Analysis of the stochastic dynamic response of Asphalt–concrete core wall dams under near-fault pulse-type stochastic ground motions

• Published in: Soil dynamics and earthquake engineering (1984) Vol. 199; p. 109703
• Main Authors: Wang, Zongkai, Song, Zhiqiang, Li, Chuang, Liu, Yunhe
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01.12.2025
• Subjects: Asphalt–concrete core wall dam; Direct probability integration method; Near-fault ground motion; Random forest; Stochastic ground motion; Stochastic seismic response; Stochastic ground motion; Asphalt–concrete core wall dam; Near-fault ground motion; Random forest; Direct probability integration method; Stochastic seismic response
• ISSN: 0267-7261
• DOI: 10.1016/j.soildyn.2025.109703

Asphalt–concrete core rockfill dams (ACCRDs) are evolving to meet the demands of larger scales, more complex geological settings, and higher seismic performance requirements, especially in near-fault regions where seismic responses become more complex due to the pulse characteristics of ground motions. Thus, investigating the seismic response of ACCRDs under near-fault pulse-type ground motions is of significant practical importance. This study uses actual near-fault ground motion records and employs a random forest algorithm to establish a regression relationship between pulse parameters and seismological parameters. A method combining high- and low-frequency components is proposed to generate near-fault pulse-type stochastic ground motions for different site conditions. Additionally, by integrating the number-theoretic point selection method with the direct probability integration method (DPIM), a stochastic dynamic response analysis approach for ACCRDs is developed. Using a real-world case study, stochastic dynamic calculations are conducted under near-fault pulse-type ground motions. The results show the probabilistic evolution of the principal tensile stress in the asphalt core and the maximum failure probabilities. This study highlights that response calculations based on a single ground motion are insufficient to reflect the actual seismic behavior of a dam. Therefore, considering the stochastic nature of ground motions is crucial in near-fault seismic response analysis. •A regression model relating pulse features to seismic parameters is built from near-fault data.•A method for generating near-fault ground motions is developed from the new regression model.•The stochastic response of the dam to near-fault ground motion is probabilistically studied.•The evolution of the response probability density of the dam was analyzed using DPIM.