Conclusions and Recommendations:

The workshop focused on issues important to the development of performance-based seismic design methodologies that can form the basis for practical guidelines, standards, and code implementation. The emphasis was on general concepts rather than issues specific to design and construction practices and code approaches in various countries. The workshop participants recognize that issues specific to different materials, innovative structural systems, existing versus new construction, and regions of different seismicity have not been addressed.

The following recommendations and conclusions have been developed by four working groups, and are based on extensive discussions of the presentations given during the first two days of the workshop.

Working Group on Loss Estimation, Fragilities and Vulnerability, and Impact on Risk Management

Co-Chairs: Comartin & Whittaker
Recorder: Miranda
Working Group Members: Bachman, Cornell, Elnashai, Hutchinson, Lowes, Manfredi, Kawashima, Porter, Reinhorn, Nakashima, Pinho, Pinto P., Sucuoglu, Franchin, Iervolino, Kante, Kramar

Conclusions and Recommendations

1. There is an urgent need to install dense arrays of instruments in selected buildings, bridges, and other structures to collect performance (loss) data. The structures and sites should be selected so that the likelihood of recording a comprehensive set of data for important types of structures within a short time frame is maximized, i.e., the emphasis should be on instrumentation of structures in urban areas of high seismicity. This will necessitate the development of plans and protocols for damage and loss (performance) data collection, and the documentation of comprehensive information on properties of the structures prior to the occurrence of an earthquake.
2. It is recommended to develop an expert system, essentially a Virtual Contractor, to aid in aggregation of capital losses for different damage scenarios, i.e., to enable the calculation of capital losses for specified distributions of engineering demand parameters. The development of the expert system knowledge base will require the systematic collection and synthesis of both loss data from past earthquakes and information from expert contractors. The expert system shell should be common to all countries and regions but the knowledge base will vary by region and country.
3. An international web-based repository should be developed and maintained for performance data and information of interest in the context of performance-based earthquake engineering, including a stakeholder encyclopedia (describing and defining performance in a manner meaningful to the stakeholders) and fragility data for structural, nonstructural, and content components and systems. As part of this effort, protocols should be developed for testing and documentation of experimental results.

Working Group on Implementation of PBEE in Engineering Practice

Co-Chairs: Hamburger & Kabeyasawa
Recorder: Bommer
Working Group Members: Alcocer, Aschheim, Aydinoglu, Bonelli, Booth, Chopra, Cosenza, Deierlein, Fajfar, Fardis, Fischinger, Ghobarah, Kowalsky, Krawinkler, Kunnath, Negro, Otani, Pinto A., Rutenberg, Shiohara, Sritharan, Tsai, Tso, Dolšek, Marušić, Peruš, Poljanšek, Sigmund, Zevnik

Conclusions and Recommendations

1. Performance-based seismic design can be viewed as a process of system conception followed by an assessment procedure in which the performance of the structural system is evaluated and improved as needed to satisfy stated performance objectives. Design tools should be developed, particularly for new structures, to assist in the conception of an effective structural system in order to provide a good starting point for subsequent assessment. Direct design, without subsequent assessment, is a feasible option for simple structures.
2. The foundation of PBSD procedures should be reliability-based. In implementation in engineering practice the reliability concepts may be incorporated implicitly through appropriate demand and capacity factors, while explicit incorporation of reliability concepts is an option to be considered primarily for major facilities with special performance requirements. Reliability concepts should also be considered to improve and transition existing code-based design methods and to improve the calibration of prescriptive rules in existing codes. They may also be partially introduced into codes to aid the transition to full use of PBSD.
3. PBSD concepts should be incorporated into codes on a world-wide basis, but with due consideration to the need for simplicity and for sound engineering judgment, and with due consideration of economic and societal priorities.
4. Opportunities should be sought out to demonstrate the feasibility and advantages of PBSD approaches compared to presently employed prescriptive approaches.
5. Future guidelines and codes must be clear regarding the limitations in the use of the different analytical procedures (linear static, nonlinear static, linear dynamic, nonlinear dynamic) and when they should not be used. Emphasis in research should be on the development of nonlinear analysis procedures. For nonlinear static (pushover) analysis additional research is needed on extension to irregular structures (particularly unsymmetrical) and structures with significant higher mode effects. Nonlinear dynamic procedures need improvement of large displacement predictions and of element hysteretic models.
6. More work needs to be done in defining appropriate performance measures, with an emphasis on providing protection against life safety hazards and excessive economic losses.
   

Working Group on Performance-Based Design Concepts

Co-Chairs: Deierlein & Fardis
Recorder: Aschheim
Working Group Members: Aydinoglu, Bachman, Bommer, Bonelli, Booth, Chopra, Comartin, Cornell, Fajfar, Fischinger, Ghobarah, Hamburger, Hutchinson, Kabeyasawa, Kawashima, Kowalsky, Krawinkler, Manfredi, Miranda, Otani, Pinho, Pinto P., Porter, Rutenberg, Shiohara, Sucuoglu, Tso, Whittaker, Dolšek, Franchin, Iervolino, Poljanšek, Sigmund, Zevnik

Conclusions and Recommendations

1. The goal of performance based seismic design (PBSD) is to assist in the engineering of cost-effective facilities, whose safety and resistance to damage from earthquakes meet the needs and expectations of key stakeholders and society at large more effectively and reliably than can be achieved with codes using prescriptive design rules. Key incentives for the use of PBSD include:
   a) Reduction in the initial capital costs of facilities designed to have comparable performance to that implied by existing standards based on prescriptive rules.
   b) Ability to design structures to higher performance that have improved safety and lower life-cycle costs associated with seismic risk.
2. The most immediate need for and benefit from PBSD are for existing structures and new facilities with special features that are not adequately addressed by existing codes (e.g., innovative new structural systems, bridges on liquefiable soils, industrial plants with complex geometries).
3. Efforts should be continued to demonstrate the benefits that PBSD will provide to key stakeholders and, thereby, to the engineering professionals who embrace PBSD in design practice. This should include pilot applications to both special facilities that cannot be reliably designed using current codes and standards, and to more conventional facilities.
4. More attention should be given to bridges, industrial facilities and other important infrastructure facilities and systems.
5. Research efforts towards improving capabilities for prediction of collapse should be emphasized. Structural collapse is defined as the state in which a structural component (for local collapse) or the structural system (for global collapse) is no longer capable to resist its tributary gravity load. Criteria for local collapse need to be established and elaborated. The extent to which local collapse propagates and conceivably leads to system collapse depends on the configuration and redundancy of the system and its ability to redistribute gravity loads from the failed component(s) to the neighboring ones. Much more experimental data on component deterioration and system collapse is needed in order to calibrate analytical models being developed for collapse prediction.
6. A transition from presently employed prescriptive design requirements to performance-based design requirements should be gradual, in order to calibrate the consequences of performance-based design and provide safeguards against its misuse. Overriding issues are societal concerns with losses of lives and excessive financial losses that may have a regional impact.Research efforts towards improving analysis capabilities for structure-soil-foundation systems should be intensified, and collaboration between structural and geotechnical engineers should be emphasized.
   

Working Group on Harmonization of Experimental and Analytical Simulations

Co-Chairs: Elnashai & Nakashima
Recorders: A. Pinto & Ghobarah
Working Group Members: Alcocer, Cosenza, Kunnath, Lowes, Negro, Otani, Reinhorn, Sritharan, Tsai, Žarnić, Kante, Kramar, Marušić, Peruš

Advanced experimental facilities have become available worldwide; for example, NEES, E-Defense, JRC, NCREE, etc. Experiments on complex structural systems at larger scales become more practicable; they provide great opportunities for more accurate characterization of various limit states of structures and ultimately for accelerated acceptance of PBEE. New experimental facilities, techniques, and devices require new approaches to research and development. The following specific recommendations are along these lines.

Conclusions and Recommendations

1. Testing procedures. Experimentation should cover the full range of behavior from damage initiation to collapse. Test structures should contain nonstructural and content systems to the extent feasible. In simplified test configurations, much attention needs to be paid to simulation of boundary conditions. Field testing should be encouraged to provide realistic performance data. A great need exists to develop testing protocols, including interaction between testing and analysis, peer review of procedures, careful selection of input motion, and specialized protocols for testing of nonstructural components and for material testing. Advanced instrumentation should be developed (including high resolution image processing) for comprehensive documentation of damage data. All experimental data should be documented, archived and shared publicly after verification, taking into account intellectual rights.
2. Analytical prediction of behavior till collapse. Improved approaches need to be developed to simulate collapse and behavior of non-structural systems, and for constitutive modeling of new and existing materials. Computer analysis programs should emphasize user-friendliness and should be developed through partnerships of researchers and practicing engineers with software companies.
3. Distributed simulations. The benefits obtained from geographically distributed simulation should be clearly advocated, including the identification of systems that necessitate distributed simulation and cannot be dealt with otherwise. To raise public awareness, news media should be utilized to inform the general public, including the technical community and policy makers, of major distributed simulation efforts and to encourage tele-observation of experimental activities.