SE21 Steel Bridge decks:Layout 1
1/11/07
12:34
Page 49
paper: bright/smith
A new design for steel bridge decks using laser fabrication Synopsis Steel orthotropic decks provide a lightweight form of construction, essential for weight-critical structures. However, their cost and poor record of fatigue durability has discouraged their use for mainstream construction. As a result, steel decks are generally considered an option of last resort, only used where the minimisation of self-weight is essential, such as long-span and moveable bridges. An innovation is proposed to overcome these problems and transform the design of steel decks. The innovation is based upon the use of laser welding to produce an enclosed ‘sandwich panel’ profile. The sandwich profile shown below overcomes many of the constraints to structural performance associated with the use of conventional orthotropic steel decks.
Deck plate
Continuous laser welds
Hot rolled I-section
Lower flange
As well as enhancing the global structural characteristics of the deck, fatigue durability is further enhanced by the laser welding process used to fabricate the panels. An extensive range of fatigue tests are presented on the welds used to join the panels. A jointing arrangement is described which offers outstanding fatigue strength and exceptional economy in assembly and fabrication. Laser welds are extremely cheap to produce and the use of proprietary rolled sections offers significant savings in
construction costs compared to conventional steel deck designs. The sandwich design has the potential to provide a competitive, lightweight alternative to the concrete decks used in general bridgeworks. Proprietary sandwich panels may also be an attractive proposition for deck replacement schemes where the introduction of a lightweight deck could facilitate an increase in the live load capacity of understrength bridges. Further testing of a complete deck assembly is required to exploit the future potential of the system.
Introduction The design of steel bridge decks has remained essentially unchanged for the last 40 years. Refinements have been made but numerous problems remain. The current generation of steel ‘orthotropic decks’ use a thin deck plate, reinforced with opensection stiffeners or closed troughs spanning 2-4m between crossgirders. A typical deck design is illustrated in Fig 1. The capital cost of a steel deck is typically around four times greater than an equivalent concrete deck1. Fabrication costs significantly exceed material costs on most steel decks. In particular, the manufacture and fitting the decks trough stiffeners requires a considerable amount of fabrication effort. The non-symmetrical distribution of weld metal about the neutral axis of the section also causes the deck to distort in both longitudinal and transverse directions during welding. Significant costs are also incurred in joining the deck to its supporting superstructure. The deck of a modern orthotropic bridge acts integrally with the main girders of the bridge and its intermediate cross-girders (see Fig 2). Maintaining continuity between the deck stiffeners and the cross-girders requires complex and cumbersome jointing details. Typical details are illustrated in Fig 3. The stress concentrations associated with the complex joint
1
S. R. Bright PhD, BEng, CEng, MICE Cass Hayward LLP, York House, Welsh Street, Chepstow
J. W. Smith† BSc (Hons), PhD, CEng, FIStructE, FICE Dept. Civil Engineering, University of Bristol, Bristol
†
Deceased
Received: 04/07 Modified: 07/07 Accepted: 08/07 Keywords: Steel, Bridge decks, Orthotropic, Design, Sandwich panels, Comparing, Lasers, Welding, Fatigue © S. R. Bright & J. W. Smith Your comments on this paper are welcome and will be published online as Correspondence.
2
3a
Deck-to-cross girder connection Cross girder
3c Open, torsionally weak stiffeners
3b Welding continuity interrupted by cross girder Closed, torsionally stiff stiffeners
Fig 1. Typical orthotropic deck design using closed trough stiffeners / Fig 2. View of underside of conventional orthotropic deck highlighting deck-to-cross girder connection / Fig 3. Typical connection details between stiffener webs and cross-girder on conventional orthotropic decks and photo showing typical weld process 6 November 2007 – The Structural Engineer|49
SE21 Steel Bridge decks:Layout 1
1/11/07
12:34
Page 50
paper: bright/smith
4a
5a
4b
4c
5b
5c
4d
patterns used in conventional orthotropic decks have given rise to numerous examples of premature fatigue cracking in steel decks worldwide. Although these do not endanger the overall structure, in situ repairs are difficult and expensive procedures. The deck-to-cross girder joints have given rise to fatigue problems on numerous bridges. Cracking has been observed at the intersection between the stiffener webs and cross-girders2,3, adding to the whole life cost of steel decks. Heavily trafficked bridges have also suffered extensive fatigue damage at the welded joint connecting the stiffener troughs to the deck plate4. The torsional stiffness of the closed troughs generates out-of-plane bending stresses in the webs as illustrated in Fig 4. The fatigue durability of the web-to-deck joint is adversely affected by the range of possible wheel positions on the deck. Moment reversals may occur for differing wheel load positions. The overall stress range experienced by the joint could thus be double the (static) values indicated in Fig 4. Modern design codes5 require the use of partial penetration butt welds to derive some improvement in fatigue strength compared to the fillet welds used on most early bridges. However, durability problems are still to be expected on many structures within a 120 year design life. Fatigue cracking has also been experienced in the surfacing of many orthotropic bridges6,7. Steel decks are inherently flexible and individual wheel loads of heavy lorries generate local bending curvatures which may be sufficient to crack ordinary highway asphalts. Surfacing cracks usually follow the pattern of the deck stiffeners, confirming the relationship between
deck flexibility and surfacing deterioration. Cracking has also been experienced in the deck plate itself as shown in Fig 5. Experience has shown the surfacing of steel decks requires repair or replacement much more frequently than stiffer concrete decks. Steel decks are generally surfaced with a heavy duty material containing a large proportion of bitumen and with more careful grading of aggregate and fines. Mastic asphalt is the preferred material on UK steel bridges in a layer of about 40mm thickness. In European practice it is more common to use two layers, of greater overall thickness, with the upper layer being a similar material to mastic asphalt. In the U.S., much use has been made of epoxy asphaltic concrete. Nevertheless, the practical life of the surfacing is much less than the steel deck itself. Replacing the surfacing on a bridge is much more disruptive than on a normal highway since the carriageway width is usually restricted. Traffic management, delay and disruption costs are thus generally expected to be greater than the costs of the repair work itself. The net result of all these manufacturing and operational problems is that orthotropic decks are considered an ‘option of last resort’, only used where minimisation of self-weight is essential, such as long-span and moveable bridges. In such cases, bridge owners expect to pay a premium in terms of capital and whole-life cost for the use of a steel deck. The innovation proposed in this paper is intended to overcome the durability and whole-life cost problems associated with current steel deck designs. By combining these improvements with a reduction in the cost of production of the deck, it is also anticipated that the new design could compete with 7
6
Fig 4. Transverse deformations and torsions in conventional trough-stiffened deck and corresponding bending stresses in webs under 2 × 20kN wheel loads /Fig 5. Surfacing cracks in conventional orthotropic deck and fatigue cracking of deck plate/ Fig 6. Deck design details proposed by Eurocode 3 / Fig 7. Typical weights of conventional US orthotropic decks (after Wolchuk8) 50|The Structural Engineer – 6 November 2007
SE21 Steel Bridge decks:Layout 1
1/11/07
12:34
Page 51
paper: bright/smith
8a 8b
9a
9b
general forms of concrete decks typically used in short-to-medium span bridging applications. The new form of deck could also be an attractive proposition for deck replacement schemes where the introduction of a lightweight deck could facilitate an increase in the live load capacity of under-strength bridges.
Even with inclined webs it is impractical to join the troughs to the deck plate with full penetration butt welds because access is restricted to the outside of the troughs. Fillet welds or partial penetration butt welds have thus traditionally been used, limiting the fatigue life of the joint. Fatigue durability problems are thus to be expected on even the most modern decks.
Current state of the art in steel deck designs
A new design concept for steel bridge decks
Since the 1960s, steel deck designs have principally evolved around the use of closed-trough stiffeners. Recommended design details proposed by the recent Eurocode 3 are illustrated in Fig 6. The roadway deck comprises a steel plate, generally 12 to 14mm thick, with a traffic wearing surface of high performance bituminous surfacing varying typically between 40mm to 80mm thick. In Europe, thinner deck plates combined with a thicker wearing course of gussasphalt and asphaltic concrete are commonly preferred. In the UK and Australia a thinner wearing course is very common but a thicker deck plate has been found to be necessary to provide sufficient fatigue durability. Trapezoidal trough stiffeners 6mm to 8mm thick are welded to the deck plate; the profile shown in Fig 6 being typical. The flat section of trough provides a bottom flange in bending and helps to keep the neutral axis at a reasonable depth below the deck plate. V-plate stiffeners were adopted for the Severn Crossing and Humber Bridge. These have the advantage of requiring only one bend in their manufacture and allow smaller cut-outs in the cross-girder to accommodate the stiffeners. However, V-troughs are much less efficient in bending. In the early history of steel bridge decks, the stiffeners were often of the open type, being I-sections, flats or bulb flats. These are more flexible in the transverse direction than trough stiffened decks since they lack torsional stiffness, which assists in the transverse distribution of local loads. Consequently, more material is required to achieve a deck of equivalent stiffness. Wolchuk8 compiled a comparison of the weights of the two types of deck based on historical US details with relatively thin deck plates. Wolchuk’s summary is reproduced in Fig 7. Rib stiffened decks are generally much heavier than an equivalent troughstiffened deck. However, rib-stiffened decks offer some long-term durability benefits since the lack of torsional stiffness of the ribs relieves transverse bending stresses at the welds, resulting in superior fatigue performance. However, twice as many welds are required to fabricate a rib-stiffened deck. Trough stiffened decks have also been preferred for their relative ease of welding. Conventional welding techniques rely upon sufficient clearance to be provided at various points around the cross section to maintain access for a welding torch or rod, see Fig 8.
Given the recorded extent of problems associated with existing steel decks, there have been many attempts to rationalise design details. However, the constraints of conventional welding processes limit the geometric profiles which can be produced, preventing any significant advances in design or construction. With current technology, it is uneconomic and impractical to fabricate a steel deck with equal (or even similar) stiffnesses in orthogonal directions. An ideal bridge deck should be as near isotropic as possible, with roughly equivalent longitudinal and transverse stiffnesses. This ensures localised wheel loads are distributed in both longitudinal and transverse directions as effectively as possible, limiting the magnitude of stress concentrations in the vicinity of the load and thereby mitigating potential fatigue ‘hot-spots’. Concrete slab decks are essentially isotropic in nature but steel decks have always been highly orthotropic with the intermittent troughs giving the deck a high longitudinal bending stiffness, but a negligible transverse stiffness between troughs. The poor load distribution of steel decks results in high peak stresses within directly loaded stiffeners, with negligible stresses elsewhere. The transverse stiffness of a steel deck could be increased significantly by adding a continuous plate along the underside of the deck. This lower flange creates an enclosed ‘sandwich profile’ with a significantly increased, and more uniformly distributed, torsional stiffness. Transverse shear stiffness would also be considerably improved, particularly in decks with triangulated (or ‘trussed’) web profiles. A range of possible sandwich profiles are illustrated in Fig 9. Fully-triangulated sandwich panels have already been trialled in other materials such as composite plastics as shown in Fig 9. The addition of the lower flange causes a downward shift in the neutral axis of the section, improving its longitudinal bending efficiency and improving the transverse distribution of loads. The results of a 3D-FE analysis comparing longitudinal and transverse bending stresses in a conventional orthotropic deck and an alternative sandwich panel deck are summarised in Fig 10. A comparison of the structural performance of the two deck types is provided in Table 1. Note that the comparison was based on a 3.6m simplysupported span for the trough-stiffened deck. A 25% greater, 4.5m simply supported span was assumed for the sandwich panel. Longitudinal bending stresses are significantly reduced in the sandwich
Fig 8. Geometric constraints associated with conventional welding techniques/ Fig 9. Addition of lower flange to create an enclosed ‘sandwich panel deck’ and view of composite plastic panel recently trialled by the UK Highway’s Agency. Various potential web configurations illustrated 6 November 2007 – The Structural Engineer|51
SE21 Steel Bridge decks:Layout 1
1/11/07
12:34
Page 52
paper: bright/smith
Table 1: Comparison of structural performance Deck type
Closed trough orthotropic deck
Longitudinal span
3600mm
4500mm (+)
Second moment of area (per m width)
1.1 × 108mm4/m
4.3 × 108mm4/m
Basic mass/m2 (average)
165kg/m2
290kg/m2
2
Typical cross girder mass/m 2
Sandwich panel
2
0*kg/m2
2
50*kg/m
Total deck weight/m (typical)
215kg/m
290kg/m2
Displacement under 2 × 20kN wheel loads
1.60mm
0.75mm
Span:displacement ratio
2250:1
6000:1
Max. longitudinal bending stress: a) in deck plate (offset wheel loads) b) in lower flange/trough
–44N/mm2 +57N/mm2
–25N/mm2 +15N/mm2
Maximum web bending stress
± 34N/mm2
± 25N/mm2
Max. transverse bending stress: a) Hogging stress in deck plate b) Sagging stress in deck plate
+100N/mm2 – 59N/mm2
+75N/mm2 – 50N/mm2
* For derivation see text
panel. A span of 6m should generate comparable deck stresses to those observed in the 3.6m long trough-stiffened deck. A span of 8.4m would generate similar stresses in the bottom flange. A typical span of 7.2m for the sandwich panel deck is thus foreseeable, exactly twice that of the conventional trough-stiffened deck. The sandwich deck is likely to be heavier than an equivalent trough-stiffened deck. However, this additional mass could be offset by a reduction in the number of cross girders required for the sandwich panel. The increased span range of the panels could be used to reduce the number of cross girders within the bridge superstructure, or alternatively, allow the cross girders to be omitted altogether. This is discussed further later in this paper. Cross girders typically contribute around 25-35% to the weight of a conventional trough-stiffened deck. These efficiencies in cross-girder design and further optimisation of the sandwich panel design could make the overall weight of the deck comparable to conventional orthotropic decks. These results show the sandwich design could significantly improve the structural performance of steel decks. However, the production of enclosed profiles is impractical with conventional fabrication techniques unless plug or resistance spot-welds are used. These welds are expensive to produce and
do not possess sufficient fatigue strength for use in bridge decks. An alternative method of manufacture has thus been devised, laser welding.
Production of steel sandwich panels by laser welding Laser welding is a relatively new technique for ‘heavy’ sections but quite well established in industries involving light fabrication. Laser welds can be produced by either Nd:YAG or CO2 lasers. Nd:YAG lasers are cheaper and more versatile, but are currently limited to maximum output powers of about 5kW. CO2 lasers are available with output powers of up to 25-30kW. CO2 lasers are thus currently best suited for heavy fabrication applications such as bridge decks. A schematic diagram of a CO2 laser welding plant is shown in Fig 11. A CO2 laser source is used to generate the beam, which is delivered to the fabrication site by means of mirrors, and finally focused at the work piece. In addition, another nozzle discharges helium gas to protect the fusion process. A low power Nd:YAG laser is also shown in Fig 12 for comparative purposes. Enough heat is produced by the focussed laser beam to create a molten pool of metal between the parts to be connected. The basic process illustrating the formation of a laser butt weld is shown in Fig 13. Welding is achieved by local fusion of the two parts of the joint. Heat input is much less than in arc welding and shrinkage is consequently less of a problem. The delivery tool is generally automatically controlled and great precision is possible. Laser welds are often referred to as ‘keyhole’ welds due to the narrow width of the joint and the remote access to the joint interface achieved by the laser. The ability of lasers to fuse components remotely by this keyhole process is pivotal to the development of (enclosed) steel sandwich panels. Such novel jointing techniques are essential for materials such as steel which cannot be extruded to form enclosed (hollow) profiles such as those available in aluminium and other materials. The design proposed makes use of laser ‘stake welds’ to produce a continuous joint between the outer flange plates and the internal core of the panel. Stake welds differ from conventional butt-joints as they are formed by extending the molten pool generated by the heat of the laser beam through the full-depth of one section and into an underlying plate. Upon the almost instantaneous cooling of the weld, the two plates become rigidly joined, see Fig 14. The laser is able to create a continuous, rigid joint between the flange and the core of the panel even though the beam has no direct access to the core material. These novel stake-welded joints are unique to the field of laser fabrication. Laser welds can also be used to produce discrete, intermittent joints,
10a
11b
10b
11a
Horizontal and vertical mirror in gantry
Four additional mirrors provided within delivery pipework from laser generator to gantry
Three mirrors in delivery head
Fig 10. Comparison of mid-span stresses in a conventional orthotropic deck with an I-section sandwich panel/ Fig 11. Schematic of 25kW CO2 laser delivering power to the workpiece by a combination of mirrors and view of practical laser delivery system 52|The Structural Engineer – 6 November 2007
SE21 Steel Bridge decks:Layout 1
1/11/07
12:34
Page 53
paper: bright/smith
12
14 Small diameter optic fibre delivery cable
Remotely operated/ programmed robotic arm
Laser delivery head
13
15
16
20kW CO2 laser Lower flange (inverted) Laser (stake) weld joining lower flange to I-section
Deck plate (inverted) Hot rolled I-section
but for fatigue durability in bridge deck applications, only continuous weld profiles have been considered. For these initial trials, prototype panels were produced with linear welds as well as continuous, sinusoidal profiles. The versatility of the laser welding process could allow sandwich panels to be produced in almost any length – their size is only limited by available rolling mill plate sizes, the geometry of the enclosure housing the laser welding apparatus, and the practicalities of transporting the finished panels. The internal core could consist of any of the profiles illustrated in Fig 9, or any profile which offers intimate contact between the top of the section and the flange plate. Almost any conceivable configuration of internal web members could thus be produced by laser welding.
Sandwich panel design for bridge deck applications Bridge decks are subjected to heavy, localised loads. The choice of web profiles for bridge applications is therefore limited by the need to provide a core with sufficient stability to resist significant compressive loads. Fully triangulated or ‘offset trough’ web profiles may be suitable but are relatively expensive to produce because of the fabrication effort associated with their plate preparation and fit-up. Discontinuous web profiles can be
formed using standard hot-rolled steel profiles which require little or no preparation before fabrication. Cold-formed sections could also be considered but are unlikely to prove as economic as hot-rolled sections. Further economies in fabrication could be derived by using web profiles which are inherently stable during assembly to limit the amount of fit-up and restraint required prior to welding. A hot-rolled I-section is an attractive proposition being economic to produce and inherently stable during fit-up. The width of flange at the top of the I-section also offers tolerance in the targeting of the laser when operated remotely from outside the panel. The laser welding of a prototype deck specimen is illustrated in Fig 16. The use of an I-section web also offers the opportunity to develop jointing details between the web and the flange plates with considerable fatigue strengths. Details of the jointing arrangements investigated by the authors are described in the following section.
Fatigue durability of laser welded joints Laser welded joints have been found in previous investigations9 to offer excellent fatigue strengths in butt-joint applications. However, the proposed sandwich panel design is reliant upon the use of ‘stake-welded’ joints which
Fig 12. A 2kW Nd:YAG laser delivering power to the workpiece by a fibre-optic cable directed by a robotic arm / Fig 13. Formation of a full- or partial-penetration butt joint using laser welding / Fig 14. Stake welded joint between flange plate and core/ Fig 15. Forms of continuous laser stake welds produced for prototype deck specimens/ Fig 16. Fabrication of prototype specimens using I-sections and linear stake welds 6 November 2007 – The Structural Engineer|53
SE21 Steel Bridge decks:Layout 1
1/11/07
12:34
Page 54
paper: bright/smith
17
18b
18a
have not previously been subjected to any form of fatigue testing. A number of fatigue tests have thus been undertaken on joint details devised by the author. The precise configuration of welds adopted in the joints could have a significant influence on their resulting fatigue strengths. The preferred design developed from this initial research utilises I-section web members to allow the panel webs to be joined to the flange plates by a pair (or more) of widely spaced welds as illustrated in Fig 17. Fig 17 illustrates the configuration of laser welds adopted in the majority of the fatigue tests undertaken as part of this research. Two types of test were undertaken to simulate the transverse bending stresses predicted in a sandwich panel deck. Diagrammatic representations of the two tests are provided in Fig 18. Further details of the tests are available in a previous paper by Bright and Smith and from a thesis by Bright10.
a) Web bending fatigue test results The first set of tests simulated the effect of web bending stresses on the fatigue durability of the welded joint. Web bending stresses are known to be a possible cause of fatigue failure in existing steel decks whereby fatigue cracking has propagated from the root of the welds joining the stiffeners to the deck plate (see Fig 19). Alternating stress patterns at the web/deck interface were simulated by connecting the web of a cut-section of panel to a hydraulic ram operating at alternate tension and compression (see Fig 18). This generated opposing axial (push-pull) forces in the welds, as well as significant coexistent shears. Despite the significant stress levels predicted by 2D-FE models of the welds, all fatigue failures in the web bending tests were found to occur in the parent metal of the webs. No weld failures were observed in any of the tests. The design proposed thus offers exceptional fatigue durability in terms of web bending stresses. A comparison of the web bending test results with the BS 5400 mean SN curves is presented in Fig 20.
b) Deck bending fatigue test results The second set of tests simulated the effect of transverse bending stresses in the deck plate. The presence of the I-section flange beneath the deck plate was shown by FE models to generate a significant reduction in transverse
deck stresses compared to a conventional orthotropic deck (see Fig 9(b)). However mobilising this composite interaction between the deck plate and the flange beneath would require significant shear stresses to be conveyed by the laser welds joining the two elements. These varying weld stresses were a potential source of fatigue damage. It was found by testing that the transverse bending of the deck plate could eventually produce fatigue failures through the welds and so provides a practical limit on fatigue durability. The typical pattern of fatigue cracking observed in the deck bending tests is illustrated in Fig 21. A comparison of the deck bending test results with the BS 5400 mean SN curves is presented in Fig 22. A Class C-D designation for the welds is suggested against bending stresses in the deck plate. A higher classification is suggested for lower stress ranges such as those which may be reasonably expected in-service. These preliminary fatigue tests thus indicate a considerable improvement in fatigue durability compared to similar joints used in conventional orthotropic decks. Full scale fatigue tests on the proposed deck design should be undertaken in due course following further design development and optimisation.
Economic viability of sandwich panel decks The simplicity of assembly and welding could make the proposed sandwich panel design an attractive proposition for mass production. The construction sequence adopted in the production of the prototype test specimens is illustrated in Fig 23. The prototype decks were welded in a continuous operation at speeds of up to 2m/minute. Fig 24 includes a sequence of photographs showing the rate of progress of the laser welds placed in the fabrication of the prototype deck specimens. Fit-up and plate preparation required the simple introduction of G-clamps at opposing ends of the deck with intermediate tack welds along each beam. For commercial mass-production, these restraints could be replaced by forced compaction ahead of the welding nozzle. The welds required no consumables such as filler wire and resulted in significantly less distortion than conventional weld processes due to the reduced heat input derived from the laser. The prototype decks were manufactured with a laser output power of 20kW. Industrial lasers generally operate at around 10% efficiency, suggesting an input power around 200kW. At an industrial user rate of 4p/kW hour,
Fig 17. Proposed arrangement of laser welds with I-section webs/ Fig 18. Transverse Virendeel deformations in discontinuous web sandwich panel and representation of web and deck transverse bending stresses in fatigue tests 54|The Structural Engineer – 6 November 2007
SE21 Steel Bridge decks:Layout 1
1/11/07
12:34
Page 55
paper: bright/smith
19
20a 20c
Bending stress range in web (N/mm2)
20b
22 Bending stress range indeck plate (N/mm2)
21
23
Fig 19. Typical fatigue crack patterns in webs of conventional orthotropic decks/ Fig 20. Typical fatigue failure in web bending test and comparison of web bending test results with BS 5400 mean SN curves/ Fig 21. Fatigue crack pattern in deck bending test/ Fig 22. Comparison of deck bending test results with BS 5400 mean SN curves / Fig 23. Proposed construction sequence and suggested sandwich panel design 6 November 2007 – The Structural Engineer|55
SE21 Steel Bridge decks:Layout 1
1/11/07
12:34
Page 56
paper: bright/smith
24
25
26
27
Fig 24. Bottom flange being welded onto I-section at 1500mm/min. Time elapsed between photographs approximately 3s / Fig 25. Superstructure design in decks Incorporating cross-girders/ Fig 26. Possible superstructure designs for decks with no cross-girders/ Fig 27. Possible jointing detail for decks with no cross girders 56|The Structural Engineer – 6 November 2007
SE21 Steel Bridge decks:Layout 1
1/11/07
12:34
Page 57
paper: bright/smith
the electrical running cost for the laser would be approximately £80 for an 8h shift. This cost is negligible considering the vast number of welds which could be completed by a laser operated on a production line basis. Furthermore, the automation of the welding process and the ease of fit-up of the panel would considerably reduce labour costs compared with a conventional orthotropic deck. Laser welded decks could thus offer significant economic as well as structural advantages over conventional orthotropic decks. The sandwich design is also far easier to integrate into the overall bridge superstructure than a conventional orthotropic deck. Various potential construction details are highlighted in the following sections.
Support and jointing arrangements for sandwich panel decks The long-span range of a sandwich panel deck could allow the panels to span between cross-girders, or allow the cross-girders to be omitted all together.
Decks incorporating cross girders Unlike a conventional orthotropic deck, a sandwich panel could be made continuous over the top of its supporting superstructure. Panel joints could comprise downhand welds above cross girders, with the cross girders themselves prefabricated on to the ends of the deck panels. Cross girder spacings would exceed 5m. The design illustrated in Fig 25 would allow both the deck plate and the lower flange of the panel to act integrally as the top flange of the supporting cross girders and main girders without inducing minor axis bending in the panel webs.
Decks with no cross girders On most structures, the long span range of the sandwich panels would allow them to span directly between the main girders of the bridge. These could be positioned outside the zone of trafficking, avoiding the ‘hard-spot’ problems associated with conventional orthotropic decks. Some suggested design configurations are illustrated in Fig 26. Panel joints could be made with prefabricated end plates connecting to the top flanges of the main girders. Deck crossfalls and longitudinal gradients could be accommodated within the end plates whilst maintaining a horizontal plane for the girder flanges as illustrated in Fig 27. The jointing details offered by the sandwich design are potentially far simpler and cheaper than the fully-welded joints used on conventional orthotropic decks. The use of bolted site joints also offers the scope for future removal or replacement of the panels in the event of deterioration or damage. The finished surface of the panels is consistent with that of conventional steel orthotropic decks. The surfacing systems to be applied to the deck could be identical to those developed over recent decades for long-span and moveable steel bridges. However, the tensile stress range experienced by the surfacing above the panel webs will be less than that experienced in a conventional orthotropic deck because the provision of the I-flange within the core of the panel improves the lateral distribution of web reactions into the deck plate (see Fig 9). The improved distribution of torsional stiffness across the width of the sandwich panel also results in a more uniform distribution of web reactions between adjacent I-beams.
Conclusions The design of steel orthotropic decks has traditionally reflected the constraints of cost and fabrication rather than the aspirations of designers. Recent advances in materials-joining technology, such as laser welding, provide an opportunity to overcome these limitations whilst improving fatigue durability. Sandwich panels provide an efficient structural profile with an improved distribution of bending and torsional stiffnesses in both longitudinal and transverse directions compared with conventional orthotropic decks. This gives the deck a more uniform response to localised wheel loads, helping to alleviate fatigue ‘hot-spots’ in the deck plate, surfacing and panel webs. A sandwich profile suitable for bridge deck applications has been devised utilising a core formed from hot-rolled I-sections. This provides an efficient use of materials and exceptional economy in assembly and manufacture. The sandwich design has the potential to provide a competitive, lightweight alternative to the concrete decks used in general bridgeworks. It could also have application to deck replacement schemes where concrete decks of old
bridges are removed to facilitate an increase in the live load capacity of the bridge. The sandwich design could allow the deck to be integrated into the existing superstructure without requiring replacement of the existing crossgirders or modification of the main girders. The system could also have more wider applications, in particular, off-shore structures, shipbuilding and ro-ro platforms. Hollow sandwich panels could also be used as an alternative to heavier composite panels which rely upon a concrete infill to maintain continuity between the outer flange plates. Laser welding allows continuity to be achieved between the flanges without the introduction of further structural infilling. The panels could therefore provide a cost-effective and more structurally efficient system for the construction of blast-protection walls and building cores. Fully cellular structures could also be produced. Dynamic testing undertaken on discrete joints used in the deck indicate considerable fatigue strengths. Full-scale testing of a complete deck assembly is expected once commercial support is forthcoming. Expressions of interest to develop the system further are welcomed.
Acknowledgments This paper is dedicated to the memory of Bill Smith who sadly died earlier this year. Bill was an inspiration to the many students at Bristol University who benefitted from his dedicated teaching of structural engineering over many years. The author is deeply indebted to Bill for his personal guidance and valuable suggestions and discussions throughout the course of this research and for providing the author with the opportunity and the freedom to explore this exciting avenue of research into new technologies and products. Bill will be sorely missed by his numerous friends and colleagues. This research was undertaken at the University of Bristol with the support of a Needham Cooper Scholarship. The kind support of the author’s sponsors, Corus UK, Fairfield Mabey and the Royal Commission for the Exhibition is gratefully acknowledged. The valuable advice offered by Geoff Booth, Peter Lloyd, Bill Ervig and Alex Cole of Fairfield Mabey is particularly noted in the development of the ideas presented in this paper and the subsequent nondestructive investigations of the test samples. The generous assistance of Chris Dolling of Corus UK and his predecessor, Bill Ramsay is gratefully acknowledged for their support of the research and the provision of test samples used in the fatigue tests. The support and guidance offered by British Waterways, The Steel Construction Institute and the Institution of Civil Engineers has also been appreciated. The advice provided by Alan Hayward and Ed Atherton of Cass Hayward and John Powell of British Waterways is specifically acknowledged.
REFERENCES 1.
Bijlaard, F. S. K., Romeijn, A. & Kolstein, M. H.: New Steel Deck Design for Highway Bridges, ECSC Steel RTD Proposal, 2000 2. Mehue, P.: Cracks in Steel Orthotropic Decks. Bridge Management I, pp633-642, Thomas Telford, 1991 3. Wolchuk, R.: ‘Lessons from weld cracks in orthotropic decks on three European bridges’, ASCE J. Struct. Eng., 116/1, pp 75-84, 1990 4. Kolstein, M.H. and de Back, J.: ‘Fatigue behaviour of field welded ribjoints in orthotropic steel decks’, IABSE Workshop, Remaining Fatigue Life of Steel Structures, Lausanne, pp237-248, 1990 5. Eurocode 3: Design of Steel Structures – Part 2: Steel Bridges. European Committee for Standardisation, ENV 1993-2, Oct. 1997 6. Gurney, T.: Fatigue of Steel Bridge Decks. HMSO, UK, 1992 7. AASHTO LRFD Bridge Design Specifications, 1st Ed. American Association of State Highway & Transportation Officials, 1994 8. Wolchuk, R.: Steel orthotropic decks: developments in the 1990s. Transportn Research Record 1688, 30-37,1999 9. Russell, J. D.: ‘The development of CO2 laser technology for welding of structural steel’, TWI, Cambridge. Make it With Lasers Conference, Corus, Scunthorpe, 21-22 November 2000 10. Bright, S. R.: ‘Evaluation and testing of a new form of lightweight deck produced by laser welding’, PhD Thesis, University of Bristol, April 2004
6 November 2007 – The Structural Engineer|57
