Numerical Modelling for Prediction of Suspended Sediment Concentration in Bidor River

Nowadays, improvement of artificial intelligence, as an estimator for hydrological phenomenon has generated an abundant change in estimations. The estimation of sediment concentrations in rivers is highly important for designing and operation of various water resources since the life of water resources structure are directly concerned to the amount of sediments in rivers. In this study, Radial Basis Function Neural Network (RBFNN) model using Thin Plate Spline function was developed to estimate the suspended sediment concentration of a Bidor River in Perak. There are many studies had been through to predict the suspended sediment concentration in the hyper-concentration river by using soft computing techniques. Such as artificial neural network (ANN), support vector machines (SVM), gene expression programming (GEP), and Wavelet – ANN approach. Usually, ANN technique is a preferable to predict the suspended sediment concentration. Due to the lack of accuracy of the result, the different numerical modelling techniques will be used and at the same time to evaluate the performance of the models and recommend the best model among the developed ones. The data for this study are given by Department of Irrigation and Drainage (DID), where 4-years data (1992 – 1995) was applied for training and 1-year data (1996) was applied for testing. The parameters that will be use during the estimation are rainfall and discharge as the input data and suspended sediment as the output data. The root – mean square error (RMSE) and coefficient of determination (R2) were expected to evaluate the model’s performance. The best model performances will be used as estimator in the future studies on designing the hydrology structures.