Aplicação de métodos de otimização para operação de redes de distribuição de água: uma revisão sistemática da literatura

Abstract

In last time research of to optimize water supply systems has long been found in the literature, whether related to the design, operation, calibration or testing of networks. In view of the growing population increase, based on the fact that water distribution systems involve energy, in addition to the finite natural resource, optimizing the processes related to the operation of the networks arises with the need to implement techniques that seek, among other objectives, to minimize costs and losses, control pressure inside the pipes, which has the purpose of guaranteeing water to the consumer in sufficient quantity and quality. The present study was carried out using the StArt software, which helps in structuring a systematic review, in order to gather relevant works from 1977 to 2020. The collection of 87 articles started from a search that originated 637 results, being 335 from the Web Of Science database, 152 works from the Scopus database, 100 results from the Science Direct database, and 49 articles from Google Academic, 1 classic article was manually added to the StArt system. The research followed a search methodology. This result originated from the pre-establishment of search keywords and the application of search filters. The funneling of the results began with the application of criteria for inclusion and exclusion of articles, in order to select the one most targeted at the theme of optimizing the operation of water supply networks. Subsequently, a quality analysis was performed, where each article was scored based on 10 pre-established quality criteria. Data related to the applied optimization method were extracted, number of objective functions (single or multi-objective), number of restrictions used, number of decision variables used, application or not of penalty function; type of system (pumping or configuration of valves), type of network studied (real or theoretical) and hydraulic simulator. The results indicate that among the evaluated articles, genetic algorithms (simple, hybrid and multiobjective) are the most employed optimization method, and most of these studies use 5 or more restrictions and decision variables. The application of penalty and non-application functions were quite balanced between the jobs, corresponding to 49% and 51% respectively. The type of system studied, the pumping systems represent the vast majority, as well as the type of theoretical network and the hydraulic simulator EPANET.