AI-driven Predictive Analysis of Seismic Response in Mountainous Stepped Seismic Isolation Frame Structures

Abstract

In this paper, we propose a unique method for rapid prediction of seismic response of stepped seismic isolation frame structures in mountainous areas using artificial intelligence (AI), based on which the results of seismic response can be used to determine the damage level of stepped seismic isolation frames in mountainous areas under seismic action, and thus to make seismic damage prediction. This study fills the knowledge gap in earthquake damage prediction for stepped isolation frame structures in mountainous areas. In this study, a number of 7-story typical RC frame structures were designed using the structural design software Midas Gen. The dynamic time-history analyses of the structures were carried out using the control variable method, and based on the results of the analyses, five factors that have a greater impact on the seismic performance of mountainous step-isolated frame structures were obtained, which are: the arrangement of seismic isolation bearings, the degree of regularity of the structure, the intensity of defense, the type of the site, and the seismic intensity. based on the results of the dynamic time course analysis, a seismic sample library with a sample capacity of 384 is established by combining these influencing factors. Each influence factor is given a suitable domain and affiliation function, and fuzzy rules are established according to the seismic sample library, and a fuzzy inference model is established by using the fuzzy logic toolbox in MATLAB. The model can directly determine the damage state of the predicted structure. Random sampling confirms the stability and accuracy of the model for different times to build a framework. The results show that the method of analysis is correct, fast and efficient and the seismic related selected factors can predict and map the seismic damage prediction of the model structure. This method can also be applied to rapid seismic damage prediction for SSIFS (stepped seismic isolation frame structures) in rocky areas.