Modelo preditivo de fugas no sistema de distribuição municipal de água de Vila Nova de Gaia

Abstract

An efficient management of water resources is one of the main society's challenges. Therefore, it is essential that entities arm themselves with tools that support decision-making in order to reduce water losses on the transportation moment to supply populations. This study proposes to analyse Vila Nova de Gaia's water distribution network, on a geographic point of view, and the occurred leaks in the years of 2017 and 2019. From the collected data and analysis, it was developed a model with machine learning algorithms and GIS tools to predict future leaks. The used algorithm was random forest regressor and it allowed to bound the most susceptible areas for the occurrence of water leaks in this municipality.