Vector-valued seismic demand analysis within earthquake loss estimation

• Main Author: Modica, Alice
• Other Authors: Stafford, Peter
• Published: Imperial College London 2013
• Subjects: 624
• Online Access: https://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.749147

A fully probabilistic framework for assessing the degree of damage that an earthquakeinduced ground-shaking at a site might in ict upon structures of a given type is computed by combining probabilistic descriptions of the ground motion with predictions of structural response. When the earthquake-induced ground-shaking is described by a single Intensity Measure (IM) and the degree of damage is related to a single Engineering Demand Parameter (EDP), only two pieces of information are required: (1) the EDP predictive equation (which takes the IM as an input) and its associated standard deviation; and (2) the IM hazard curve which is obtained from a probabilistic seismic hazard analysis. However, a single IM does not reflect many of the aspects of the ground motion that affect inelastic stiffness and strength degradation. More accurate estimates of damage can be achieved by describing multiple attributes of the ground motion, as well as multiple indices of the dynamic response of the structure. The vector-valued seismic demand analysis framework presented herein considers innovative vectors of two or more IMs and investigates the joint occurrence of these parameters for the prediction of scalar and vector EDPs. The efficiency of scalar and vector IMs is quantifed via single and multiple regression analyses performed upon demand measures obtained from nonlinear time-history analyses of frame models. The dataset used to develop these relationships is obtained by conducting a large number of NTHA using a suite of reinforced concrete moment-resisting frames representative of those within the Euro-Mediterranean areas. The uncertainty in the calculation of the probability of exceeding different levels of demand is reduced by minimising the standard deviation of the EDP prediction, through the use of a vector model. The framework proposed permits the derivation of fragility functions for Earthquake Loss Estimation that take a vector of IMs as input and provide a vector of EDPs as output. This vector of EDPs can then be used to obtain a robust estimate of the damage associated with the scenario under consideration.