Sediment model for natural and man-made channels using general regression neural network

Abstract

This paper presents a new sediment transport model using general regression neural network (GRNN) that are applicable for both natural and man-made channels. GRNN is a supervised network that trains quickly sparse data sets. The network architecture responses very well with data that is spasmodic in nature than back propagation algorithm. Field data (499 data) extracted from rivers in Selangor, Perak and Kedah are used in the training and testing phases. The model is further tested using hydraulics and sediment data from rivers in the United States namely Sacremento, Atchafalaya, Colorado, Mississippi, Middle Loup, Mountain Creek, Niobrara, Saskatchewan, Oak Creek, Red, Rio Grande rivers and Chop Irrigation Canal. Four independent variables, namely, relative roughness on the bed (R/d50), ratio of shear velocity and fall velocity (U*/Ws), ratio of shear velocity and average velocity (U*/V) and the Froude Number (V/√gy) are used as input variables in the input layer and the total sediment load QT as the output variable. The proposed GRNN sediment model had accurately predicted 89% of the river data sets (local and foreign rivers) with 90% of the predicted values lie in the discrepancy ratio of 0.5 – 2.0. For the sake of illustrations, accuracy of the derived sediment transport model is also measured using smaller range of discrepancy ratios.