Modeling and Assessment of Discharge Coefficient of Arc Labyrinth Weir Using Experimental and Meta-model Methods

While having economic advantages, nonlinear labyrinth weirs have more passing flow capacity than linear weirs. Having a high capability of extracting hidden complex relationships among dependent and independent variables besides saving financial and time, intelligent algorithms are economic and time-saving and have dedicated a remarkable role among researchers. In this research, the performance of support vector machine (SVM) and gene expression programming (GEP) algorithms is figured out to predict the discharge coefficient (Cd) of the arched labyrinth weir using 226 experimental data series. Involved geometric and hydraulic parameters are total head (Ht), weir height (P), cycle arc angle (θ), Froud number (Fr), cycle wall length (Lt), the width of a cycle (w), weir nose length (A), an increase of weir height of 15% and change of weir crest shape change to quarter circle (U). Results showed that the maximum values of the Cd belong to arc labyrinth weir of arc angle 40 degrees. Numerical simulation illustrated that combination of (c، u،  ، ، ،) and (c، u،  Fr، ، ،) parameters have optimum performance in the SVM and GEP algorithms of assessment indices as (R2=0.9791, RMSE=0.03, DC=0.9776) and (R2=0.9871, RMSE=0.0231, DC=0.9856), respectively; showing highly accurate performance of two algorithms in the prediction of the Cd.