Comparison of methods to develop risk-targeted seismic design maps

Abstract

The seismic design of structures according to current codes is generally carried out using a uniform-hazard spectrum for a fixed return period, and by employing a deterministic approach that disregards many uncertainties, such as the contribution of earthquake ground motions with return periods other than that assumed for the design. This results in uncontrolled values of the failure probability, which vary with the structure and the location. Risk targeting has recently emerged as a tool for overcoming these limitations, allowing achievement of consistent performance levels for structures with different properties through the definition of uniform-risk design maps. Different countries are implementing the concepts of risk targeting in different ways, and new methods have recently emerged. In the first part of this article, the most well-known approaches for risk targeting are reviewed, with particular focus on the one implemented in recent American design codes, the one based on the use of risk-targeted behaviour factors (RTBF), and an approach based on direct estimation of hazard curves for inelastic response of single-degree-of-freedom systems. The effect of the linearization of the hazard curve is investigated first. A validation of the RTBF approach is then provided, based on comparison with the results of uniform-risk design spectral accelerations for single-degree-of-freedom systems with elastic-perfectly plastic behaviour for two different sites. The effectiveness of the current risk-targeting framework applied in the United States is also investigated. In the last part of the paper, uniform-risk design maps for Europe are developed using the RTBF approach, showing how the seismic design levels may change when moving from a uniform-hazard to a uniform-risk concept.