Support vector regression for modified oblique side weirs discharge coefficient prediction

Abstract

Accurate determination of discharge coefficient is one of the major concerns in the process of the designing of side weirs. Relation between the modified side weirs discharge coefficient to various geometric and hydraulic situations leads to a high flow complexity around the weirs. In this study, two types of support vector regression (SVR) methods were employed to model the discharge coefficient of a modified triangular side weir. Two types of SVR are obtained by using the radial basis and polynomial as the kernel functions. Six different non-dimensional input combinations with different input variables were used to find the most appropriate one. The results show that both SVR-rbf and SVR-poly methods perform better when the number of input variables is higher, and there is no compaction in the non-dimensional input variables. Comparison between the investigated models shows that the SVR-rbf by RMSE of 0.063 performs much better that SVR-poly by RMSE of 0.084.