Calibration and optimal leakage management for a real water distribution network

Abstract

The aim of the paper is to present two optimization methodologies applied to a real water distribution system. The first one involves model calibration, while the second addresses the problem of optimal leakage management. Model calibration is based on a single objective, real-coded genetic algorithm, which provides optimal values of pipe friction factors and a leakage-dependent coefficient. Leakage management is accomplished through the introduction and regulation of pressure reducing valves. To this end, the writers adopted an approach based on multiobjective genetic algorithms, searching for the optimal placement and regulation of such valves. In particular, the first criterion is represented by the minimization of the number of valves, while the second is the minimization of the total leakage in the system. The algorithm provided several trade-off alternatives between leakage reduction and installation costs: the water utility managing the system chose to place and regulate the first four valves identified by the algorithm, resulting in a water loss reduction (as evidenced by the continuous monitoring of the system after valve installation) and in a decrease in pumping costs.