Optimal Leakage Detection Model of Water Distribution Systems Using Semi-Pressure Driven Analysis and Harmony Search

During maintenance and managing of water distribution systems, reducing water losses is a critical component. To achieve reduction of water losses, active leakage management is required. In this study, we propose optimal leakage detection model, which is the combination of simulation of hydraulic analysis and optimization technique. Through the use of hydraulic analysis method can able to achieve in creating a real-life based Semi-Pressure Driven Analysis (Semi-PDA) model without having to use the highly uncertain Head Outflow Relationship (HOR) is applied. EPANET2 created a model of emitter coefficient`s optimization, and it decides leakage location. For optimization technique, Harmony Search (HS) is applied that was developed in Korea and has been used in various cases recently. When we apply proposed method to sample network, it practically searches the leakage location. Especially, in optimization technique, we have a research of comparing the results of Harmony-Search (HS) and Genetic-Algorithm (GA); HS` result is more accurate. Also we verify that increasing of leakage location detection uncertainty by decreasing of number of observed data is less than GA. Therefore, suggested technique can be substituted to existing experiential leakage detection method. Therefore, this basic technology can be used as future proactive water resource management environment like Smart Water Grid.