Key points are not available for this paper at this time.
The essential role of water and the need to protect this valuable resource are well recognised worldwide. Leakage in water distribution networks (WDNs) has serious economic, social, and environmental consequences. Because conventional leak detection methods face several technical and practical limitations, recent studies have increasingly focused on field-simulation-based methods. Among these, hydraulic model calibration has proven to be one of the most effective approaches. However, many existing studies depend on idealised or fully defined hydraulic models, which restrict their applicability to actual WDNs. This study presents a calibration-based leak detection method that employs a modified simulated annealing (SA) algorithm to identify both the location and magnitude of leaks in a WDN. To represent model uncertainty more realistically, two approaches – Monte Carlo simulation (MCS) and simultaneous calibration (SC) – were incorporated into the proposed framework. The method was first tested on a benchmark WDN, where the results confirmed its ability to detect leaks effectively. The SC approach demonstrated higher accuracy and computational efficiency compared with MCS. The methodology was also applied to an actual WDN, where it successfully located a deliberately introduced leak. The findings indicate that the proposed method offers a practical and reliable tool for water utilities aiming to enhance leak detection and improve WDN management.
Moasheri et al. (Sun,) studied this question.