Ranking Method for Strong Ground Motion Based on Dynamic Response of Underground Structures

Abstract

AbstractIt is well known that different input ground motions result to significantly different underground structural response. Hence, rational selection of input Intensity Measures (IMs) is the basis for accurate seismic design of underground structures. In order to find out the most destructive ground motion for seismic design of underground structures, a ranking method for strong ground motions was presented to get their failure potential sequence which may result to maximum structural response of underground structures against ground motion excitations. In this paper, dynamic time-history analyses under a ground motion cluster were carried out on a typical subway station structure in Class IV site. In the analysis, the general finite element code ABAQUS was used to set up a refined numerical modeling, 64 real ground motion records were selected as seismic wave input, and 12 IMs of 3 groups were selected as original IMs. The moment, shear, axial force of central columns and relative displacement of central column of a subway station structure was used as the Structural Response Parameters (SRPs). Based on the Partial Least Squares (PLS) regression method, a series of composite IMs were developed and investigated. The results show that the composite IMs constructed by PLS regression method have better efficiency. This method can be used to rank the design input ground motion according to the degree of SRPs, which represents the failure potential of structure, and the results can provide a reference for the selection of input ground motions for seismic design of underground structures.KeywordsUnderground structuresComposite intensity measuresDynamic time-history analysisPartial least squares regressionRanking method