Evaluation of ground motion scaling techniques

Abstract

Nonlinear Time History Analysis (NTHA) is a rigorous technique requiring the analyst to assemble a suite of ground motions to perform structural analyses. In current practice and literature, there are several techniques available to select and scale records: however, there is little or no consensus among researchers or designers regarding the accuracy of these techniques towards acceptability in structural design. The selection criteria for ground motions are unclear: parameters such as earthquake magnitude, epicentre distance, soil type, geotectonic setup and, the ratio of peak ground acceleration to velocity (A/V) have been used to develop ground motion suites. There is ambiguity in the current ground motion modification procedures as well: three techniques, namely, amplitude scaling, spectral matching and normalisation of records to the expected peak ground acceleration (PGA) are employed variously by researchers. Some codes allow the use of only one record while others permit three to seven records. In this study, the effect of selection and scaling of records on the structural response is examined using a single degree of freedom (SDOF) model of an unreinforced masonry wall. The structural response parameter is wall displacement. Data from major Himalayan earthquakes is used to develop a suite of records. Statistics of the storey displacement indicate that the amplitude scaling of records to match the PGA leads to inaccurate estimates of displacements. Amplitude scaling of records at the structure’s fundamental period and spectral matching of records in a defined period range lead to better estimates of actual storey displacements. The use of one or three records for analyses leads to an incorrect estimation of seismic demand.