Estimation of the Scouring Depth of the Plunge Pool of the Symmetrical Crossing Jets by Support Vector Machine

Abstract

At the downstream end of the large dam spillways, the plunge pool is used for energy dissipation due to the flow. This type of structure has a scouring hole downstream of the dam, which needs to be explored for the issue. The present research work aimed to examine the performance of the support vector machine (SVM) method as one of the familiar soft computing techniques used for estimating the scouring depth of symmetrical crossing jets of a plunge pool. Firstly, an appropriate percentage run and Kernel of the SVM were determined by using laboratory data. Then, the parameters involved the scouring depth including Dencimetric Froude Number, Tailwater Depth, vertical jet angle, angle of the impact point of the jet, and the relative distance between of impact of the jet up to the water surface (Frd90 ،Tw ،αv ،αc ،δ), were divided into different models and the runs were performed. Numerical values of the evaluation criteria as R, RMSE, and NRMSE related to the test stage results for the best model to estimate the scouring depth were found as 0.9563, 0.688 and 18.47%, respectively. The results of this model confirm the acceptable performance of the SVM for estimating the scouring depth studied in present work. To find a best model, the results of sensitivity analysis reveals that the parameters Tw and αv have the highest and lowest effects on the accuracy amount of estimation for scouring depth, respectively.

Keywords

• scouring
• dencimetric froude number
• tailwater depth
• soft computing