Displacement-Based Seismic Design of Structures

Abstract

The concept of designing structures to achieve a specified performance limit state defined by strain or drift limits was first introduced, in New Zealand, in 1993. Over the following years, and in particular the past five years, an intense coordinated research effort has been underway in Europe and the USA to develop the concept to the stage where it is a viable and logical alternative to current force-based code approaches. Different structural systems including frames, cantilever and coupled walls, dual systems, bridges, wharves, timber structures and seismically isolated structures have been considered in a series of coordinated research programs. Aspects relating to characterization of seismic input for displacement-based design, and to structural representation for design verification using time-history analysis have also received special attention. This paper summarizes the general design approach, the background research, and some of the more controversial issues identified in a book, currently in press, summarizing the design procedure. INTRODUCTION Viewed through the historical prism of the past 100 years, seismic structural design can be seen to have been in constant evolution – much more so than design for other load cases or actions such as gravity, wind, traffic etc. Initially, following structural damage in the seminal earthquakes of the early 20 century (Kanto, Long Beach, Napier), seismic attack was perceived in terms of simple mass-proportional lateral forces, resisted by elastic structural action. In the 1940’s and 50’s the influence of structural period in modifying the intensity of the inertia forces started to be incorporated into structural design, but structural analysis was still based on elastic structural response. Ductility considerations were introduced in the 1960’s and 70’s as a consequence of the experimental and empirical evidence that welldetailed structures could survive levels of ground shaking capable of inducing inertia forces many times larger than those predicted by elastic analysis. Predicted performance came to be assessed by ultimate strength considerations, using force levels reduced from the elastic values by somewhat arbitrary force-reduction factors, that differed markedly between the design codes of different seismically-active countries. Gradually this lead to a further realization, in the 1980’s and 90’s that strength was important, but only in that it helped to reduce displacements or strains, which can be directly related to damage potential, and that the proper definition of structural vulnerability should hence be related to deformations, not strength. This realization has lead to the development of a large number of alternative seismic design philosophies based more on deformation capacity than strength. These are generally termed “performance-based” design philosophies. The scope of these can vary from comparatively narrow structural design approaches, intended to produce safe structures with uniform risk of damage under specified seismicity levels, to more ambitious approaches that seek to also combine financial data associated with loss-of-usage, repair, and a client-based approach (rather than a code-specified approach) to acceptable risk. This paper does not attempt to provide such ambitious guidance as implied by the latter approach. In fact, it is our view that such a broad-based probability approach is more appropriate to assessment of designed structures than to the design of new structures. The