Modeling and Prediction of Embarkment Dam Displacement under Earthquake Loading Using PSO-ANN

Abstract

Abstract Evaluation of embankment dam displacement (D) under earthquake loading can contribute to safe design of the dam. Due to the complexities encountered when modeling this problem, soft computing can be seen as an appropriate solution for predicting the embarkment dam displacement under earthquake loading. In this research, ANN and PSO were integrated in an attempt to present a relationship for predicting the displacement of embankment dam (D). for this purpose, data from 102 real cases was utilized. Input parameters included the height (H) and natural period of the dam (Td), minimum required yield acceleration to slide a block of the dam body (ay), and magnitude (Mw), dominant frequency (Tp), and peak acceleration (amax) of the earthquake. It was figured out that PSO-ANN outperforms PSO in estimating earthquake-induced dam displacement. Compared to other soft-computing methods for predicting embankment dam displacement under earthquake loading, the hybrid PSO-ANN rendered more powerful and suitable.