Reliability-Based Simulation for Performance-Based Design

Johns Hopkins University, Baltimore MD

Investigators

Abstract

During the past decade, significant changes have occurred in building design and construction: the move toward limit states design; revolution in the use of advanced computation as a tool in structural design; revolution in the use of advanced computation as a tool in structural design: and the introduction of innovative high-performance materials. At the same, time, the performance of many structures designed by traditional methods and criteria and subjected to extreme natural hazards has been judged increasingly inadequate. The problem is that in current design codes the prevention of structural collapse has been used as the primary criterion of acceptable performance. It has been found, however, that less drastic forms of structural damage as well ass the functional disruptions resulting from natural hazards can lead to economic losses that approach the cost of the structure itself. Current design practice focuses on the characteristics and behavior of a new facility. However, the continued satisfaction of the building owner and occupants depends on the performance of the building over an extended period. In the past, durability of materials usually has been addressed through material selection, detailing and finishing at the design and construction stage. This approach has not always led to satisfactory results in service, and is likely to even less successful as innovative materials are introduced into building construction. To achieve better solutions in performance-based engineering, material selection and durability issues must be better integrated into the design process. Methods to do this are only in a rudimentary stage of development. For performance-based engineering to achieve its full potential, the following ingredients must be available: computational tools to evaluate the performance of integrated systems; relations that define performance in terms of structural response quantities that engineers can compute using available methods; databases to measure uncertainties in the engineering parameters of interest; system reliability analysis procedures to provide the framework for managing these uncertainties and quantitative risk targets.