Placement Strategies for Water Quality Sensors Using Complex Network Theory for Continuous and Intermittent Water Distribution Systems

Abstract

Water quality sensors are used to detect contamination in water distribution systems (WDSs) to help supply safe and quality drinking water to society. However, identifying the optimal location to place sensors is still an open challenge. Many approaches have been proposed in literature to solve this problem. Complex network theory-based approaches (CNW) to sensor placement are simple, easy to implement, take less computational time even for large-scale networks, and do not require calibrated hydraulic and water-quality models. However, existing CNW methods perform well for only a subset of common objectives. Optimization-based approaches offer better placement but are computationally costly and require detailed knowledge of the WDS. We proposed a new method, “EQ-Water,” to identify the locations to place water quality sensors for continuous and intermittent WDS with variable demand patterns. EQ-Water is based on complex network theory, but uses minimal additional hydraulic information. We validate the performance of EQ-Water on four real networks: BWSN 1, BWSN 2, JPN, and D2B networks. We have compared the performance with a number of approaches from literature, including the popular TEVA-SPOT tool. The comparison is based on the four objective functions, Z1–Z4, which are commonly used, and also on a weighted cumulative objective function. Our results indicate that EQ-Water is among the top three methods for BWSN 1, JPN and D2B when using diverse weights for the objective functions, and it shows median performance for BWSN 2. We also observe consistently superior performance against other CNW, and it is competitive with simulation-based approaches while taking lesser time and effort.