Effective Slab Width for Composite Steel Bridge Members (2005)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

66 CHAPTER 4 CONCLUSIONS AND SUGGESTED RESEARCH 4.1 CONCLUSIONS This study has resulted in the recommendation that full width may be used for effective width in composite steel bridge members for most situations of practical interest. This recommendation was determined to be suitable for the Ser- vice as well as Strength limit states, for exterior as well as interior girders, and for skewed as well as right alignments. The simplicity of this recommendation results from an extensive set of analyses on various bridge configurations cul- minating in the following two consistently observed trends: • Full width was typically acting at cross sections where it was most needed, i.e., where moments and hence per- formance ratios would be highest; in the cases where the effective width was less than full width at such cross sections, that cross section had considerable excess flex- ural capacity, and • An extensive “impact analysis” based on Process 12- 50 principles revealed that more cumbersome curvefit expressions for effective width, although more accurate, were not significantly so in terms of the governing rat- ing factor (RF) of the bridge investigated. Based on a limited number of studies of prestressed con- crete girder configurations producing similar results, the above simple criterion is thought to be reasonable for such config- urations as well. The very notion of effective width presumes composite behavior. A question addressed in one of the experiments performed herein is whether composite behavior can legiti- mately be assumed in negative moment regions without con- tinuous shear connectors. It would appear that composite behavior can be attained in negative moment regions, even without shear studs being distributed throughout the negative moment region, as long as the longitudinal reinforcing steel is properly anchored and developed. This observation, how- ever, is based on only a single experimental specimen. 4.2 IMPLEMENTATION PLAN It is recommended that the AASHTO Subcommittee on Bridges and Structures (SCOBS) consider the draft revisions to Article 4.6.2.6.1 of the LRFD Specifications and Com- mentary as developed herein. These provisions in LRFD two- column format are provided herein on pp. M-7 and M-8 of Appendix M. 4.3 SUGGESTED RESEARCH 4.3.1 Bridge Types and Geometries Not Considered Herein 1. Tied-arch bridges were not explicitly modeled in the FEM studies performed herein. It is not known how the presence of net tension in the floor system of such bridges will affect the effective width. Various deck- ing options should be considered in this context (e.g., cast-in-place and precast prestressed longitudinally post-tensioned). 2. Curved bridges present another situation of interest. The forthcoming 2005 Interims to the AASHTO LRFD Specifications add curved girder analysis provisions to the curved girder resistance provisions that were in the 3rd Edition of the Specifications in 2004. It is generally agreed that a curved girder bridge should be analyzed as a system, such that line-girder simplifications (where effective width is used) would not apply. Approximate methods, however, are explicitly permitted in the 2005 Interims for use in analyzing curved girder bridges. One such approximate analysis method is the V-Load method. The V-Load method idealizes the curved girder as a straight girder subjected to vertical (V) loads applied at diaphragm locations to complement gravity loads. Engineers using the V-Load method will want to know what value of effective width to use for resisting super- imposed dead-load and live-load effects on the com- posite section. The research reported herein simply does not address that question. 3. A third situation not explicitly investigated herein is where decks are longitudinally prestressed. Such decks would be designed not to crack under service loads. Whether the effective width of such decks remaining uncracked at the critical cross section in negative moment regions extends to the full width is an impor- tant question. This question was beyond the scope of

67 the parametric study conducted herein and should be investigated. Although it is worth mentioning that while investigating the negative moment region prior to crack- ing, the observation was that an apparent smaller beff than that of a cracked section whereby the rebars are fully engaged in taking the tensile stresses. 4. A fourth situation involves bridge decks that do not have solid thicknesses that meet or exceed the mini- mum depth of 175 mm (7 in.) specified in Article 9.7.1.1 of the AASHTO LRFD Specifications. The FEM analy- ses upon which this report’s recommendations are based all presumed solid deck thicknesses meeting or exceed- ing this minimum depth requirement. Situations with other types of decks are thus not included within the scope of the recommendations and therefore should be investigated independently. 5. The research results presented herein focused primarily on slab-on-girder systems. Whether the effective width provisions for slab-on-girder systems can or should be reconciled with the effective width provisions for seg- mental prestressed concrete box girders is a reasonable question. Posited another way, for example, why should a deck in a segmental box girder experience shear lag differently than a deck in a tub girder? One would not expect a difference, which means that one would not expect a different criterion for effective width. Yet there was no attempt in the present study to reconcile its results with existing provisions for segmental box girders. 4.3.2 Bridge Types and Geometries Considered Herein Within the framework of the parametric study and addi- tional cases examined herein, further investigation may be appropriate beyond the range limits adopted in this study, i.e., • Girder spacings farther apart than 4.8 m (16 ft), or in fact greater than 3.6 m (12 ft) for prestressed girders, • Span lengths greater than 60 m (200 ft), and • Skew angles greater than 60 deg. Although prestressed concrete girder configurations and cable-stayed bridges were considered in this study, only a few such cases were explicitly examined. Expanding the num- ber of analyses on these types of bridges in order to modify or increase the credibility of the recommendations contained herein for these types of bridges may be desirable. For cable-stayed bridges, the following are suggested as areas for further research regarding effective width: • Live-load placement influence on effective width: No live load was considered in the present study because the dead load is very large in such structures with respect to live load and because Byers (whose results served as a basis for comparison with ours) did not use live load either. Thus, it would be of interest to investigate the live-load influence on the effective width. • Influence of cable tensioning during construction on beff: Construction steps were not considered in this work. Investigating their influence on beff is recommended. • Negative shear lag in cable-stayed bridges: The phe- nomenon has been observed in various types of struc- tures, and further research on the subject is recom- mended as far as cable-stayed bridges are concerned. • Single versus dual effective width: A single value of effective width was evaluated herein. Attempts to sepa- rate normal stresses into their “axial” and their “flexural” components are also recommended for future research although the difficulty of the task is recognized. • Impact assessment in terms of rating factor: As was done for the more common slab-on-girder cases, it would be of interest to investigate how the proposed values of effective width in cable-stayed bridges affect analysis results as measured by rating factor. • A wider range of bridge geometries and cable config- uration: For example, in this project the bridges investi- gated (other than the Cooper River Bridge) were no more than 30 m (100 ft) wide. It would be of interest to deter- mine values of effective width for additional bridges wider than 30 m and for bridges with different cable patterns, cable spacing, floorbeam spacing, slab thick- ness, and so forth. 4.3.3 Recommendations Originating from Experimental Investigations From the experimental studies conducted as part of this research, the following recommendations for further research arise: • Research is recommended for evaluation of instrumenta- tion used for measuring strain on rebars that are embed- ded in concrete. • More extensive study, including evaluation of rebar strains, of intentionally composite versus noncomposite slab-on-girder specimens would be valuable. Ideally the specimens would be multi-girder systems. Investiga- tions and comparisons of global and local composite/ noncomposite behaviors would be useful. As mentioned in the literature review, AASHTO is confusing on the point of composite behavior relating to shear stud design. It is recommended that research in this area includes eval- uation of situations and/or loading that allow noncom- posite beams to be evaluated as composite. Surely the details of the steel-concrete bond surface would be of interest since various conditions exist at that location in the field. Perhaps a FEM model with interface elements

68 • Further investigation of crack patterns and how they relate to composite beams versus slip regions of non- composite beams with developed rebar may be useful in developing and verifying refinements to concrete and rebar material modeling assumptions and friction modeling assumptions used in finite element models of slab-on-girder bridges. only in the shear stud cluster region would be in order since there was no contact in certain regions of the non- composite specimen during later levels of loading in the experiment. This report presents only comparison of one composite specimen to one noncomposite specimen, so it is unreasonable to assume that the material presented here applies to all conditions.