Estimating discharge coefficient of side weirs in trapezoidal and rectangular flumes using outlier robust extreme learning machine

Abstract Using the outlier robust extreme learning machine (ORELM) method, discharge coefficient of side weirs placed on rectangular and trapezoidal canals was simulated for first time in this study. The parameters governing including Froude number (Fr), ratio weir length to main channel ( L/b ), flow depth at upstream width y 1 / b ) crest height W L/y wall slope parameter m were initially detected. governing, eight different input combinations defined. By randomly selection approach, 65% data considered train ORELM models rest samples applied test them. correlation coefficient, Nash–Sutcliffe efficiency Scatter Index model calculated be 0.937, 0.869 0.092, respectively. results sensitivity analysis indicated more sensitive W/y than Fr . also compared with support vector optimized genetic algorithm (SVM-GA) (ELM)) four multiple linear regression models, a better performance model. demonstrated higher precision experimental values.