

-nodefile=file -nooverload

node1 node2 node3

Select additional resource for running external moving mesh code. The default is to overload the STAR master CPU with the external moving mesh code, when one is being used. Select nodes to use for running STAR in a file. This can be specified on a single line or multiple lines. Disable overloading of the STAR master processor with the external moving mesh code. The number of STAR domains plus one extra process is needed in the resource line. The nodes to use for running STAR. The node is specified in the format “hostname,np”, where “np” is the number of processes to use. The local host will be assumed if the “hostname” is not specified and a single process will be used if the “,np” parameter is not supplied.

Default Options The environment variable STARFLAGS can be set to include some default STAR options that will be processed before any command line options. Its value is normally set in the software initialization file (software.ini) to cater for site-specific STAR solver options that are always used. Examples are: STARFLAGS=-dp STARFLAGS=-set VARIABLE="Some Value" STARFLAGS=-mpi=mpich -noshmem -distribute -timer The user can reset STARFLAGS manually or use a different .ini file to change its value. The options defined in STARFLAGS are always processed first and can be over-written by additional command-line options, but only if an alternative option exists. Thus, if STARFLAGS=-mpi=mpich the user can still use LAM MPI as follows: star -mpi=lam However, if STARFLAGS=-dp this setting cannot be modified because a single-precision option is not available at the command line. Another example is: STARFLAGS=-set GTIHOME=/users/netapps/gt GTISOFT_LICENSE_FILE=27005@heraclitus

Using STARFLAGS, the software administrator can set things up so that ordinary Version 4.02

F-7

STAR RUN OPTIONS

Appendix F

Cluster Computing

users need to do less work. Other examples are to make everybody run in double precision, to always use the -distribute option, etc.

Cluster Computing Cluster computing is widely adopted by STAR-CD users and typically consists of computing nodes connected by network interconnect devices such as Gigabit Ethernet, Myrinet and InfiniBand. CD-adapco have been actively working with computer hardware and software vendors to ensure that STAR-CD takes full advantage of progress in cluster technology. An important aspect of this work is STAR-CD’s integration with MPIs that support various network interconnect devices. Users can type star -h to check the available MPI options for the port they are using prior to issuing one of the "star -mpi=" commands in their session window. By checking which network interconnect devices are supported by each MPI, users can determine whether STAR-CD works with a particular MPI/interconnect combination. STAR-CD’s performance on a cluster is influenced by numerous factors, such as MPI performance, interconnect latency, interconnect bandwidth and file system performance. We have been working with our hardware and software partners to provide benchmark data on various clusters and such data are available on CD-adapco’s web site. Due to the extensive range of cluster configurations and the rapid developments in cluster technology, it is not possible to test all MPI/interconnect combinations and to measure their performance. Users are advised to contact the relevant system vendors to check whether a particular combination of MPI implementation and network interconnects works with STAR-CD.

Batch Runs Using STAR-NET STAR-NET 3.x is a new, lightweight tool for running applications in sequential and parallel modes under a batch environment using a resource manager. It is a completely new design, not compatible with the previous STAR-NET 2.0.x versions (which only work with STAR-CD in parallel mode). Currently, the IBM Loadleveler, LSF, OpenPBS, PBSPro, Sun Grid Engine and Torque resource managers are supported through STAR-NET 3.x compliant plug-ins. Therefore, you must install STAR-NET 3.x in order to run in batch mode or to use any of the above resource mangers. Note also that the PBSPro and Torque are only supported in OpenPBS compatibility mode. Concise guidelines for running under each system are given below, assuming prior configuration as detailed in the Installation and Systems Guide. Running under IBM Loadleveler using STAR-NET To run STAR-PNP under Loadleveler: 1. Create a batch.sh script by specifying the -batch option: star -batch

F-8

Version 4.02

Appendix F

STAR RUN OPTIONS Batch Runs Using STAR-NET

where represents all the normal STAR-PNP flags for your job, as described in the sections above. Note that you cannot assign a node list for resource allocation in batch mode as this will be performed automatically by Loadleveler. 2. Submit your job using the llsubmit command. For example: star -batch -chktime=60 llsubmit batch.sh The llsubmit command does not allow any resource selection and so this must be specified correctly in the batch.sh script. The following shows the most useful settings: # # # #

@ @ @ @

node_usage class node total_tasks

= shared = = 3 = 8

The above requests 3 nodes and a total of 8 CPUs for running the batch job. 3. The llsubmit command does not support automatic restarts and checkpointing, so you will need to enable application-level check-pointing by STAR-PNP as follows: star -batch -chktime=60 llsubmit batch.sh Other useful Loadleveler commands: •

Show all my Loadleveler jobs llq -u username

•

Continuously monitor the output of job number 123 tail -f batch.o123

•

Terminate job number 123 under Loadleveler llcancel 123

•

Use the built-in GUI interface for submitting and monitoring jobs xloadl

Running under LSF using STAR-NET To run STAR-PNP under LSF: 1. Create a batch.sh script by specifying the -batch option: star -batch where represents all the normal STAR-PNP flags for your job, as described in the sections above. Note that you cannot assign a node list for Version 4.02

F-9

STAR RUN OPTIONS

Appendix F

Batch Runs Using STAR-NET

resource allocation in batch mode as this will be performed automatically by LSF. 2. Submit your job to the queue using the bsub command. For example: (a) To submit to queue starnet requesting 2 to 4 processors: bsub -q starnet -n 2,4 batch.sh (b) To submit to queue starnet requesting 2 to 4 processors with LSF-controlled automatic restarts and enabling check-pointing every 60 minutes: bsub -q starnet -n 2,4 -r -k "CHECK 60" batch.sh It is recommended that you always enable check-pointing and automatic restarts to allow time-windowing/high-load-enforced job migration to work. (c) To submit to a subset of hosts: bsub -q starnet -m "host1 host2 host3" -n 2,4 -r -k "CHECK 60" batch.sh

Other useful LSF commands: •

Show all my LSF jobs bjobs

•

Continuously monitor the output of job number 123 peek -f 123

•

Terminate job number 123 under LSF bkill 123

•

Use the built-in GUI interface for submitting and monitoring jobs xlsbatch Alternatively, command starnet can be used to display a brief summary of the current LSF status.

Running under OpenPBS using STAR-NET To run STAR-PNP under OpenPBS: 1. Create a batch.sh script by specifying the -batch option: star -batch where represents all the normal STAR-PNP flags for your job, as described in the sections above. Note that you cannot assign a node list for resource allocation in batch mode as this will be performed automatically by OpenPBS. 2. Submit your job to the queue using the qsub command. For example, to F-10

Version 4.02

Appendix F

STAR RUN OPTIONS Batch Runs Using STAR-NET

submit to queue starnet requesting 3 nodes with 2 processors each: qsub -q starnet -l nodes=3:ppn=2 batch.sh 3. OpenPBS does not support automatic restarts and check-pointing, so you will need to enable application-level check-pointing by STAR-PNP as follows: star -batch -chktime=60 qsub -q starnet -l nodes=3:ppn=2 batch.sh Other useful OpenPBS commands: •

Show all my OpenPBS jobs qstat -u username

•

Continuously monitor the output of job number 123 tail -f batch.sh.o123

•

Terminate job number 123 under OpenPBS qdel 123

•

Use the built-in GUI interface for submitting and monitoring jobs xpbs

Please note that only the OpenPBS features of PBSPro and Torque are supported. Running under PBSPro using STAR-NET PBSPro is supported in OpenPBS compatibility mode. This means that only OpenPBS features are supported (see the description above). Running under SGE using STAR-NET To run STAR-PNP under Sun Grid Engine: 1. Create a batch.sh script by specifying the -batch option: star -batch where represents all the normal STAR-PNP flags for your job, as described in the sections above. Note that you cannot assign a node list for resource allocation in batch mode as this will be performed automatically by Sun Grid Engine. 2. Submit your job to a queue using the qsub command. For example: (a) To submit to parallel environment starnet requesting 2 to 4 processors: qsub -pe starnet 2-4 batch.sh (b) To submit to a subset of queues: Version 4.02

F-11

STAR RUN OPTIONS

Appendix F

Batch Runs Using STAR-NET qsub -pe starnet 2,4 -q queue1,queue2,queue3 -ckpt starnet batch.sh

3. Sun Grid Engine supports automatic restarts but not check-pointing, so you will need to enable application-level check-pointing by STAR-PNP as follows: star -batch -chktime=60 qsub -pe starnet 2-4 -ckpt starnet batch.sh Please note that Sun Grid Engine versions earlier than 5.3 do not support automatic restarts when the master host fails. Other useful SGE commands: •

Show all my Sun Grid Engine jobs qstat -u username

•

Continuously monitor the output of job number 123 tail -f batch.sh.o123

•

Terminate job number 123 under Sun Grid Engine qdel 123

•

Use the built-in GUI interface for submitting and monitoring jobs qmon Alternatively, command starnet can be used to display a brief summary of the current SGE status.

Running under Torque using STAR-NET Torque is supported in OpenPBS compatibility mode. This means that only OpenPBS features are supported (see the description above).

F-12

Version 4.02

Appendix G

BIBLIOGRAPHY

Appendix G BIBLIOGRAPHY [1] [2] [3] [4] [5]

[6]

[7]

Version 4.02

Kee R.J., Rupley F.M. and Miller J.A. 1990. ‘The Chemkin Thermodynamic Data Base’, Sandia Report No. SAND87-8215B. “CET89 — Chemical Equilibrium with Transport Properties”. 1989. NASA Lewis Research Center. Liepman H.W. and Roshko A. 1957. “Elements of Gas Dynamics”. John Wiley & Sons, New York. Shapiro A.H. 1953. “The Dynamics and Thermodynamics of Compressible Fluid Flow — Vol. 1 and Vol. 2”. Ronald, New York. Gordon S. and McBride B. J. 1994. “Computer Program for Calculation of Complex Chemical Equilibrium Compositions and Applications, Part I. Analysis”, NASA Ref. Publ. 1311, NASA Lewis Research Center. McBride B. J. and Gordon S. 1996. “Computer Program for Calculation of Complex Chemical Equilibrium Compositions and Applications, Part II. Users Manual and Program Description”, NASA Ref. Publ. 1311, NASA Lewis Research Center. Harten, A., Lax, P.D. and Van Leer, B. 1983. ‘On upstream differencing and Godunov-type schemes for hyperbolic conservation Laws’, SIAM Rev., 25, pp. 35-61.

G-1

INDEX

INDEX Commands are listed in a separate index immediately following this one

A abbreviations of commands 16-1 absorption coefficient 7-4 absorptivity of baffle 4-26 thermal 4-25 of wall 4-22 solar 4-21 thermal 4-20 accuracy numerical 1-5 temporal 1-14 view factor 7-5 adaptive mesh refinement. See mesh, adaptive refinement add set selection option 2-21, 4-3 aeroacoustics 13-3 aerodynamics problems, boundaries for 4-36 all set selection option 2-21, 4-3 angle internal 1-5 warp 1-5 angular velocity 12-1, 12-8 defined by user subroutines 14-16 animation 2-32 anode 8-33 append mode 2-13 aspect ratio cell 1-5 to 1-6 patch 7-6 atomisation models 9-1 attachment boundaries. See boundary conditions, attachment autoignition (double-delay model) 8-37 to 8-39 average nuclear radius 11-7 axis of rotation 12-4, 12-8 axisymmetric flow. See flow, axisymmetric

B background colour 2-32 material, in chemical reactions 8-17 background fluid 13-3 baffle 4-23 to 4-27 conducting 3-18 expanding 3-18 in restart after mesh changing 5-17 porous 4-24, 4-25, 6-5 radiation 4-25 side numbering 4-24 thickness 3-18 transparent 4-26 batch mode 2-34 beams 7-6 black and white printing 17-22 Version 4.02

black body 4-20 blending factor 14-19, 15-2, 17-19 block set 2-21, A-2 boiling. See cavitation boundary cells 1-4 defining 4-2 layer 1-6 in compressible flow 3-10 turbulent 1-7, 3-14 list 4-3 location 1-11, 4-1 to 4-5 maximum number of 2-18 modifying 4-2 monitoring behaviour 5-3 monitoring regions 4-40 notional 1-4 region 4-1 changing type 4-2, 5-4, 5-11 compress 4-7 default (region no. 0) 4-7 defining 4-5 set 2-21, A-2 selection 4-3 symmetry 1-3 types 4-5 visualisation 4-41 boundary conditions 1-11 attachment 4-38 to 4-39 displaying 12-30 baffle 4-23 to 4-27 for solar radiation 4-26 cyclic 1-3, 1-12, 4-27 to 4-32 anticyclic 4-29 partial 4-30 defined by GT-POWER 4-9 defined by tables 2-24, 4-7 baffles 4-27 free-stream 4-33 inlet 4-10 non-reflective 4-19 outlet 4-12 pressure 4-14 Riemann 4-37 stagnation 4-16 transient wave transmissive 4-35 wall 4-22 defined by user subroutines 4-7, 14-5 defined using load steps 5-6 to 5-14 degassing 4-40, 10-1 for cavitating flows 11-5 for compressible flow 3-9 for free surface flows 11-1 for liquid films 13-6 for rotating frames 12-6 for subsonic compressible flow 3-9 1

Index for supersonic flow 3-9 for transonic flow 3-10 free-stream 1-12, 4-32 to 4-34, 14-6 inlet 4-9 to 4-11, 14-5 non-reflective 1-12, 4-16 to 4-19, 14-6 outflow 1-11, 4-9, 4-10, 4-12 outlet 1-11, 4-11 to 4-12, 14-6 plots of 4-8 prescribed flow 1-11 pressure 1-12, 4-12 to 4-14, 14-6 for cavitating flows 11-9 for free surface flows 11-4 radial equilibrium 4-9 on pressure boundaries 4-12, 4-14 radiation 4-39 to 4-40, 7-8 Riemann 1-12, 4-36 to 4-38, 14-6 stagnation 1-12, 4-14 to 4-16, 14-6 for rotating reference frames 12-1, 12-4, 12-9 symmetry 1-12, 4-27 table input 2-24 to 2-31 time-varying 5-4 transient wave transmissive 1-12, 4-34 to 4-36, 14-6 wall 4-19 to 4-23, 14-6 for solar radiation 4-20 in turbulent flow 3-12 See also wall buoyancy 1-9 See also density buoyancy driven flow. See flow, buoyancy driven byte ordering 16-1

C calculations, checking 1-20 catalytic converters 8-1 cathode 8-33 cavitation 11-5 to 11-10 defined by user subroutines 14-11 in free surface flows 11-3, 11-6, 11-8 initialisation 11-8 solution algorithms 1-13, 11-9 steady-state flows 11-6 temperature calculation 11-8 vapour properties 11-7 CEA (Chemical Equilibrium with Applications) program 8-3 cell 1-2 attachment 12-18, 12-23 to 12-28 data 17-3 detachment 12-23 to 12-28 face boundary 4-1 matching 4-29 index 3-3 interface 12-18 layer addition 12-14 2

removal 5-17, 12-14 list 2-24, 3-4 maximum number of 2-18 monitoring behaviour 5-3 near-boundary 1-4 near-wall 3-12, 3-13, 3-14 number of 15-1 plot 2-23 properties 3-1 set 2-21, A-2 volume 3-22 shape 1-4 shape changing 5-15 size 1-6 table 3-1 to 3-3 compress 3-3 editor 3-1 to 3-3 radiation 7-8 number 3-2 porosity 6-1 tool 3-3 type 3-3 change fluid type 12-23 change grid (CG) operation 12-9 characteristic length. See length, characteristic velocity. See velocity, characteristic check tool 3-5 checking model and problem data 15-1 results 1-20 chemical reaction. See reaction CHEMKIN 8-11 clearing entire geometry 17-14 coal combustion 8-41 to 8-47 default models 8-42 inlet mass fractions 8-44 coal particle size distribution 8-45 colour background 2-32, D-1, D-2 foreground D-1, D-2 options 2-4 palette 3-2 table index 3-2 colour tool 17-22 combustion. See reaction and coal combustion command abbreviations 16-1 arithmetic in A-1 conventions A-1 help A-3 history 2-13, 2-19, 17-2 input 2-13 echo 2-19 output 2-13 number of lines 2-19 commands 2-36, 2-41 Version 4.02

Index compressibility 3-9 compressible flow Courant number 5-9 model setup 3-9 to 3-11 outlets 1-11 to 1-12 pseudo-transient approach 5-1 stagnation boundaries 4-14 transient, boundaries for 4-34 compression wave 4-32 condensation to liquid films 13-6 conduction thickness 3-19 conduction through baffles 4-25 conductivity defined by user subroutines 14-6 in chemical reaction problems 8-17 in multi-component mixing 13-3 connectivity 12-18 control keys A-4 controlling STAR with StarWatch 17-17 convergence 1-19 in steady-state calculations with SIMPLE 1-16 in transient calculations with SIMPLE 1-17 rate of 4-13 coordinate system in attachment boundaries 12-19, 12-23 in porous media 6-2 to 6-3 corrector step tolerance 1-14 couple set 2-21, A-2 couples across cyclic boundaries 4-29 Courant number 5-9, 14-18 for Lagrangian multi-phase flow 9-10 for LES turbulence models 3-15 for pseudo-transient calculations 5-2 crank angle 9-9 cursor select 2-19, A-2 customisation of pro-STAR 2-18, 16-1 cyclic boundary pair. See boundary conditions, cyclic set list 4-31

D deleting entire model 17-14 density at Riemann boundary 4-37 calculation 1-9, 3-9 defined by user subroutines 14-8 in aeroacoustic analysis 13-4 in buoyancy driven flow 3-20 in free surface flows 11-3 in PPDF scheme reactions 8-4, 8-19 reference 3-21 in buoyancy driven flow 4-14 under-relaxation 1-15, 17-19 dependent variable in tables 2-27 initialisation 1-10, 4-42 Version 4.02

monitoring 1-19 printout 15-3 differencing schemes 15-2 for free surface flows 11-2 for steady-state runs 5-2 for transient runs 5-11 for use with DES turbulence models 3-15 for use with LES turbulence models 3-15 diffusion reaction system 8-1 diffusivity characteristic 1-14 molecular, defined by user subroutines 14-8 porous 6-4 directory, working 2-3, 2-9 discrete fourier transform (DFT) algorithm 4-18 discrete transfer/ordinates radiation. See radiation discretisation error 1-6, 1-21 schemes 1-21 temporal for cavitating flows 11-9 for DES turbulence models 3-15 for free surface flows 11-4 for LES turbulence models 3-15 with the SIMPLE algorithm 1-18 time 5-11 volume 1-3 distance, normal dimensionless. See near-wall, dimensionless normal distance and y+ values distributed resistance 6-1, 15-2 user subroutines 14-8 divergence 2-6, 2-7 double precision mode 1-18 drag coefficient 14-19 force 14-14 droplet age 9-7 collision models 9-1 defined by user subroutines 14-12 diameter distribution function 9-1 defined by tables 2-26 gravitational effect 3-21 information 9-8 mass transfer 14-13 momentum transfer 14-13 number density 14-12 positions 9-7 reading data 9-5 set 2-21, A-2 set selection 9-6 track list 9-7 transfer to/from liquid films 13-5, 13-6, 13-7 user subroutines 9-1 to 9-3, 14-12 to 14-13 volume 9-10

3

Index

E emissivity at escape boundaries 4-39 at radiation boundaries 4-39 of baffle 4-26 solar 4-26 thermal 4-25, 7-1, 7-4 of wall 4-22 thermal 4-20, 7-1, 7-4 with FASTRAC 7-2 engine data 9-9 enthalpy defined by user subroutines 14-7, 14-10 in PPDF scheme reactions 8-19 stagnation 3-11 temperature dependence 14-7 environment variables 17-12 equations of state 1-9 error messages 2-19 sweep limits 1-14, 1-15 numerical discretisation 1-6, 1-21 recovery 2-20 round-off 1-18 splitting 1-13 escape surfaces 7-2, 7-4 Eulerian multi-phase flow boundaries 10-1 interphase 10-2 model setup 10-1 to 10-4 phase-escape boundary 4-40 response coefficient 14-14 user subroutines 14-14 evaporation from liquid films 13-6 event steps cell attachment 12-22 cell inclusion/exclusion 12-28 cell removal/addition 12-14 deleting 12-12 listing 12-12 modifying 12-12 moving mesh 12-9 moving pistons 12-13 reading 12-12 regular sliding 12-18 turning off 12-30 writing 12-12 exhaust gas recirculation 8-11, 8-17 exhaust valve 5-11 expansion wave 4-32 exposure of baffle 4-26 of wall 4-20, 4-22

F facets 1-4 4

FASTRAC 7-1 boundary conditions 7-2 escape boundaries 7-5 patches 7-2 symmetry and cyclic boundaries 7-5 user subroutines 7-6 view factors 7-3 with moving mesh 7-7 favourites menu 2-17 file menu 2-16 case name 17-1 edit file 17-10 model title 2-19 resume 17-10 resume from 17-2 resume model 17-2 save as coded 17-3, 17-10 save model 17-2, 17-10 system command 2-18 write geometry file 17-5, 17-10 write problem file 17-6, 17-10 files 2-36, 17-1 to 17-12, B-1 to B-3 boundary 17-3 cell 17-3 coded 17-3 command 17-10 data repository 17-5 droplet data 9-5 echo 17-2 editing 17-11 engines 9-9 event steps 12-12, 12-19 format of B-3 geometry 17-5 load step 5-9 macros 16-6 manipulating 17-9 mapping 5-15 model 2-4, 17-2 monitoring engineering data 5-3 output 2-7 panels 16-5 param.prp 17-13 plot 17-4 problem 2-6, 17-6 PRODEFS 16-1 PROINIT 16-1 reaction mechanism 8-11 relationship between 17-7 residual 17-6 restart 2-7 scalar properties 13-3 scene 17-23 set-up 16-1 solution 17-5 STAR-CD 3.2x equivalents 17-6 temporary 17-14 Version 4.02

Index transient 5-5, 5-13, 17-4, 17-7 compressing 5-14 vertex 17-3 view factors 7-3, 7-6 film stripping 13-5, 13-7 flame kernel 8-36 flamelet library 8-20 flow axisymmetric 1-12 buoyancy driven 1-13, 1-15, 3-20, 5-4 pressure boundaries 4-14 under-relaxation 3-21 unstable 3-21 cavitating. See cavitation chaotic 5-4 compressible. See compressible flow cyclic 5-4 free surface. See free surface flow impingement 3-15 inviscid 4-19 enthalpy 3-10 non-Newtonian 3-11 periodic 5-4 prescribed 1-11 split 4-11 steady 1-10, 1-15 to 1-17 analysis controls 5-1 to 5-4 output controls 5-2 solution controls 5-1 subsonic 3-9 supersonic 1-11, 3-9 mesh at inlet boundaries 4-10 transient 1-13 to 1-15 analysis controls 5-4 to 5-14 output controls 5-5, 5-12 solution controls 5-5, 5-9 transonic 1-11, 3-10 residuals 3-10 turbulent 3-12 unsteady 1-10 fluid background 13-3 injection 3-21 to 3-22 defined by user subroutines 14-10 mixture 13-1 non-Newtonian 3-11 properties 13-3, 15-2 stream 15-2 multiple 3-5 font size 2-33 fonts D-1, D-2 force, body 1-9 FORTRAN conventions 14-4 free surface flow 11-1 to 11-5 defined by user subroutines 14-11 density 11-3 differencing schemes 11-2 Version 4.02

initialisation 11-3 pseudo-transient approach 5-1 solution algorithms 1-13, 11-4 steady-state 11-1 surface tension 11-2 temperature calculation 11-3 free-stream boundary. See boundary conditions, freestream function keys 16-9 to 16-11, A-4

G gas ideal 3-9 law. See ideal gas law geometry, modifying 5-15 graph menu 2-17 graphics driver 2-3 group number 3-2 GT-POWER 4-9

H hard copy 17-21 heat conductivity defined by user subroutines 14-6 in chemical reaction problems 8-17 in multi-component mixing 13-3 transfer coefficient 3-17, 14-17 in baffles 3-18 in porous media 6-4 solid-fluid 3-16 to 3-20 solid-solid 3-20 help menu 2-17 on-line help 2-2 pro-STAR help 17-7

I I/O window 2-13 IC setup panels 8-23 to 8-37 ideal gas 3-9 law 3-20, 8-4 ignition 8-10, 8-15, 8-21, 14-15 advanced ICE models 8-24, 8-27, 8-29, 8-31 AKTIM 8-33 to 8-36 for simulations involving cell layer removal 12-18 imbalance 7-5 independent variables in tables 2-27 inflow at outlet boundaries 1-11 INFO button 2-22 initial conditions 1-10, 1-20, 17-6 defined by tables 2-24 for liquid films 13-6 for transient analyses 1-18 5

Index initialisation procedure in Lagrangian flow using user coding 14-13 in moving meshes 14-17 steady-state run 4-42 transient run 4-42 injection groups 9-2 injection. See fluid, injection inlet. See boundary conditions, inlet inner iterations. See iterations, inner input/output window 2-13 instability numerical 1-10, 1-14, 4-13 physical 1-9 interface sliding 12-19 solid-fluid 3-18 radiative 7-5 invert set selection option 2-21, 4-3 iterations inner 1-13 number of 15-1 outer 1-13, 1-17 iterative calculation 1-19

K Kirchoff’s law 4-20, 4-26

L Lagrangian multi-phase flow atomisation models 9-1 mesh 9-10 model setup 9-1 nozzle models 9-1 static displays steady-state 9-5 transient 9-8 trajectory displays 9-8 user subroutines 9-1 to 9-3, 9-10, 14-12 with liquid films 13-5 See also droplet length, characteristic 15-2 lift coefficient 14-19 lighting material 3-2 liquid films 13-5 to 13-7 boundary conditions 13-6 boundary regions 13-5 evaporation/condensation 13-6 film stripping 13-5, 13-7 gravitational effect 3-21 initial conditions 13-6 multi-component 13-6 results 13-6 velocity 13-7 with Largrangian multi-phase flow 13-5 lists 6

boundaries 4-3 cells 3-4 cyclic sets 4-31 droplet tracks 9-7 tracks 9-7 lists menu 2-16 load steps 5-6 to 5-14, 14-18 definition 5-8 identifying number 5-11 in multi-component mixing 13-3

M macros 16-6 to 16-9 creating 16-7 menus 16-8 modifying 16-7 mass conservation 8-3, 8-10 flow rate defined by tables 2-26 defined by user subroutines 14-10 flux 3-22 in excluded cells 12-28 transfer coefficient 14-17 droplet 14-13 in porous media 6-4 material number 3-2, 3-7 properties 3-1 maximum plot screen 2-32 memory requirements of pro-STAR 17-13 menus 2-16 to 2-17 mesh adaptive refinement 5-17 to 5-19 at non-reflective boundaries 4-18 block. See block set distortion 1-5 to 1-6 problems caused by 1-16 distribution, near walls 1-7 mean dimension of 1-14 moving 5-14, 12-9 to 12-13 defined by user subroutines 14-16 in porous media 6-5 mesh preview mode 12-13, 12-18 parameters 12-10 post-processing 12-29 pre-processing 12-28 restoration to original state 12-29 with radiation 7-3, 7-5 polyhedral at boundaries attachment 4-39 free-stream 4-33 pressure 4-14 Riemann 4-37 Version 4.02

Index stagnation 4-15 supersonic inlet 4-10 transient wave transmissive 4-35 walls 4-22 refinement 5-15 to 5-17 rotating. See rotating reference frames 12-1 sliding 12-18 to 12-22 defined by user subroutines 14-16 mesh preview mode 12-22 parallel processing 12-22 regular interface 12-18 to 12-22 without shearing 12-21 tetrahedral, at boundaries free-stream 4-33 pressure 4-14 Riemann 4-37 stagnation 4-15 supersonic inlet 4-10 transient wave transmissive 4-35 walls 4-22 visualisation colour setting 3-1 lighting effect 3-1 message passing routines E-1 mixing, multi-component 13-1 mixture fluid 13-1 fraction 8-2, 8-3, 8-10, 8-16 model checking 1-20 title 2-19 modelling strategy 1-1 modifying cell type 3-3 to 3-4 modules menu 2-16 transient 5-6, 17-4 monitoring engineering data 4-40, 5-3, 5-13 boundary behaviour 5-3 cell behaviour 5-3 field data 5-12 field variables 2-7, 5-13, 14-19, 15-2 to 15-3, 17-18 numerical solution 5-3, 17-8, 17-19 to 17-21 scalars 5-12 multi-component liquid films 13-6 mixing 13-1 setting up models 13-1 multi-phase flow. See Lagrangian multi-phase flow and Eulerian multi-phase flow multiple streams 3-5 to 3-9 of fluid mixtures 13-1

N natural convection. See flow, buoyancy driven NavCenter 2-38 near-wall Version 4.02

cell 3-12, 3-13, 3-14, 5-18 for compressible flow 3-10 dimensionless normal distance 1-7, 3-13 See also y+ values layer (NWL) 1-7, 3-12, 3-14 mesh distribution 1-7 region 3-13 neutral plot file 2-31, 17-4 new set selection option 2-21, 4-3 none set selection option 2-21, 4-3 non-Newtonian flow. See flow, non-Newtonian no-slip condition 4-19 NOVICE mode A-1 NOx modelling 8-39 defined by user subroutines 14-15 nozzle models 9-1 nuclei, number of 11-7 numerical discretisation error 1-6 Nusselt number 14-15

O on-line help 2-2 operate utility 13-4 operating mode 2-18 outer iterations. See iterations, outer outflow. See boundary conditions, outflow outlet. See boundary conditions, outlet output controls 5-2, 5-5, 5-12

P panels 16-2 to 16-6 creating 16-2 environment 16-7 files 16-5 manipulating 16-6 menus within 16-3 modifying 16-2 panels menu 2-17 define macros 16-7 define panel 16-2 parallel processing 2-5 for sliding mesh 12-22 run options F-3 to F-7 user subroutines 14-22, E-1 with cell layer removal/addition 12-18 with moving mesh 12-13 parameters 2-36 varying during run 17-18 to 17-20 parcels 9-2, 9-6 particle radiation. See radiation, coal particles patch number 4-4 radiation 7-2 to 7-7 surface 7-7 7

Index permeability function 1-10 PISO algorithm 1-13 to 1-15 under-relaxation 17-19 plot menu 2-17 alternate plot mode 2-31 background 2-32 cell display 4-41, 7-2, 9-5 maximum plot screen 2-32 plot to file 17-4 standard plot mode 2-31 standard plot screen 2-32 plotting hard copies 17-21 porous baffles 4-23, 4-25, 6-5 media in Eulerian multi-phase flow 10-2 in moving mesh 12-13 in multi-component mixing 13-2 user subroutines 14-8 pressure drops 6-5 region modelling 6-1 to 6-5 post menu 2-17 get droplet data 9-3 post register 13-4 post-processor 2-1 power law of viscosity 3-11 Prandtl number 8-17 defined by user subroutines 14-18 precision. See solver, precision pre-processor 2-1 pressure 4-5 boundary. See boundary conditions, pressure correction 1-13 to 1-15 drop across porous region 6-5 prescribed. See boundary conditions, pressure saturation 11-7 product 8-2 prolinkl script 17-14 prosize script 17-13 pro-STAR 1-2, 2-1 customisation 16-1 display D-1 to D-3 executables 17-14 launching 2-3, 2-10 layout D-3 to D-5 memory 17-13 on-line help 2-36, 17-7 quitting 2-21 resizing 17-13 size 2-18 pseudo-transient calculation 5-1 for compressible flow 3-11

Q quitting pro-STAR 2-21

8

R radiation 7-1 to 7-8, 15-2 analysis methods discrete ordinates 7-3 to 7-5, 7-7 to 7-8 discrete transfer 7-1 to 7-7 at walls 4-20 boundaries. See boundary conditions, radiation cell table editor 7-8 coal particles 7-5, 8-43 cpu time 7-6, 7-7 escape boundaries 7-5 FASTRAC. See FASTRAC in coal combustion 8-44 in Eulerian multi-phase flow 10-2 on baffles 4-25 participating media 7-3 patch 7-2 to 7-7 properties, defined by user subroutines 14-11 solar 7-1 baffle boundary conditions 4-26 discrete ordinates 7-8 in particpating media 7-3 wall boundary conditions 4-20 sub-domains 7-8 surface exchanges 7-1 transparent solids 7-3 to 7-8 user subroutines 7-6 with STAR-HPC 7-6 Rayleigh model 11-7 reactant 8-2 leading 8-2, 8-18 reaction advanced ICE models 8-22 to 8-39 CFM 8-24 to 8-25 ECFM 8-26 to 8-27, 8-30 ECFM-3Z 8-28 to 8-30 compression ignition 8-29 to 8-30 spark ignition 8-28 to 8-29 level set 8-31 to 8-32 saving data 8-32 background material 8-17 complex chemistry models 8-11, 8-21 coupled 8-16 eddy break-up reaction 8-13 Landau-Teller reaction 8-12 Lindemann fall-off reaction 8-12 SRI fall-off reaction 8-13 sub-timestep 8-21 three-body reaction 8-12 Troe fall-off reaction 8-12 conventions 8-18 copying 8-17 EGR systems 8-17 heterogeneous 8-1 homogeneous 8-1 in Eulerian multi-phase flow 10-2 Version 4.02

Index local source model 8-2, 8-16 models 8-1 NOx formation 8-39 partially premixed 8-1 PPDF scheme 8-3 to 8-11, 8-18 to 8-19 multi-fuel 8-9 single-fuel 8-3 equilibrium models 8-3 with dilutants 8-9 premixed 8-1 rate, defined by user subroutines 14-15 regress variable models 8-10, 8-16 eddy break-up models 8-10 flame-area models 8-10 CFM-ITNFS 8-10 Weller 8-10 Weller 3-equation 8-10 schemes 8-2, 8-14 soot modelling. See soot modelling source term 14-11 temperature limit 8-21 turning off 8-17 types 8-1 unpremixed/diffusion 8-1 user subroutines 14-15 real constants 5-3 recovery 2-20 re-executing commands 2-20 reference temperature 15-2 reflectivity of baffle 4-26 solar 4-26, 7-2 thermal 4-25, 7-1, 7-4 of wall 4-22 solar 4-21, 7-2 thermal 4-20, 7-1, 7-4 regress variable 8-10 relaxation factors 1-15, 1-16, 15-2 residuals 1-15, 1-19, 2-6, 2-7, 15-3, 17-19 for transonic flow 3-10 inner 1-13, 1-15 oscillations 1-19, 2-7 tolerance 8-16, 10-4, 15-1, 17-19 resistance, distributed 6-1 user subroutines 14-8 resizing pro-STAR 17-13 resource allocation F-6 restart 4-42 aeroacoustic analysis 13-4 after mesh changes 5-15 coal combustion 8-44 data 5-13 files 2-7 flamelet calculations 8-20 Lagrangian multi-phase 9-10 moving mesh 12-13 multiple runs 5-14 Version 4.02

non-reflective boundaries 4-18 run options F-2 steady-state runs 4-42, 5-4 transient runs 4-43, 5-6, 5-10 turbulence models 3-16 view factors 7-5 with INITFI 14-17 restoring sets 2-22 results checking 1-20 RESULTS sub-directory 2-7 resume mode 2-13 rotating reference frames arbitrary interface 12-4 coupling 12-8 defined by user subroutines 14-16 multiple explicit method 12-5 to 12-9 non-reflecting explicit option 12-9 implicit method 12-2 to 12-5 single 12-1 rotation 1-9 rotational speeds, defined by tables 2-25 rothalpy, in rotating reference frames 12-5 roughness 14-6 run time controls, defined by tables 2-25 running simulations 2-2, 2-4, 2-11 in parallel 2-5 on other hosts 2-5

S Sauter mean diameter 9-9 saving model 2-42 screen 2-33 sets 2-22 scalar CAV 11-6, 11-8 in fluid mixtures 13-1 initialisation 4-42 numbering 13-3 printing 13-2 properties, defining 8-15 variable 8-16 VOF 11-1, 11-6 initialisation 11-4, 11-9 scalar solver 1-19 scattering coefficient 7-4 Schmidt number 6-4, 8-17 defined by user subroutines 14-8, 14-18 screen capture 2-33 high-resolution 2-33 display control 2-18 dump 2-33 size 2-32 storage 2-32 9

Index set active cell type 3-3 sets restoring 2-22 saving 2-22 set-up files 16-1 shock wave 4-32 short cut keys. See function keys short input history 2-15 SIMPLE algorithm 1-13, 1-16 to 1-18 single precision mode 1-18 sliding mesh. See mesh, sliding solid regions 15-2 initialisation 4-42 solid-fluid heat transfer 3-8, 3-16 to 3-20, 15-2 hints 3-19 in free surface flows 11-2 radiative 7-3, 7-5 solid-solid heat-transfer 3-20 solution algorithms 1-13 to 1-18 for buoyancy driven flow 3-21 for cavitating flows 11-9 for use with DES turbulence models 3-15 for use with LES turbulence models 3-15 in free surface flows 11-4 controls 5-1, 5-5, 5-9 domain 1-2 mapping 5-15, 5-16 procedure 15-1 solver conjugate gradient 1-19 precision 1-18 for Eulerian multi-phase flow 10-3 for liquid films 13-6 tolerances 1-14, 1-15, 1-16, 15-2 soot modelling 8-39 to 8-41 flamelet library model 8-39 PSDF moments model 8-26, 8-29, 8-40 sound, speed of 14-12 source aeroacoustic 13-4 defined by tables 2-24 enthalpy 3-9 defined by user subroutines 14-10 in cavitating flows 11-9 in Eulerian multi-phase flow 10-3 in free surface flows 11-4 mass 3-8 defined by user subroutines 14-10 momentum 3-8 defined by user subroutines 14-10 scalar, defined by user subroutines 14-11 turbulence 3-9 defined by user subroutines 14-10 species mass fraction 8-4 defined by user subroutines 14-16 10

in coal combustion 8-44 reacting 8-2 specific heat 8-17, 8-19, 15-2 defined by user subroutines 14-9 in multi-component mixing 13-3 spin index 12-3, 12-6 parameters 12-1, 12-4 to 12-8 velocity 12-1, 12-4 spline set 2-21, A-2 stability numerical 1-5, 1-10 dependence on time step 5-9 stagnation boundary. See boundary conditions, stagnation STAR 1-2 defaults F-7 run options F-1 to F-12 running 2-4, 2-5, 2-11 switches 2-5, F-1 to F-12 STAR-GUIde 2-38 check everything panel 4-6 favourites 2-40 STAR-HPC, with radiation problems 7-6, 7-7 STAR-Launch utility 2-8 to 2-12 STAR-NET F-8 to F-12 STAR-View 17-23 StarWatch utility 17-15 to 17-21 states 17-6 steady-state calculation 1-15 to 1-17 stoichiometry, checking 8-16 strain rate, at inlet 8-20 subset set selection option 2-22, 4-3 surface tension 11-2 coefficient 14-12 surface set selection option 2-22 sweep limits 1-14, 1-15, 1-16, 15-2 sweeps 1-13, 17-19 switches 5-3 for ‘prostar’ system command 2-4 for ‘star’ system command 2-5, F-1 to F-12 symmetry plane. See boundary conditions, symmetry system commands, entering in pro-STAR 2-18

T tables dependent variables 2-27 editor 2-26 to 2-31 graphs of 2-29 hints 2-31 independent variables 2-27 title 2-27 usage in boundary conditions 2-24, 4-7 baffles 4-27 free-stream 4-33 Version 4.02

Index inlet 4-10 non-reflective 4-19 outlet 4-12 pressure 4-14 Riemann 4-37 stagnation 4-16 transient wave transmissive 4-35 walls 4-22 initial conditions 2-24 injectors and sprays 2-26 rotational speeds 2-25 run-time controls 2-25 source terms 2-24 tcl/tk interpreter 2-35 temperature at free-stream boundary 4-33 at Riemann boundary 4-37 at transient wave transmissive boundary 4-35 defined by user subroutines 14-7 devolatilisation 8-43 distribution 3-17 functional dependence 14-7 in cavitating flows 11-8 in free surface flows 11-3 limit on reaction 8-21 radiation 7-2, 7-4 reference 15-2 in restart runs 5-17 stagnation 3-11 under-relaxation 17-19 temporal discretisation. See discretisation, temporal thermal resistance 3-17 runaway 1-10 thermal radiation. See radiation time characteristic 1-14 cpu 2-18, 15-3 reducing 1-16 elapsed 2-7 elapsed computational 15-3 ignition delay 8-37 scale 5-9 heat/mass transfer 8-44 step 1-10, 5-6, 5-9 adjusting 1-14, 1-18 defined by tables 2-25 number of 15-1 specification 5-11 variable 14-18 varying during run 17-18 See also Courant number tools menu 2-16 cell tool 3-3 check tool 3-5 colour tool 17-22 convert 17-8 Version 4.02

users tool 2-35 transient calculation 1-13 to 1-15, 3-11 completion 2-7 full 5-6 to 5-14 single 5-4 to 5-6 transient wave boundary. See boundary conditions, transient wave transmissive transient waves 4-34 transmissivity at solid-fluid interface 7-5 of baffle 4-26 solar 4-26, 7-2 thermal 4-25, 7-1, 7-4 of wall 4-22 solar 4-21, 7-2 thermal 4-20, 7-1, 7-4 turbomachinery, boundaries for 4-16 turbulence 3-12 to 3-16 changing model 3-16 DES models 3-15 hybrid wall functions 3-12, 3-14 in aeroacoustic analysis 13-4 in ECFM combustion models 8-30 in porous media 6-4 in rotating reference frames 12-2, 12-4, 12-9 initialisation 4-42 length scale 14-9 LES models 3-15 low Reynolds number models 3-12, 3-14 models 3-12, 15-2 Reynolds stress models 3-15 conditions at boundary free-stream 4-33 inlet 4-10 Riemann 4-37 stagnation 4-15 transient wave transmissive 4-35 two-layer models 1-8, 3-12, 3-13 to 3-14 wall functions 3-12, 3-13 non-equilibrium 3-13 tutorials 2-37 two-dimensional flow, axisymmetric. See flow, axisymmetric two-phase flow. See Lagrangian multi-phase flow and Eulerian multi-phase flow

U under-relaxation 1-10, 1-15, 3-20 density 1-15 for compressible flow 3-10, 3-11 for moving mesh 12-13 for steady-state calculations with PISO 1-15 for steady-state calculations with SIMPLE 1-16 for transient calculations with SIMPLE 1-17 in buoyancy driven flow 3-21 in cell layer removal/addition 12-18 11

Index in chemical reaction problems 8-16 in Eulerian multi-phase flow 10-4 in multi-component mixing 13-3 pressure 1-16 pressure correction 1-14 varying 17-19 velocity 1-16 viscosity 1-15 units 2-36, C-1 unselect set selection option 2-21, 4-3 unsteady calculation. See transient calculation user interface 2-35 user subroutines 2-4, 2-18, 2-36, 14-1 to 14-22, 15-2 activating 14-2 checking 1-20 defining material properties 14-6 to 14-9 editing 14-4, 17-10 for boundaries 4-7, 14-5 for chemical reactions 14-15 for Eulerian multi-phase flow 14-14 for heat and mass fluxes 14-6 for heat and mass transfer coefficients 13-3, 14-17 for Lagrangian multi-phase flow 9-10, 14-12 for moving mesh 12-9, 14-16 for porous media 6-4, 14-8 for rotating reference frames 12-1, 12-4, 12-8, 14-16 for solar radiation 7-6 for turbulence 14-9, 14-10 in droplet injection 9-1 to 9-3 in parallel processing runs 14-22, E-1 users tool 2-35 utility menu 2-17 calculate volume 3-22 capture screen 2-33 count 4-2 function keys 16-10 save screen as 2-33 solution mapping 5-16 user subroutines 14-2 write STAR-CD scene file 17-23

V vaporization rate 14-12 vapour in cavitating flows 11-7 variables 2-36 vector solver 1-19 velocity angular. See angular velocity at stagnation boundaries 4-15 boundary values 4-27 characteristic 1-14 in porous media 6-5 injection 3-22 defined by user subroutines 14-10 of baffles 4-27 of liquid films 13-7 12

of walls 4-22 vertex coordinate in moving mesh 5-14, 12-9 data 17-3, 17-9 maximum number of 2-18 set 2-21, A-2 view factor 7-3, 7-5 viscosity defined by user subroutines 14-9 in chemical reaction problems 8-17 oscillations 1-15 power law 3-11 turbulent 14-9 under-relaxation 1-15 viscous sublayer 3-13 volume of fluid (VOF) model 11-1 volume, of droplet 9-10

W wall 1-3 boundary layer 1-6 data 5-3 functions 1-7, 3-12 hybrid 1-8 heat flux 5-5 defined by user subroutines 14-17 moving 4-19 no-slip 4-19 patch 7-2 radiation 4-20 to 4-22 transmissive external 7-5 transparent 4-21 velocity 4-22 See also boundary conditions, wall wave compression 4-32 expansion 4-32 shock 4-32 transient 4-34

Y y+ values 1-21, 3-13, 3-15 See also near-wall, dimensionless normal distance

Z Zeldovich mechanism 8-39

Version 4.02

INDEX OF COMMANDS

INDEX OF COMMANDS This User Guide does not contain comprehensive information on all commands used in pro-STAR. The Meshing and Post-Processing Guides and on-line STAR GUIde Help should also be consulted

A

CZONE 3-4

ABBREVIATE 16-1 ABORT A-1

D

B BATCH 2-18 BCROSS 4-2, 7-6 BDEFINE 4-2, 7-2 BDELETE 4-4 BDISPLAY 4-41, 7-2 BGENERATE 4-2 BLIST 4-4 BLKSET 2-24 BMODIFY 4-2, 4-4, 7-6 BSET 2-24, 4-3, 4-4 BSHELL 4-2, 7-2

C CASENAME 17-1 CAVERAGE 12-29 CAVITATION 14-11 CAVNUCLEI 14-11 CAVPROPERTY 14-12 CBEXTRUDE 3-18, 3-19 CCLIST 17-6 CCROSS 3-4 CDELETE 12-29 CDISPLAY 4-41, 7-2, 9-5 CDSAVE 17-3 CDSCALAR 13-3 CDTRANS 5-11, 17-4 CFIND 3-4 CJOIN 5-19 CLOSE 17-10 CLRMODE 2-32 CLRTABLE 17-22 CMODIFY 3-5, 16-5 COKE 14-18 CONDUCTIVITY 14-6 COUNT 4-2 CPLOT 12-12 CPOST 5-11 CPRANGE 5-11 CPRINT 5-11 CPSET 2-24 CREFINE 16-4 CRMODEL 14-16 CSET 2-24, 12-29 CTABLE 3-2, 12-15, 12-24 CTCOMPRESS 3-2, 3-4 CTDELETE 3-2 CTLIST 3-2 CTMODIFY 3-2 CTNAME 3-2 CTYPE 3-4, 16-5 CURSORMODE 2-19 CYCLIC 4-29 CYCOMPRESS 4-31 CYDELETE 4-31 CYGENERATE 4-29 CYLIST 4-31 Version 4.02

DAGE 9-7, 9-8 DCOLLISION 14-12 DELTIME 14-18 DENSITY 14-8 DIFFUSIVITY 14-8 DLIST 9-8 DRAVERAGE 14-12 DRCMPONENT 14-13 DRHEAT 14-13 DRMASS 14-13 DRMOMENTUM 14-13 DRPROPERTIES 14-13 DRUSER 14-13 DRWALL 14-13 DSCHEME 14-19 DSET 2-24, 9-6, 9-8 DTIME 9-5, 9-7

E EACELL 12-16 EACOMPRESS 12-20 EADELETE 12-20 EAGENERATE 12-20 EALIST 12-20 EAMATCH 12-19, 12-21, 12-27, 12-29 EATTACH 12-20, 12-21, 12-27 ECHOINPUT 2-19 ECLIST 12-16 ECONDITIONAL 12-26 EDATA 5-3 EDCELL 12-16 EDCOMPRESS 12-26 EDDELETE 12-26 EDDIR 12-16 EDETACH 12-21, 12-28 EDLIST 12-26 EDRAG 14-14 EECELL 12-28 EFLUID 12-26 EGRID 12-11, 12-13, 12-16, 12-29 to 12-30 EHTRANSFER 14-15 EICOND 14-17 ETURB 14-14 EVCHECK 12-30 EVCND 12-26 EVCOMPRESS 12-12, 12-17, 12-20 EVDELETE 12-12, 12-17, 12-20, 12-26 EVEXECUTE 12-29 to 12-30 EVFILE 12-11, 12-16, 12-25 EVFLAG 12-30 EVGET 12-12, 12-17, 12-20, 12-26 EVLIST 12-12, 12-17, 12-20, 12-27 EVLOAD 12-29 to 12-30 EVOFFSET 12-12, 12-17, 12-20, 12-27 EVPARM 12-10 EVPREP 12-12, 12-17, 12-20, 12-27 EVREAD 12-12, 12-17, 12-21, 12-27 EVSAVE 12-11 EVSLIDE 14-16 1

Index of Commands EVSTEP 12-11, 12-16 EVUNDELETE 12-12, 12-17, 12-20, 12-27 EVWRITE 12-12, 12-17, 12-21, 12-27 EXPERT A-1

F FSTAT 2-23, 17-10

G GEOMWRITE 12-12, 12-17, 12-20, 12-27, 17-5 GETCELL 12-12 GETD 8-36

H HCOEF 14-17 HISTORY 2-19 to 2-20 HRSDUMP 2-33

I IFILE 13-3, 17-2, 17-4, 17-10, 17-12 IGNMODEL 14-15 INITIAL 14-17 ITER 5-14

O OFILE 17-10 OPANEL 16-1, 16-6 OPEN 16-10

P PAGE 2-19 PATCH 7-7 PLATTACH 12-30 PLTBACK 2-33 PLTYPE 12-12 PMATERIAL 12-25 POPTION 12-12 POREFF 14-8 POROSITY 14-8 PORTURBULENCE 14-8 PRESSURE 12-18, 12-25 PRFIELD 14-19 PROBLEMWRITE 12-12, 12-17, 12-20, 12-27, 17-7 PROMPT 16-7 PRTEMP 3-18 PTCONV 17-8 PTPRINT 9-7, 9-8 PTREAD 9-7

Q QUIT 2-21, 2-42

K KNOCK 14-15

L LFQSOR 14-14 LFSTRIP 14-14 LQFBC 14-14 LQFINITIAL 14-14 LQFPROPERTY 14-14 LSCOMPRESS 5-11 LSDELETE 5-11 LSGET 5-11 LSLIST 5-10 LSRANGE 5-11 LSSAVE 5-10 LSTEP 5-10, 14-18 LVISCOSITY 14-9

M MACRO 16-9 MEMORY 17-14 MFRAME 14-16 MLIST 3-9 MMPISTON 12-13 MONITOR 12-18, 12-25 MVGRID 5-11, 12-9, 12-10, 12-16, 14-16, 14-17

N NFILE 17-4 NOX 14-15 2

R RADPROPERTIES 14-11 RCONSTANT 12-18 RDEFINE 12-19, 12-23, 14-5, 14-6 RECALL 2-20 RECOVER 2-20, 2-42 REPEAT 16-10 REPLOT 16-3 RESET 2-32 RESUME 2-42, 17-3, 17-10 REWIND 17-10 RGENERATE 4-7 RRATE 14-15 RSOURCE 14-10 RSTATUS 8-17

S SAFETY 2-20 SAVE 2-42, 17-2 SC 13-3 SCCONTROL 13-3 SCDUMP 2-33 SCENE 17-23 SCPOROUS 14-8 SCPROPERTIES 13-3, 14-16, 14-18 SCRDELETE 2-33 SCRIN 2-32 SCROUT 2-32 SCSOURCE 14-11 SCTRANS 5-11, 13-2 SETADD 2-23 SETDELETE 2-22 Version 4.02

Index of Commands SETENV 16-6 SETFEATURE 16-1 SETREAD 2-23 SETWRITE 2-22 SIZE 2-18 SMAP 5-15, 5-16 SMCONV 17-8 SOLAR 14-11 SPECIFICHEAT 14-7, 14-9 SPIN 14-16 SPLSET 2-24 STATUS 2-42, 12-25 STENSION 14-12 STORE 12-12 SUCCEED 2-20 SYSTEM 2-18

T TBCLEAR 2-28 TBDEFINE 2-28 TBGRAPH 2-28, 2-30 TBLIST 2-30 TBMODIFY 2-30 TBREAD 2-30 TBSCAN 2-31 TBWRITE 2-28 TDSCHEME 5-11 TERMINAL 2-31, 17-4, 17-21 TEXT 2-18 TIME 12-10, 12-15, 14-18 TITLE 2-19 TLMODEL 14-9 TPRINT 2-18 TRFILE 2-42, 5-10, 17-4 TRLOAD 12-12 TSCALE 2-33 TSMAP 5-16 TURBULENCE 14-9

U USER 2-18 USUBROUTINE 14-3

V VAPORIZATION 14-12 VFILL 12-11 VLIST 16-5 VMOD 12-11 VMODIFY 2-24 VOLUME 3-22 VSET 2-24, 12-11 VSMOOTH 12-30

W WHOLE 2-32 WIPEOUT 17-14 WPOST 5-11 WPRINT 5-11

Version 4.02

3

Star CCM+ User Guide

Recommend Documents