Multivariate Adaptive Regression Splines Model for Prediction of Local Scour Depth Downstream of an Apron Under 2D Horizontal Jets

Abstract

Sluice gates commonly control water levels and flow rates in rivers and channels. They are also used in wastewater treatment plants and to recover minerals in mining operations and in watermills. Hence, scour phenomena downstream of sluice gates have attracted the attention of engineers to present a precise prediction of the local scour depth. Most experimental studies of scour depth downstream of sluice gates have been performed to find an accurate formula to predict the local scour depth. However, an empirical equation with appropriate capacity of validation is not available to evaluate the local scour depth. This study presents the application of multivariate adaptive regression splines (MARS) to evaluate the local scour depth downstream of sluice gate using 228 experimental case studies of the scour depth downstream of sluice gates with an apron. MARS is used to develop empirical relations between the scour depth and various control variables, including the sediment size and its gradation, apron length, sluice gate opening, and the flow conditions upstream and downstream of the sluice gate. Six non-dimensional variables were given to determine a functional relationship between the input and output parameters. The efficiency of MARS model is investigated with ANN model in the training stages. On the other hand, performances of the testing results for this model are compared with the ANN model and traditional approaches based on regression methods. The uncertainties prediction of the MARS was quantified and compared with ANN model. Also, sensitivity analysis was performed to assign effective parameter on the scour depth prediction.