A novel method for the dynamic reliability analysis of slopes considering dependent random parameters via the direct probability integral method. (September 2022)

• Record Type: Journal Article
• Authors: Zhou, Yang; Jing, Mingyuan; Pang, Rui; Xu, Bin; Yao, Haoyu
• Is Part Of: Structures. Volume 43(2022)
• Page Start: 1732
• Page End: 1749
• Publication Date: 2022-09
• Subjects: Stochastic dynamic response -- Vine Copula -- Direct probability integral method -- Seismic reliability analysis -- Parameter correlation; Structural engineering -- Periodicals; 624.1
• DOI: 10.1016/j.istruc.2022.07.074
• Languages: English
• ISSNs: 2352-0124

Abstract

Abstract: The uncertainty of ground motions and soil properties is the focus of stochastic dynamic analysis of slopes due to its significant impact on the safety evaluation of seismic performance. This paper analyzes and discusses the influence of stochastic ground motions and correlated random parameters on the stochastic dynamic system of slopes and studies the necessity of considering uncertain factors in slope performance evaluation from the perspective of probability. A reliability evaluation framework based on the direct probability integral method (DPIM) is proposed to solve the difficult problem of stochastic dynamic solutions of complex geotechnical structures. Firstly, a point set selection strategy that generates more representative and multidimensional dependent variables is described. This method not only uses the technology of minimizing GF discrepancy (Generalized F discrepancy), but also uses the vine Copula model to construct the correlation between multiple geotechnical parameters for the first time. Then, the method of generating nonstationary stochastic ground motions is introduced as stochastic excitation to construct a coupled stochastic system of soil parameters and ground motions. Finally, the stochastic dynamic analysis framework based on the DPIM is established, and the average large failure probability is further proposed to evaluate the seismic performance of slopes. Three classical slope cases with material parameter correlations are used as examples of the application of the proposed method. The effectiveness and accuracy of the proposed framework are verified by comparison with the probability density evolution method (PDEM), and it is proven that the computational efficiency of the DPIM is significantly higher than that of the PDEM. Furthermore, the results also indicate that the correlation of slope parameters has an impact on the stochastic system, and this correlation cannot be completely ignored.