Cavitation Damage Prediction on Dam Spillways using Fuzzy-KNN Modeling

Abstract

The present paper deals with a numerical method for prediction of cavitation damage level and location on dam spillways. A method was applied to predict the intensity of cavitation damage to spillways, using the fuzzy k-nearest neighbor algorithm. Five levels of damage intensity were considered to predict cavitation damage in the spillway of Karun-1 Dam in Iran. According to the results, the proposed model could properly predict the location and intensity of damage in comparison with the actual damage reports of past floods. According to the Pearson's correlation coefficient, mean absolute error, coefficient of residual mass, and normalized root mean square error, the fuzzy k-nearest neighbor model is efficient and suitable.