JWSS - Journal of Water and Soil Science

The water supply networks have always been of significance to researchers as a hydraulic system of transferring and distributing water. The pressure gradient is the main reason of water transfer in networks, and in case of non-standard pressure increase, the undesirable phenomenon of leakage occurs in the network. Leakage in urban water distribution networks causes water waste and enormous financial losses. Therefore, there exists the need to manage and minimize the amount of leakage. In this study, a water distribution network is modeled using the potentialities of hydraulic analysis model, the EPANET 2.0, and, by presenting a new model, the leakage location is recognized. In order to do this, we, firstly, entered all the network parameters into the software. Then, the network was analyzed supposing the non-occurrence of leakage and the amounts of nodal pressures were measured. Moreover, the nodal pressures were estimated by creating a hypothetical leakage in one of the network nodes and analyzing the network. Finally, the position of leakage was determined by defining the leakage index and comparing it in various nodes. The results show that the suggested method is efficiently capable of predicting the leakage position in the network so it can safely replace other methods, especially destructive methods used in recognizing the leakage position in the network.

A water distribution network (WDN) may not be able to satisfy all required demands when it’s in the pressure deficit mode or under over-loaded demand conditions. Analysis of the network in this mode requires pressure dependent analysis (PDA). Unlike demand driven analysis (DDA), PDA needs an extra equation for every node to relate the nodal demand and the nodal pressure; so it should be solved with the other network’s equations simultaneously. In this paper, based on the Particle Swarm Optimization (PSO) algorithm, a decision support system has been developed by using MATLAB and EPANET for PDA simulation in WDNs. A four-loop network selected from the literature was analyzed using different scenarios and different pressure dependent functions presented by the previous investigations. The results showed that the proposed model (PSO-PDA) was as accurate as the previous ones and provided better convergence. The results of the nodes’ pressure and discharge also indicated minor differences obtained by different PDA functions. However, the differences between the results of PDA and DDA were considerable.