Reservoir bank slope stability prediction model based on BP neural network

Abstract

Safety monitoring and stability analysis of high slopes are essential for construction of concrete dam in precipitous gorges or mountainous areas. The estimate of slope stability is a difficult engineering shortcoming with a number of variables. Thereafter, a hybrid model of Support Vector Regression (SVR) and Teaching–learning-based optimization technique (TLBO) is proposed to develop the predicting function. TLBO was used in obtaining the best SRV factors to improve the prediction accuracy. Few essential factors, such as the installation height of instruments, classification of rock masses, modulus of elasticity, the complete measuring time cycle, the excavation height of slope, the start measuring time, and the actual excavation height after measurement are considered as the input parameter, but the slope displacement is regarded as output. The outcomes showed SRV-TLBO a reliable hybrid accurate prediction of slope stability, then it was effectively used to the left abutment slope of Jinping I hydropower station located in Yalongjiang concrete dam reservoir as a novel method for this purpose.