Efficient Fragility Analysis of Cross-Fault Hydraulic Tunnels Combining Support Vector Machine and Improved Cloud Method
Benbo Sun, Mingjiang Deng, Sherong Zhang, Weiying Liu, Jia Xu, Chao Wang, Wei Cui
Abstract
One of the primary concerns in the field of seismic risk assessment for underground structures is establishing an accurate connection between seismic intensity and structural responses. The objective of this work is to conduct a rational support vector machine (SVM) model for generating mass data and improved fragility curves of cross-fault hydraulic tunnels (CFHTs). The results highlight that the 900 sets, multiple earthquake intensity measures, and cubic polynomial kernel function of the SVM model can improve reliability in evaluating structural performance. Additionally, the improved Cloud analysis method is more suitable for seismic performance than the typical Cloud analysis.