Abstract:The reservoir bank slopes of high-dam and large-reservoir hydropower stations are significantly affected by complex geological conditions. Traditional prestressed anchor cable reinforcement designs often neglect the uncertainty of soil strength parameters, and the limit equilibrium method struggles to effectively address complex geological environments, resulting in insufficient reliability of the analysis results and low computational efficiency. To address the aforementioned issues, this study innovatively proposed an efficient reliability analysis method integrating active learning support vector machine (AL-SVM) with the strength reduction method. The support vector machine could efficiently approximate the instability criterion of the slope, replacing the traditionally time-consuming strength reduction process to achieve rapid reliability assessment. Additionally, the active learning algorithm iteratively and actively selected critical sample points near the decision boundary, substantially reducing the computational cost required for training the support vector machine. Taking the reservoir bank slope of the Lianghekou hydropower station as a case study, the proposed method showed an improvement in computational efficiency of approximately 98% compared to the traditional Monte Carlo method, while maintaining high accuracy and quantitatively revealing the influence of reinforcement parameter variations on slope failure probability. This method significantly improves the efficiency and accuracy of reliability analysis for the design of slopes reinforced with prestressed anchor cables under complex geological conditions, and provides an innovative solution to the challenges of reliability analysis under complex working conditions.
