Neural Network Modeling of Scour Depth Downstream of Spillways Available to Purchase

The evaluation of the depth of scour downstream of spillways is of utmost importance for stability analysis purposes and safety assessment of hydraulic structures. To model the problem a multilayer perceptron network was adopted with a hidden layer. Three input parameters were used to describe the hydraulic conditions: type of spillway, specific discharge, and hydraulic head; two parameters were used to describe the rock mass, one for the intact material and the other for the spacing of the joints. The output parameter was the spillway depth of scour. The network training process was successful and the error evaluation shows that the result is promising for reliable engineering predictions of new projects.