Predicting submerged hydraulic jump characteristics using machine learning methods

Abstract

Hydraulic jump typically occurs downstream of hydraulic structures by converting the supercritical to subcritical flow regimes. If the tail-water depth is greater than the secondary depth of the hydraulic jump, the jump will be submerged (SHJ). In these conditions, the momentum equations will not have an analytical solution and a new solution is required. In this study, after dimensional analysis, an experimental study was conducted in a rectangular flume with a length of 9 m, a width of 0.5 m and a depth of 0.45 m in a wide range of Froude numbers (Fr = 3.5 to 11.5) and submergence ratios (Sr = 0.1 to 4). The data were then normalized and divided into two parts of training and testing. A new technique, DGMDH, was used to predict the submerged hydraulic jump characteristics. The results were then compared with the GMDH model. The results showed that the DGMDH model estimated the relative submergence depth, jump length, and relative energy loss with accuracy of R2 = 0.9944 and MAPE = 0.038, R2 = 0.9779 and MAPE = 0.0387, and R2 = 0.9932 and MAPE = 0.0192, respectively. While the accuracy of the GMDH model for relative submergence depth, jump length, and relative energy loss was respectively R2 = 0.9923 and MAPE = 0.043, R2 = 0.9671 and MAPE = 0.0527, and R2 = 0.9932 and MAPE = 0.0192. Due to superiority of the DGMDH model over the GMDH model, it is recommended to use this model to estimate the submerged hydraulic jump characteristics. Highlight The results showed that the DGMDH model has more accurate results than the GMDH model in predicting the relative submergence depth, jump length, and relative energy loss.;

Keywords

• artificial intelligence
• data mining
• experimental study
• submergence hydraulic jump