Predicting Discharge Coefficient of Triangular Side Orifice Using LSSVM Optimized by Gravity Search Algorithm

Side orifices are commonly installed in the side of a main channel to spill or divert some flow from source lateral channels. The aim present study is accurate estimation discharge coefficient for through triangular (Δ-shaped) by applying three data-driven models including support vector machine (SVM), least squares (LSSVM) and improved gravity search algorithm (LSSVM-GSA). was estimated utilizing five dimensionless variables resulted experimental data (570 runs). Five different scenarios were applied based on input variables. evaluated several statistical indices graphical charts. results showed that all could successfully estimate Δ-shaped with adequate accuracy. However, LSSVM-GSA produced best performance combination highest coefficients determination (R2) Nash–Sutcliffe efficiency (NSE), equal 0.965 0.993, root mean square error (RMSE) absolute (MAE), 0.0099 0.0077, respectively. RMSE SVM LSSVM 26% 20% estimating coefficient. Furthermore, ratio orifice crest height (W/H) identified as having influence among various