Mathematical and artificial neural network models for simulation and optimization of chlorine residuals in water distribution systems

Management of adequate chlorine residuals in the WDS (Water Distribution System) has proven difficult for large cities because of the diversity of retention time and water quality properties. This thesis describes a management tool called EQM (Effluent Quality Manager) that overcomes these difficulties. EQM analyzes critical water quality properties and reservoir retention times to guide plant managers and operators to find adequate chlorine concentrations in the plant effluent to ensure sufficient residuals in the WDS. The EQM program can operate "on the fly" because it has the ability to directly analyze on-line PI (Production Information) about the effluent water quality properties using various sensors. Consequently, EQM can quickly analyze new circumstances independent of a large volume of information or complexity, thus helping to avoid decision ambiguity. In addition, formation of DBPs (Disinfection By-Products) such as TTHM (Total Trihalomethane) can also be predicted using the EQM program.; Development and verification of mathematical and ANN (Artificial Neural Network) algorithms used in the EQM program was based on extensive lab experiments of chlorine decay and TTHM formation-decomposition, as well as collection of field data on chlorine residuals in the WDS using on-line analyzers. Efficient simulation and optimization of chlorine residual and TTHM concentrations in the WDS requires modeling algorithms that are both simple and accurate. Therefore, among the main challenges of this work was to verify and improve the existing kinetic models of chlorine decay and TTHM formation, as well as to link these models to the critical water quality properties.; Interpretations of various operational and water quality scenarios using EQM, as well as mass balance models of chlorine in reservoirs indicate that there is a strong need to more efficiently reduce TOC (Total Organic Carbon) in the plant effluent, which is the main cause of water quality related problems. This thesis also presents a number of operating tools that can be used in chlorine management, which include pH and TOC control, as well as structural modifications to reservoirs.