Resumo:
The disorderly use of water resources leads them to depletion, resulting in the emergence of environmental and social issues. Thus, a relevant aspect to be observed in water supply systems are the losses that occur throughout the operating process of the system. However, the largest water losses are in the distribution networks, which lead water to homes, businesses and industries. This fact causes financial losses and mainly natural losses. The concern that the theme brings is relevant to the use of measures that allow its reduction. For this, the use of computer simulations of networks is intended to study their behavior, especially to identify areas prone to water loss. Thus, it was developed, in the present study, methodology of operation of water distribution network, using two methods: genetic algorithms and random search, aiming at the optimization of the system through the optimization of network pressures and thus the reduction of losses. The parameters of the theoretical study network were changed, such as: reservoir level, roughness, diameter, demand, pumps and pressure reducing valves simultaneously, using the tools R and EPANET. The loss reduction found with the simulation using the Genetic Algorithm was 15.75%. In the random simulation, the loss reduction found was 9.79%. In addition, the values of pressures found in the optimal generation, presented by the genetic algorithm, were much better than that of the random, because the high pressures that the network presented before the simulation were below 50 m.c.a. Thus, the simulation of the study network on the use of the genetic algorithm method was much better, with optimized loss and pressure values.