Modeling of scour depth and length of a diversion channel flow system with soft computing techniques

Abstract This study employed soft computing techniques, namely, support vector machine (SVM) and Gaussian process regression (GPR) to predict the properties of a scour hole [depth (ds) length (Ls)] in diversion channel flow system. The considered different geometries channels (angles bed widths) hydraulic conditions. Four kernel function models for each technique (polynomial function, normalized polynomial radial basis kernel, Pearson VII kernel) were evaluated this investigation. Root mean square error (RMSE) values are 8.3949 training datasets 11.6922 testing datasets, confirming that function-based GP outperformed other predicting Ls. Regarding ds, SVM outperforms models, recording RMSE 0.5175 0.6019 datasets. sensitivity investigation input parameters shows angle had major influence Ls ds.