Leak Detection and Localization Based on Search Space Reduction and Hydraulic Modelling

Reducing the cost and time required to isolate leakages occurring in Water Distribution Networks (WDNs) is the main task for resilient and sustainable management of these systems. The paper presents a systematic model-based leak detection and localization framework using optimization. The approach prerequisites a well calibrated WDN hydraulic model. The leakage localization model splits into two stages: (a) the search reduction stage where the number of decision variables and the range of possible values are reduced, and (b) the leak detection and localization stage for isolating the fault. The leakage localization method is formulated to optimize the leakage node locations and their associated emitter coefficients, such that the differences between the model predicted and the field observed values for pressure and flow are minimized. The optimization problem is solved by using a non sorting genetic algorithm. A real case from a UK system is presented with the outcome showing that the method reduces the leak search space within 10% of the WDN, while contributing to earlier leakage hotspot detection and localization. The framework for predicting leakage hotspots can be effective despite the recognized challenges of model calibration and the physical measurement limitations from the collected pressure and flow data.