Multi-objective Sensor Network Placement Model for Integrated Monitoring of Hydraulic and Water Quality Parameters

1. Abstract Near real-time continuous monitoring systems have been proposed as a promising approach for helping drinking water utilities detect and respond quickly to threats related to the normal operation of the water network. Water quality sensors may detect contamination events that pose a growing threat to public health, while pressure and flow sensors are used for estimating the hydraulic state of the system. So far, optimization algorithms for the placement of sensor nodes in water networks were implemented separately for water quality and hydraulic applications. Deploying two independent sensor networks within one distribution system is not a cost-effective approach, since an in-line sensor tapping point is expensive to install and maintain, and is bounded by physical constraints such as limited access to underground pipelines and/or power sources. Therefore, there is a need to consider a sensor network with a dual hydraulic water quality monitoring capability (i.e., sensor nodes that collect both hydraulic and water quality data at the same tapping points). An additional benefit from this integrated monitoring approach is the possibility of reducing false positive detections of contamination events by associating abnormal water quality records with observable changes in the hydraulic data. In this study, a multi-objective sensor network placement model for integrated monitoring of hydraulic and water quality parameters using the nondominated sorted genetic algorithm II method is demonstrated on two water distribution systems of increasing complexity. Tradeoffs between competing hydraulic and water quality design objectives are explored and analyzed.