Water Quality Modelling and Control in a Water Treatment Process

Abstract

Drinking water quality is an important issue around the world since the low quality water causes health-related problems and economic losses. To ensure high quality water an efficient monitoring and control of a water treatment process is essential. In this study, two common quality variables of treated water, turbidity and residual aluminium, are modelled using the cross-validation method. Selected variables for developing the models are easy and reliable to measure on-line from raw water source. The linguistic equation (LE) approach based on nonlinear scaling and linear interactions produces models, which can be used in addition to predicting the water quality, for monitoring and controlling the water treatment process. The goal of the control simulation was to minimize the turbidity by controlling the coagulation chemical dose and see how this affects the residual aluminium level in drinking water. The results showed that the developed models were accurate and followed the changes in measured water quality variable. Results of the control simulation suggest that the water quality can be improved by proper control and optimizing the chemical dosing, as minimizing the turbidity reduces the residual aluminium level.