Author(s): Haitham Abdulmohsin Afan; Wan Hanna Melini Wan Mohtar; Ahmed El-Shafie
Linked Author(s):
Keywords: No Keywords
Abstract: The need to predict suspended sediment load (SSL) is accelerated due to the unsustainable approach of anthropogenic activities such as deforestation, poor agricultural practices, and massive development. The transportation of sediment to and through the river system consists of a number of complex phenomena including fluid-sediment interaction and the characteristics of both flow and sediment should be taken into account, of which the prediction through conventional method might be exhaustive. This research introduces an artificial intelligence prediction model for SSL. A combination of genetic algorithm with radial basis neural network presented to select the relevant input and predict the SSL in daily time scale based on SSL and water discharge data for Kelantan River in Malaysia. The combined model (neuro-genetic) simplifies the task of selecting the relevant input variables for better SSL prediction. The neuro-genetic model is able to analyse up to10 input combinations and eliminate the conventional trial-and-error input selection procedure. The proposed model proved its ability to accomplish satisfactory results. The best model shows satisfactory results within 8-input combination (with R2=0.9088 and relative-error of SSL cumulative 0.53%).
Year: 2018