Prediction and Simulation of Multimedia Filter Performance Using Artificial Neural Network

The objective of this paper is to simulate and Predict suspended solid removal efficiency in multimedia filter depend on experimental work for (43) case studies. The experimental work was carried out on pilot unit by changing each one of the input parameter and get results for each change. The suspended solid removal efficiency obtained from laboratory was used as target function in ANN while the other properties of filter such as raw water quality, operation conditions and media characteristics was used as input parameters. The optimal number of neurons in ANN network structure was investigated in this study. As a result, ANN with back propagation algorithm is a good tool that can use to predict removal efficiency of multimedia filter the results indicated that the BP model had showed a good convergence performance during training, and the model predictions of outflow suspended solid removal efficiency coincided well with the measured values.