Resumo:
The developed research on water distribution networks are of great relevance,
considering the social, environmental and economic impacts occurred due to the scarcity of
hydric resources. Therefore, all the cientific efforts denoted on the researches that study the
distribution water network are of great relevance. Techniques as mathematical modeling,
computer simulation and statistic methods are widely used to obtain more reliable answers for
the actual situation of the network and also for the prediction of scenarios, failure events,
increased demand, etc. The main objective of this research is to carry out a bibliometric and
systematic analysis that contemplates researches that use statistic models for the control and
reduction of water volume losses in the distribution system, identifying which were the models
applied, resulting the actual scenarios, which will be as guide for the futures works to be
structured in the most relevant research’s that studies the theme. The developed methodology
was able to analyze a metadata composed of 4150 documents from the base of Web of Science.
As result was obtained a geographical view of the subject pointing out the major countries,
affiliations, journals and researchers, as well as pointed out the main documents, the relevance
of the theme and the statistical models applied. It is concluded that the most applied statistical
models are regression, survival analysis, Monte Carlo technique and Poisson technique
respectively. Eighteen other techniques were used, but in smaller quantities, which do not
render them ineffective, water distribution networks are complex mechanisms and sometimes
require different forms of approach that require other models of statistical analysis.