(1) Water Technologies Department, Kuwait Institute for Scientific Research, P.O. Box 24885, Safat, 13109, Kuwait;(2) Department of Electrical Engineering, Florida Atlantic University, Boca Raton, Florida, 33431, U.S.A.
Abstract:
Prediction of critical desalination parameters (recovery and salt rejection) of two distinct processes based on real operational data is presented. The proposed method utilizes the radial basis function network using data clustering and histogram equalization. The scheme involves center selection and shape adjustment of the localized receptive fields. This algorithm causes each group of radial basis functions to adapt to regions of the clustered input space. Networks produced by the proposed algorithm have good generalization performance on prediction of non-linear input–output mappings and require a small number of connecting weights. The proposed method was used for the prediction of two different critical parameters for two distinct Reverse Osmosis (RO) plants. The simulation results indeed confirm the effectiveness of the proposed prediction method.