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.

No abstract

No abstract

A B S T R A C TPipelines are a very important component of water supply systems. Specially, the pipe bursts and leaks are very useful indicators to show the condition of the network. To keep and improve the performance of the system, much accumulated know-how for inspections, operation, maintenance, and suitable rehabilitation to achieve the best performance is needed, as well as a logical method that can estimate the optimal time and range of replacement/rehabilitation work with an understanding of deterioration factors of pipe networks. Therefore, in this study, a statistical probability model for pipe burst risk was developed with various data from leak-repairing records and local characteristics of the circumstance on the real-scale distribution system in Seoul in order to utilize this method for management and operation of the water pipe network, including prioritization of pipe replacement/ rehabilitation work.

No abstract