Establishment of Relationship Between Coagulant and Chlorine Dose Using Artificial Neural Network

Abstract

Multiple treatment phases are involved in a water treatment plant (WTP), but coagulation and disinfection are the most crucial for producing safe and clear water. Determining the optimal coagulant and chlorine doses in the laboratory is time-consuming and poses a significant challenge in water treatment. To streamline this process, artificial neural network (ANN) models have been developed to predict the chlorine dose based on the coagulant dose. Studies comparing various ANN models indicate that the radial basis function neural network (RBFNN) model provides excellent predictions (R = 0.999). In modeling with radial basis function neural networks (RBFNN) and generalized regression neural networks (GRNN), the spread factor was varied from 0.1 to 15 to achieve a stable and accurate model with high predictive accuracy. Employing soft computing models to define the coagulant and chlorine doses has proven highly beneficial for the management of WTPs, significantly enhancing the efficiency and accuracy of dosing predictions.