Using a hybrid artificial intelligence approach to estimate length of the hydraulic jump caused by novel kind of Sharp-Crested V-notch weirs

Abstract

In this study, the SCVW (a novel type of Sharp-Crested V-notch weirs) is employed for more experimentally and theoretically investigations. The length of the hydraulic jump at downstream of the SCVW (LjSCVW) is measured via new supplementary experimental datasets via most popular vertex angles θ (128° and 60°). The experiments are conducted under aerated, steady, and free overflow conditions in an open channel for large physical models. To assess the variations of LjSCVW versus the θ, extensive laboratory work is performed at different discharges (Q), diverse triangular segments number (Nst) in the SCVW, and various tailgate angles (ɸ). Via dimensional analysis and pre-processing method, most effectual parameters on the LjSCVW were obtained as Fr1 (approaching Froude number), Re1 (incoming Reynolds number), and (relative sequent depths). Based on the experimental results, by increasing the values of Fr1 and Re1, the values of LjSCVW noticeably increased. Three types of data-driven models (DDMs), namely, support vector regression (SVR), gene expression programming (GEP), and a robust hybrid model entitled hybrid (SVR-ACO) are developed to estimate the LjSCVW. The suggested hybrid model is a coupled form of SVR with ant colony optimization (ACO) algorithm, which is used to enhance the estimation precision of the LjSCVW. According to the attained statistical indices (determination coefficient (R2) = 0.98, root mean square error (RMSE) = 0.191, and value of total grade (TG) = 19.78) and scatter plots, the hybrid SVR-ACO model was determined as the most superior method to estimate the LjSCVW with high accuracy.