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.
47 CHAPTER 4 CONCLUSIONS AND SUGGESTED RESEARCH 4.1 CONCLUSIONS The conclusions presented here go beyond the scope of Project NCHRP 12-61. These conclusions principally iden- tify deficiencies that the research team considers will remain, even if the change proposals recommended in this document are adopted. The conclusions are presented in four cate- gories: basis of design provisions; role of experimental research and field experience; role of design database and numerical tools; and differences in shear design provisions. 4.1.1 Basis of Design Provisions 1. Although researchers may agree about the components that contribute to shear resistance, there is considerable disagreement about the relative magnitude of these con- tributions, the factors that influence these contributions, and their significance for different design conditions. 2. The diagonal cracking strength is not a measure of the concrete contribution at ultimate for members with shear reinforcement. Thus, provisions in which Vc is related to the diagonal cracking strength for members with shear reinforcement are purely empirical and need to be validated by comprehensive test data. 3. The parallel chord truss model provides a direct means of calculating the contribution of shear reinforcement to shear capacity. That contribution can be calculated as the yield strength of a stirrup times the number of stirrups in one leg of the idealized truss (see Equation 35 where the angle θ is the angle of diagonal compres- sion in the idealized truss relative to its longitudinal axis); however, there is no agreement on how to calcu- late the angle θ. In ACI 318-05 and the AASHTO Stan- dard Specifications, θ is assumed to be 45 degrees. In the Eurocode, in which the shear design provisions are partially based on plasticity theory, the angle can be selected by the designer to be as low as 18 degrees or that when the diagonal compressive stress reaches a limit equal to about 60 percent of the concrete com- pressive strength. In the LRFD Sectional Design Model, this angle is calculated using the MCFT and using the longitudinal strain at mid-depth. In the pro- posed simplified provisions, this angle is calculated in the web-shear region using Mohrâs circle of stress to find the angle of diagonal cracking. It has further been argued by researchers that the number of stirrups that should be considered to form a given leg in the ideal- ized truss should be dvcot(θ)/s â 1 because diagonal cracks often form from the top of one stirrup to the bot- tom of another. (Eq. 35) 4. For members with shear reinforcement, the equation developed for Vc in provisions must account for the rules used for evaluating the angle of diagonal com- pression. For example, in current U.S. practice where the angle θ is assumed to be 45 degrees for prestressed concrete structures, the calculated contribution of the shear reinforcement is less and Vc can afford to be larger than in other codes such as the LRFD Sectional Design Model where θ may be as low as 18 degrees. 5. For members without shear reinforcement, there is a large debate about how to evaluate the contribution of the concrete at the ultimate limit state. Some re- searchers argue that it should be the load required to form or propagate a diagonal crack. Others suggest that it should be based on the shear-slip resistance of the diagonal crack while others suggest that it is best eval- uated by considering the shear force that can be trans- mitted in the uncracked compression zone. Regardless of which method is used, there is a significant depth effect in shear for members without transverse rein- forcement and little depth effect for members with shear reinforcement; members without shear reinforce- ment and with a unit depth of three can fail at one half the stress of a geometrically similar member with a unit depth of one. However, there is significant debate over the types and sizes of members for which the depth effect in shear must be considered. 4.1.2 Role of Experimental Research and Field Experience 1. What researchers have tested and continue to test in laboratories is not representative of what is built using V A f d s s v y v = cot( )θ
48 design codes. The most typical laboratory test structures are small (less than 15 inches deep), have rectangular cross sections, do not contain shear rein- forcement, are simply supported, are stocky, are loaded by point loads over short shear spans, and are supported on bearings positioned underneath the member. In addition, nearly all members are designed so that shear failures occur near supports. By contrast, a large frac- tion of the bridge members in the field are large, con- tinuous, have top flanges, are subjected to uniformly distributed loads and are built integrally at their ends into diaphragms or piers. In addition, members in the field are designed for shear over their entire length and away from simple supports where there can be sub- stantial effects of flexure on shear capacity. 2. Because most code provisions are ultimately validated by test data, and because most members in the experi- mental database do not represent what is built with pro- visions, there is great uncertainty about the safety, economy, and validity of these provisions for most shear design regions in most structures. A particular case in point is the region of contraflexure in a contin- uous beam. The wide spread in the shear requirements found in Example 5 of Appendix J for this region for different provisions is a direct reflection of the uncer- tainty of the safety of those provisions for that region. 3. Most experimental researchers fail to collect or report detailed information about the performance of the test members before failure. This information consists of material properties, member or test set-up geometry, crack patterns and widths, stirrups strains, measured diagonal compressive stresses, and shear deformations. Thus, most tests are not useful for assessing the condi- tion of members under service loads or for evaluating the accuracy of complete behavioral models for resistance. 4. It is difficult to judge the overall safety of design code provisions from field experience because most struc- tures in the field have redundant load paths, additional load resisting elements not accounted for in design, and are unlikely to be subjected to loads approaching their factored design loads. Further, many of the difficulties observed in the field are dominated by an interaction of deterioration, environmental and repeated loading effects. 4.1.3 Role of Design Database and Numerical Tools 1. In this project, a comparison was made of the required strength of the shear reinforcement (Ïvfy) by four dif- ferent design approaches and by Response 2000 for about 500 design cases. The research team chose these design cases in an effort to capture the range in design cases for which shear design specifications would be applied. The results of these comparisons were considered useful for evaluating the safety and econ- omy of design provisions, particularly for the types of structures and regions for which there is little experi- mental test data. 2. Although these comparisons were useful, the dataset selected by the research team may not well represent the types and frequency of structures to be designed by provisions. 3. The assessment of the effect of the proposed changes on bridge design practice would also have been more reliable if the design database well represented the types and frequencies of structures to be designed by these provisions. 4. The results would also have been more useful if there were more computational tools (other than Response 2000) for predicting the required strength of shear rein- forcement in these design cases. 5. The NCHRP Process 12-50 helped establish a frame- work for addressing the three foregoing shortcomings, but the design example database has yet to be populated with representative types and frequencies of members designed with provisions. 4.1.4 Differences in Shear Design Provisions 1. There is a wide variation in the forms of shear design specifications used in different influential codes of practice such that the amount of shear reinforcement required by one code may be two to three times that required by another code for the same section and fac- tored sectional forces. 2. There remains considerable disagreement in codes of practice on the minimum required amount of shear reinforcement and when this minimum reinforcement is required. There is a factor of about 2 in the minimum required amounts of shear reinforcement. Some codes required minimum shear reinforcement when the fac- tored design shear force exceeds one half of the design strength provided by concrete alone while others do not require minimum shear reinforcement until the fac- tored design shear force exceeds this design strength. The types of members exempt from more stringent minimum shear reinforcement requirements also vary. 3. There is a large variation in the maximum allowable shear stress by different codes of practice. The differ- ence can be a factor of two and one-half between the AASHTO Standard Specifications and LRFD Sec- tional Design Model. 4. The depth effect in shear that has been strongly observed in members without shear reinforcement is captured in some codes of practice by making the allowable design stress a function of the overall depth of the member. The depth effect can change the allow- able shear design stress by more than a factor of two for different sized members.
5. The bases of shear design provisions include experi- mental test data, the equilibrium condition of members in the ultimate limit state, and comprehensive behav- ioral models for capacity. 4.2 RECOMMENDED RESEARCH Several significant shortcomings in shear design practice were presented above. Of particular concern are the large dif- ferences between codes of practice in the required amount of shear reinforcement, the maximum allowable shear design stress, minimum shear reinforcement requirements and how the depth effect in shear is addressed. Equally important are the lack of experimental validation for practical design cases and the lack of understanding and consensus on how struc- tural concrete members carry shear. To address these con- cerns, the following research efforts are recommended. 1. Process 12-50 should be populated with the range and frequency of members commonly designed using the AASHTO Bridge Design Specifications. This will enable a much more accurate assessment of the effect of proposed changes to the specifications on the safety and economy of the nationâs current and future inven- tory of bridge structures. 2. A web-based national database of shear test results should be established. This can be used by researchers and funding agencies to understand where research is most needed. It will also ensure that the resources spent in conducting experiments are used when it is 49 time to revise code specifications. Process 12-50 devel- oped some of the structure for creating this archive, but additional coordination and efforts are required. An example of an experimental test archive being devel- oped by the earthquake engineering community as part of a large new initiative by the National Sciences Foun- dation is available at http://nees.org 3. Because provisions principally are validated by test data, shear tests are needed on the types of members built with provisions but for which there is little or no test data. This missing population principally consists of large members, continuous members, members sup- porting distributed loads, and members that fail in regions other than adjacent to a support. 4. Where testing of members is not practical, suitable numerical approaches should be used to obtain the best possible estimates of shear capacity and behavior. 5. Standards for shear testing should be developed so as to ensure that material test data and the detailed struc- tural Response 2000 are measured in such ways that they can enable the evaluation of structures under service load levels, as well as ultimate load levels, and the validation of numerical methods and behavioral models for analyzing Response 2000. 6. Although the depth effect in shear has been well demonstrated, the range of applicability of this effect and its relation to minimum shear reinforcement requirements needs to be better understood. It is likely that depth rather than concrete compressive strength is a better parameter for establishing minimum shear rein- forcement requirements.
