Modelling suspended sediment transport using an integrated numerical and ANNs model

An integrated sediment transport model, combining a deterministic numerical model with the data driven artificial neural networks (ANNs), has been developed to predict the distribution of suspended sediments under non-uniform flow conditions. In this model the transport process of sediments within the water column is simulated by numerically solving the advective-diffusion equation, with the bottom boundary condition being specified using a near-bed reference concentration. The reference concentration is estimated based on an ANNs approach. The finite volume method is used to solve the hydrodynamic and sediment transport equations with the sigma coordinate system being used to represent the model bathymetry in the vertical plane. The integrated model has been applied to three tests under both uniform and non-uniform flow conditions and the modelling results are very encouraging.