Application of Artificial Neural Network for Reducing of Chlorine Residual Concentration in Water Distribution Network

It is difficult to estimate residual chlorine at the dead-end area of the water distribution network because chlorine consumption is influenced by various factors. Therefore, there are many water utilities that control the amounts of chlorine in reservoirs using empirical trial-and-error methods to maintain safe levels of residual chlorine in the distribution system. In this study, an ANN model of residual chlorine concentration is proposed which could be used to reduce in chlorine use in water distribution system. The ANN model with best performance was selected by training and verification. The five scenarios for the reduction in chlorine use were analyzed by setting the input chlorine as low as 0.05~0.25 mg/L compared with the input chlorine observed in the time series. Case 4 is the best to be satisfied with the input condition (0.4 mg/L or more) and output condition (0.34 mg/L or more) at the same time. It is possible to reduce chlorine in use up to 0.2 mg/L in the maximum amount.