Evaluating performance of various methods in predicting triangular sharp-crested side weir discharge

Abstract

This paper investigated the efficiency of the traditional weir equation (TWE), Domínguez, adjusted Domínguez, and Schmidt approaches, as an alternative to the De Marchi procedure, for computing discharge of a sharp-crested triangular side weir. Comprehensive experimental data were used for the analysis, including 342 data from the present study and 140 data from other sources. The effects of approach Froude number Fr1, the ratio of weir height to upstream flow depth p/y1, and weir apex angle θ on the discharge coefficients obtained from different methods were studied. Sensitivity analysis using the partial swarm optimization-support vector regression method indicated that Fr1, p/y1, and θ affect the discharge coefficients. It was found that Fr1 with sensitivity indices equal to 1.89, 3.74, and 4.04 has the most substantial effect on the De Marchi coefficient, TWE coefficient, and adjusted Domínguez coefficient; meanwhile, p/y1 has the most significant impact on Domínguez coefficient and Schmidt coefficient with sensitivity index equal to 1.57. In addition, it was found that θ had the lowest sensitivity indices in estimating discharge coefficients. New equations for forecasting sharp-crested triangular side weir discharge coefficient were presented based on dimensional analysis. The new De Marchi coefficient executed better for calculating triangular side weir discharge than earlier De Marchi coefficients. Moreover, TWE, Domínguez, adjusted Domínguez, and Schmidt methods performed better than the De Marchi procedure (with MSE = 4.581) in calculating sharp-crested triangular side weir discharge. However, considering the simplicity of the TWE approach compared to other methods, this approach with R2 = 0.975, NSE = 0.975, MSE = 3.610, MRE = 0.097, and CP10% = 71.36 was introduced as the superior procedure.