The contribution of near-fault ground motion velocity pulse to the seismic response of high-speed railway bridge-track system

Abstract

Velocity pulses are often considered a critical factor leading to significant structural responses and more severe damage under near-fault ground motions. This paper investigates the impact of near-fault ground motion velocity pulse (NGMVP) on the contribution of high-speed railway bridge-track system's seismic response and establishes predictive models, incorporating eight ground motion parameters. With wavelet packet transformation, the original ground motions are decomposed into high-frequency and low-frequency pulse components. These components, along with the original ground motions, are then separately input into the system. Subsequently, the seismic responses of the system are obtained through nonlinear time history analysis. For a more comprehensive characterization of NGMVP, 19 ground motion and pulse parameters are selected. Through regression analysis, the parameters with high correlation with near-fault ground motion response ratio are preliminarily screened. Then, four prediction models are established to realize the prediction from the parameters to the response of each component of the system. Furthermore, this paper also examines how the number of parameters impacts model performance. The results indicate that the near-fault ground motion parameter ratio has a high correlation with the response ratio of low-frequency pulse component, while the correlation with the response ratio of high-frequency residual components is low. Among the four prediction models, the exponential product prediction model has the best prediction effect. The final model formula and parameter values are also established, enabling a quantitative analysis of the contribution of NGMVP to the seismic response of the system.