USO DE INTELIGÊNCIA ARTIFICIAL PARA PREVER O CONSUMO DE ÁGUA À CURTO PRAZO DE UMA REGIÃO RESIDENCIAL DA CIDADE DE PATO BRANCO-PR

Abstract

One of the most critical factors in the operation of the water supply system is to maintain a supply of water equal to or larger than the population's consumption. The water must be supplied constantly to the population being served by the water supply system. By considering that this consumption is variable throughout the day and that every water supply system is subject to operational problems that can directly impact supply, forecasting water consumption can help to improve the operation of the supply system. This work aims to use artificial neural networks (ANN), in particular, the Radial Basis Function (RBF) network to predict short-term water consumption. The forecasting of the consumption was made based on historical consumption data of an area of a city of the Parana state, Pato Branco. The area was predominantly residential with medium consumption. The results of this study were satisfactory and efficient to describe the consumption forecast model.