Contemporary issues in earthquake engineering research: processing of accelerometric data, modelling of inelastic structural response, and quantification of near-fault effects

Abstract

This study focuses on three contemporary issues related to the fields of EarthquakeEngineering and Engineering Seismology: (1) processing acceletometric data andassessing permanent ground displacements, (2) modelling the response spectra ofinelastic structures, and (3) quantifying near-fault ground-motion effects relevant toengineering structures.Most of the noise in accelerometric data can be removed by standard signal processing,for example, low-pass and high-pass filters. High-pass filters, if used to adjustbaseline errors, can result in inadvertent loss of long-period signal due to, for example,permanent ground displacements in the near-fault area. Despite several efforts,a consistent and well-defined method to adjust baseline errors without losing a partor the whole of permanent displacements (if any) is not available in the literature.This study develops and thoroughly tests such a method by using several near-faultaccelerograms.An inelastic response spectrum is required in the design of earthquake-resistant structures.In the current practice, it is obtained by scaling an elastic response spectrumby some ad-hoc factors, such as, structural behaviour factor, or force reduction factor.This approach lacks a rational basis and is highly uncertain. This study showsthat inelastic response spectra can be modelled in a manner similar to elastic responsespectra by calibrating Ground-Motion Prediction Equations (GMPEs) forinelastic structures. This is achieved by regressing peak responses of inelastic structuresagainst earthquake parameters (such as, size, faulting mechanism, etc.), pathparameters (such as, source-to-site distance), site parameters (such as soil conditions),and structural parameters (such as, undamped natural period of vibration,damping ratio, displacement ductility, yield strength, etc.). The resulting GMPEscan be applied in Probabilistic Seismic Hazard Assessment (PSHA) and ProbabilisticSeismic Demand Assessment (PSDA) to estimate uniform-hazard representations ofdesign forces and ductility demands with greater reliability and accuracy than bytraditional methods.Near-fault ground motions affected by forward-directivity effects, which are on thefocus of this study, contain most of their energy in a narrow frequency band, andtherefore affect certain structures more severely than others. To ensure reliable design/evaluation of structures, important properties of near-fault ground motions, suchas their amplitude and frequency content, need to be properly quantified. In thisstudy, peak ground velocity (PGV) and predominant period (Td) of ground motionare used to characterize near-fault ground motions. Mathematical models relatingPGV and Td to earthquake size, source-to-site distance, etc, are calibrated using alarge set of near-fault ground-motion data. These parameters, along with structuralproperties, such as, undamped natural period and damping ratio of an oscillator, areused to quantify the elastic as well as inelastic response spectra of pulse-like groundmotions in the near-fault area more adequately than was possible using earlier approaches.