Optimal Intensity Measures for Probabilistic Seismic Response Analysis of Bridges on Liquefiable and Non-Liquefiable Soils

Abstract

One of the key steps of the Performance-Based Earthquake Engineering (PBEE) framework is the probabilistic seismic demand model (PSDM) analysis which provides a probabilistic relationship between a hazard intensity measure (IM) and seismic demands of structures. Determining the optimum IM is still a challenge especially for coupled bridge-soil-foundation (CBSF) systems when soil failure and liquefaction conditions are considered, since soil-structure interaction and liquefaction make the seismic response of the CBSF system more complex. The objective of this study is to identify an optimal IM to estimate the seismic response of a bridge on liquefiable and non-liquefiable soils with soil structure interaction. A case study CBSF with deterministic geometry and soil and structural modeling uncertainties is used to test alternative intensity measures on the basis of selection criteria such as efficiency, practicality, proficiency and hazard computability. In addition, an alternative measure to obtain the coefficient of variance of the seismic response of the components is proposed. This measure provides a method to compare dispersions across different components and potentially inform monitoring and maintenance decisions. Results show that of the 18 IMs considered, velocity-based IMs (e.g. velocity spectrum intensity, VSI) and structure-based IMs (e.g. spectral acceleration at the fundamental nonlinear period, Sa-nonliner) are optimal IMs. However, when hazard computability is considered, structure-specific IMs are not a good choice and VSI and peak ground velocity (PGV) are recommended for practical use. In addition, although duration has a great effect on liquefaction, all IMs related to duration are the least efficient. Conventional studies for evaluating the seismic response of CBSF systems are often based on peak ground acceleration (PGA), but the findings of this study indicate that VSI and PGV display a good correlation between the seismic demand and seismic intensity, thus reducing the uncertainty in predicting the seismic response of the CBSF systems.