Probabilistic Nonlinear Displacement Ratio Prediction of Self-centering Energy-absorbing Dual Rocking Core System under Near-fault Ground Motions Using Machine Learning

Abstract

ABSTRACT Near-fault pulse-like ground motions can lead to significant seismic demand on building structures due to velocity pulses. The self-centering energy-absorbing dual rocking core (SEDRC) system is a newly developed seismic resilient structural system. This paper investigates the seismic demand of SEDRC systems subjected to near-fault pulse-like ground motions by determining their nonlinear displacement ratios. Two hundred and five near-fault pulse-like ground motion records are used to consider the uncertainties of seismic events. The influences of design hysteretic parameters and near-fault ground motion characteristics on the nonlinear displacement ratio of the SEDRC system are investigated through parametric dynamic analysis of single-degree-of-freedom (SDOF) systems. The dynamic analyses results indicate that the stiffness hardening ratio α and energy-absorbing ratio β of the SEDRC system, predominant period, pulse period of ground motions, and earthquake magnitude show obvious effects on the nonlinear displacement ratio responses of SDOF systems, while the unloading stiffness ratio ε, site condition, and source-to-site distance show limited or negligible influence on that. The past studies mainly used the mean or median responses of single-degree-of-freedom systems to predict the nonlinear displacement ratio of structures, which may not be enough to guide the design of buildings with great importance (e.g., hospital and fire station). This paper proposes an innovative framework for predicting the nonlinear displacement ratio of structures underground motions using a probabilistic estimation method and machine learning technique. Based on the dynamic analysis results of SDOF systems, a probabilistic model for predicting the nonlinear displacement ratio of the SEDRC system under near-fault pulse-like ground motions is developed through the proposed framework. The proposed framework is also applicable to estimate the seismic demand of other structures under near-fault or far-field ground motions to facilitate the development of the performance-based seismic design.