Source identification for contamination events involving reactive contaminants

The problem of contaminant source identification in a water distribution system can be solved as an inverse problem using a simulation-optimization approach. The optimization method searches for contaminant source characteristics which lead to matching observations at the sensors. Accuracy of identification depends on the quantity and quality of data available at the sensors. The present state of the art in water quality monitoring sensors does not always allow for the detection of different kinds of contaminants in the system and they do not provide continuous contaminant concentration measurements. Some sensors provide an event detection trigger based on a specific concentration threshold yielding a binary detection/no-detection signal. Sensors also routinely monitor water quality parameters such as chlorine and pH. For example, a contaminant present in the system may react with chlorine leading to changes in chlorine-based water quality indicators. These anomalies (or deviations) in the observed water quality in the distribution system can be used as indicators of presence of contaminants in the system. A methodology for identifying the source characteristics using sensors outputting binary signals was presented by the authors recently. In the present study we investigate the interaction of reactive contaminants with chlorine in the system and its effect on water quality indicators. These anomalies indicating the presence or absence of contaminants will be used for determination of source characteristics using an evolutionary algorithm-based simulation-optimization approach.