Seismic vulnerability analysis of adobe structures considering historical Chinese seismic intensity standards

Abstract

To study and compare the differences in evaluating the vulnerability of adobe structures under different seismic intensity standards, a seismic vulnerability assessment model for adobe structures considering typical international seismic intensity standards was proposed. An instrument intensity clustering cloud model considering multidirectional ground motion synthesis parameters (real array monitoring data within multiple intensity zones) was generated. In different historical periods, several versions of seismic intensity standards have been released by the China Earthquake Administration (CEA) to evaluate the degree of damage to engineering structures. This paper innovatively considers the vulnerability assessment scale of adobe structures considering six versions of the Chinese earthquake intensity standard, and a seismic damage probability model of the adobe structure is established considering the impacts of six destructive earthquakes during different periods (China). A macroscopic intensity and actual vulnerability distribution of adobe structures considering six destructive historical earthquakes were generated. A novel vulnerability prediction model for adobe structures considering the evaluation of historical Chinese intensity standards was proposed via Gaussian process regression and algorithms. Using the actual seismic loss dataset of adobe structures (the Wenchuan earthquake in China) and the proposed vulnerability model, seismic vulnerability prediction curves for adobe structures considering six versions of macrointensity standards were generated. A numerical model of an adobe structure was established using numerical simulation methods. Considering the ground motion parameter values specified in different intensity zones of the three Chinese earthquake intensity standards, dynamic response and failure risk analyses were conducted on the numerical model, and earthquake damage stress clouds were generated. A comparison of different macrointensity scales and regression algorithms revealed that the latest version of the Chinese macrointensity standard (CMS-20) can accurately quantify the seismic intensity level and seismic vulnerability of adobe structures. Gaussian process regression has excellent prediction accuracy and can be used for seismic vulnerability assessment of adobe structures.