Performance-Based Seismic Bridge Design (2013)

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.

84 CHAPTER TWELVE IDENTIFICATION OF KNOWLEDGE GAPS There does not appear to be a single entity that is driving the industry toward PBSD. Instead, there is a perception that the bridge industry could better predict likely performance in large, damaging earthquakes than it is doing at the present. This perception is driven from within the engineering and research communities as the state of knowledge advances. However, there are gaps in that knowledge base that need to be closed. There is also a perceived need from outside the engi- neering community, because the ability to clearly describe to the public what performance to expect in earthquakes is less developed than public decision makers would prefer. Knowledge gaps certainly exist in all facets of PBSD; however, key knowledge gaps that need to be closed in order to implement PBSD are covered here. Gaps related to seis- mic hazard prediction include the following: • Currently, probabilistic ground motion data are related on a uniform-hazard basis. This basis also includes mean response. For PBSD to work, the entire chain of seismic design calculation must be formulated in terms of a central tendency—median or mean—and a measure of dispersion about that central tendency— coefficient of variation, standard deviation, and so on. Probabilistic density functions or distributions of failure rates must be quantified. This has been done already with the ground motion data; thus, there are fewer knowledge gaps with seismic hazard than from the other portions of the PBSD process. The science and art of seismic hazard characterization continue to evolve, and it is reasonable to expect that they will con- tinue to do so. • The potential impact of developing risk-adjusted spec- tral accelerations for the 1,000-year design earthquake is not known. The use of risk-adjusted accelerations could reduce the design accelerations in some parts of the coun- try, based on the experience of the building industry. Gaps related to structural analysis include the following: • Improved nonlinear static analysis procedures that pro- vide median response. Additionally, means for estab- lishing the dispersion of structural response around the mean or median (β factors) will be required for implementation of PBSD. Guidance is needed for each analysis technique or methodology, including the coef- ficient and the capacity spectrum methods, to ensure consistent application of probabilistic data. • Modeling guidelines for nonlinear analysis specific to the analytical technique used and the objectives selected, in terms of detail and refinement. Such guide- lines are to focus on areas where simplification can be employed without loss of accuracy and on the uncer- tainty that is introduced with simplification. Gaps related to damage prediction include the following: • A national consensus document relating damage lev- els and performance descriptors. Currently, there is inconsistency in the use of terms and the use of dif- ferent terms. A more uniform application of PBSD could eventually be implemented if researchers report consistent data; state DOTs use consistent terms for performance, damage, and repair; and the design com- munity uses consistent terms. • Lack of sufficient fragility relationships and the asso- ciated uncertainty regarding damage levels (e.g., strain, curvature, rotation, displacement). Some fra- gility information has been developed, and these rela- tionships may provide approximate data until more detailed information becomes available. • Lack of fragility relationships for different types of structural systems, particularly substructure systems where energy dissipation is expected. In some areas, sufficient data have already been developed, and they simply need to be put into a usable form. Some of the PEER databases already have addressed this for some elements, for instance circular and rectangular col- umns, although relevant data may always be added as they become available. For other substructure types, insufficient data exist to formulate reliable fragilities. • The influence of loading history. The sequence of load- ing cycles may affect the available deformation capac- ity (and ductility capacity) owing to concentration of strains in reinforced concrete members. • The quantification of seismic performance of innova- tive or novel designs. Useful items currently missing include standardized formats for proof testing, specifi- cations, and design procedures or methodology. • Clearly written descriptions of damage types. These could be limited to EDPs so that designers may clearly understand the damage limit states that are

85 being considered. In general, it would be useful if design codes could provide detailed commentar- ies describing the physical damage states that each prescriptive requirement or performance method- ology is intended to control. Currently, the further the design specification migrates from the origina- tor of the knowledge—researchers, field investiga- tors, reconnaissance teams—the more likely it is that knowledge of the failure state is lost. This must be corrected for PBSD to be successful. If the design engineer cannot identify and confidently suppress potential failure modes, the design is likely not to be successful. Performance-related gaps include the following: • Performance information related to permanent defor- mations of a bridge. This includes deformations of col- umns and other substructure units and deformations, offsets, slopes and misalignment of the roadway, barri- ers, approach slabs, and other features that are impor- tant to the driving public, emergency responders, and continued use of bridges. • Performance related to clearly defined damage states that are unambiguous regardless of location within the country or type of structure used. • Expected performance and damage states in smaller earthquakes. Sufficient data may not be available today to describe the expected damage in smaller earth- quakes; so the only alternative is to analyze the struc- ture for the smaller events. It is expected that this will continue to be the case. However, development of clear definitions of the performance that would be preferred in smaller earthquakes would be useful. For instance, is it acceptable to be above first yield, but below full yield of the section, or below a small displacement duc- tility factor, say 1.5? Gaps related to loss prediction are the following: • Cost data are inconsistent and generally lacking for performance-based design, although many agencies are willing to pay somewhat more to achieve improved seismic performance. Some cost data exist, but are perhaps not generally available. For example, Caltrans potentially has relevant repair and replacement cost data from the Loma Prieta and Northridge earth- quakes, which could be used to support this effort. • There is a need for objectively developed data, which are calibrated against cost models, for quantifying the risks of losing a structure by collapse or of losing some or all levels of service for the structure. • Methodologies for involving the public or other share- holders are either not available or are not sufficiently developed. Again limited data exist, but may not be generally available. Gaps related to regulatory oversight and training are the following: • Current PBSD guidelines, for example the FHWA Retrofitting Manual and the ICC Performance Code, require user- or project-specific decisions to link performance with EDP) and DMs, and such guide- lines may benefit from additional prescriptive provi- sions or additional guidance to augment the basic performance data. • Many engineers are not prepared, either by practice or by their educational training, to deal with inelastic response as a result of earthquakes. This gap is gradu- ally being closed as younger engineers are trained in inelastic behavior, primarily in graduate school, and as younger faculty, who are similarly trained, enter the workforce. Their students then enhance public agen- cies’ and private practices’ skill sets in this area. Gaps related to decision makers are the following: • Development of clear and agreed-upon definitions of critical, essential, or any other designation of impor- tant bridges. Currently, little is stated in the AASHTO LRFD, and the definitions that are included may need clarification. Decision makers need a national consen- sus approach or guideline for defining the importance of bridges and then for defining the appropriate seismic design criteria for such bridges. • The integration or lack thereof of seismic effects with other hazards is not considered in the AASTHO docu- ments. This falls under the category of multihazard performance. Based on the observations of the direc- tion in which PBSD is likely to move, where calcula- tion of the risk of loss of a bridge or loss of use of a bridge is a likely goal, there may be an opportunity to integrate seismic effects with other hazards. This might be accomplished on the basis of uniform risk (i.e., some percentage risk of loss) for combined haz- ards, rather than on the basis of the occurrence of the hazard itself (i.e., chance of an earthquake or flood occurring). Analytical procedures for this combined risk approach would need to be developed. • Guidance regarding the development of consistent criteria for nonseismic bridge improvement projects needs to be developed. An example is widening of bridges where the existing bridge may be older and have seismic deficiencies, and the widening may be used to improve the existing bridge. However, fund- ing requirements often restrict the scope of improve- ments for such cases. Therefore, a logical framework for deciding when, if, and how seismic issues will be addressed on such nonseismic projects needs to be developed to assist decision makers. Such decisions could be made on the basis of bridge performance, thus drawing on principles of PBSD.