Suspended load routing using artificial neural network and 1D fully coupled model (Case study: Ahwaz Station, Karoon, Iran)

Abstract

Sediment load estimation is one of the challenges of river engineering. More researches have been conducted to develop a perfect model to sediment transport simulation. Analytical and data-iven models are two main groups of models. In this paper, one dimensional fully coupled model and artificial neural network models performance is compared in sediment rating curve simulation in Ahwaz station, Karoonriver, Iran. 1D fully coupled model has calibrated and validated using ash-Sutcliffe coefficient. The magnitude of 0. 15 and 0. 19 of NS coefficient for calibration and validation periods of coupled model represent good agreement of the model with average condition of river. According to calculation, derived sediment rating curve using ANN with FFBP algorithm, has good agreement with measured rating curve. In high flows, both two models have difference with measured data. In general ANN model has more accuracy than coupled model.