Artificial neural networks as a corrector of hydrodynamic modelling results

In this paper, the application of artificial neural networks (ANN) to optimise the results obtained from a hydrodynamic model of river flow was evaluated. The study area is Reynolds Creek Experimental Watershed in southwest Idaho, USA. A hydrodynamic model was constructed to predict flow at theoutlet using time series data from upstream gauging sites as boundary conditions. In the second stage, the model was replaced with an ANN model but with the same inputs. Finally the error of the hydrodynamic model was predicted using an ANN model to optimise the outputs. Simulations were carried out for two different conditions (with and without data from a recently suspended gauging site) to evaluate the effect of this suspension in hydrodynamic, ANN and the combined model. Using ANN in this way, the error produced by the hydrodynamic model is predicted and thereby, the results of the model are improved. In addition, the results of hydrodynamic modelling affected by the suspension of the flow gauging is appropriately improved by neural networks. Combination of these two techniques for this specific application uses the potential of both methods and shows a good performance

Keywords:

Combination of hydrodynamic and ANN , model results optimisation , error prediction , flow prediction by ANN , Neural networks for flood prediction