A novel hybrid entropy-clustering approach for optimal placement of pressure sensors for leakage detection in water distribution systems under uncertainty

This study presents a novel hybrid entropy-clustering framework for placing pressure sensors in water distribution systems (WDS) to detect leakage. Leakages are simulated at all potential nodes of WDS, and then potential pressure sensors (PPS) in WDS are classified using a K-means clustering algorithm. Transinformation entropy for each potential pair of PPS was also computed, which in turn helped to reduce redundant information. PPS locations were subsequently optimized using a multi-objective optimization model. Furthermore, to capture the sensitivity of sensors' layout in WDS to sensor error, a fuzzy-based analysis is integrated with a multi-objective optimization model. Finally, the best compromise solution of PPS placement in each category was selected using an ELECTRE multi-criteria decision making model. Reducing redundant information of pressure sensors based on information theory and choosing the best possible solution based on the ELECTRE model are the main novelties of this study. Results of C-Town WDS attest to the proposed framework' efficiency.