Assessment of Sediment Concentration Through RBNN and SVM-FFA in Arid Watershed, India

Abstract

For scheduling, designing and management of hydraulic structure and hydrologic model prediction of sediment concentration play a crucial role in water resources engineering. Here, prediction of the sediment concentration using SVM-FFA and radial basis function network (RBFN) models are done. Parameters like rainfall, temperature, and sediment concentration on a monthly basis from Puintala, Odisha, India watershed are considered. SVM-FFA model shows better performance than the RBFN model. Among five input scenarios, St−1, St−2, St−3, St−4, St−5 give most excellent WI value 0.965 and 0.943 for testing and training phases. In case of RBFN, best WI value is 0.8636 and 0.8911 for training and testing phase when 3-0.5-1 architecture is used as an input.