| Resumo: | Nowadays, the major expenses with water supply systems (WSS) correspond to energy consumption. In WSS, the pumps are only activated when the reservoirs, responsible for supplying certain populations, reach their minimum levels. The introduction of a pump pattern adapted to energy prices variation and consumption patterns of populations can minimize energy costs significantly. In this paper, two different examples of water supply networks simulations are introduced. For these WSS the application of different optimization methods that minimize the costs associated to energy consumption in water pumping are presented. The selected optimization methods were the method of Levenberg-Marquardt (LM) and an evolutionary algorithm. In both simulated examples, the optimization of pump pattern allows significant reductions in costs associated to the pumping (up to 71%). Reductions of energy consumption are also observed in one of the examples. It was found that the optimized pump patterns take into account the variation of the energy cost throughout the day. The classic method of LM proved to be the most efficient in this kind of optimization problems. The sequential use of both methods allows further reductions with the drawback of large CPU time.
|
|---|