Modelling sediment concentration using back propagation neural network and regression coupled with genetic algorithm

Abstract

Prediction of minimum sediment concentration is vital for planning, designing and management of hydraulic structures. This work focused on the prediction of the sediment concentration using regression and Back Propagation Neural Network (BPNN) models. Parameters like discharge, temperature and sediment concentration had been collected on daily basis from different basins on River Suktel. BPNN and Regression models had been used to map the sediment concentration with discharge and temperature. Mutually regression and BPNN models are into consideration for predicting the fitness of models. Regression and BPNN model are then coupled with GA to acquire sediment concentration. For minimum sediment concentration, optimum discharge and temperature were obtained from coupled GA. Comparison between GA-BPNN and GA-Regression models are computed for knowing the sensitivity of models at regional scale. This work is unique in predicting minimum sediment concentration.