Municipal water distribution network is prone to failures of very different types, such as water contamination, broken or fractured pipes, and water leakage due to distribution system aging and deterioration. To overcome such faults in the network, and for better fault detection and maintenance, the present research proposes a systematic sensor placement method to ensure network monitoring and observability that speeds up the process of fault detection and diagnosis. The proposed monitoring system guarantees that the system preserves its connectivity in case of several failures occurrences in the water monitoring network and also whether the observability radius remains at the desirable threshold. The proposed method is based on establishing a communication network over the water distribution network by installing wireless sensors over the water pipes. Sensors could communicate with each other and send packet data for reporting their status and conditions. The sensor placement methodology has four objective functions which are the minimum number of sensors, maximum communication range, the maximum number of failures that could occur while the network remains connected, and also minimizing the observability radius while some of the sensors fail. The problem of sensor placement regarding four objectives functions was solved using a multi-objective genetic algorithm. A hierarchical clustering algorithm was used to cluster the water distribution network into several smaller clusters to manage the computational costs and ensure the scalability of the proposed algorithm. The proposed method is demonstrated on the problem of a benchmark municipal water distribution problem.

Keywords