مريم نصيري دهج

Leakage causes water loss, pressure drop, increase in productivity costs, and entry of pollutants in water distribution networks. Therefore, leakage detection and action to fix it is one of the main challenges of water and wastewater utilities. Since common leak detection methods are expensive and time-consuming and require expert personnel, in recent years, the use of field-simulation methods for detecting leaks in networks has been of special interest. In this research, leak detection is done using the calibration-optimization method, and Optimization is done by meta-heuristic algorithms. Considering that the more powerful the algorithm used, the less the error resulting from leak detection, it is necessary to introduce and apply the best algorithm in the field of leak detection. For this purpose, seven meta-heuristic algorithms were investigated in three water distribution networks in 45 leakage scenarios. After finding the best algorithm, areas prone to leakage are introduced by using k-means and two stage method. Also, in this study, the issue of the uncertainty of the base demands was also studied and included in the calculations. The reason for introducing a leak-prone area is that 1) the nodes that are introduced leaky are approximate, and 2) if we consider a real network, leaks can exist at scattered nodes. Therefore, it is necessary first to identify and investigate the limited areas where larger leaks have occurred. And 3) the existence of uncertainties makes it impossible to accurately guess the location of the leaks, and the found nodes should only Know the approximate location of real leaks. Therefore, it is necessary to identify the area that is prone to leakage. The two proposed methods were eva‎luated and validated by defining 20 leakage scenarios. The results indicate the superiority of the two stage method in finding the leak-prone area compared to the k-means method. In 17 scenarios, the two-stage method has a successful performance and the areas introduced by this method contain an average of 67% of the network leaks.

شبكه هاي توزيع آب , واسنجي , نشت يابي , بهينه يابي , الگوريتم هاي فرا ابتكاري

Water distribution networks , Calibration , Leak detection , Optimization , Metaheuristic algorithms

Maryam Nasiri Dahaj

Dr. Ebrahim Jabbari /Dr. Mohammadreza Jalili Ghazizadeh