Modelling minimum night consumption as a function of long-run average outflow to users and applications to real world case studies

As accurate monitoring and mitigation of water leakage are essential in the context of smart and resilient cities, this paper proposes a novel methodology for modelling users’ minimum night consumption (MNC), a key parameter in the estimation of district leakage levels. Unlike traditional models, which relate MNC to the number of users supplied by the network district, a new model with a bounded and more interpretable coefficient is conceived to introduce and evaluate the dependence of MNC on the yearly average user outflow, which can be evaluated based on long-run aggregate billed consumption. A methodology is proposed to characterize MNC starting from smart metering data of hourly user water consumption, which are progressively aggregated and analysed at the hour of minimum consumption for the district. Applications to three case studies in Italy, each with distinct characteristics, demonstrate that the novel methodology is beneficial. In fact, the results show that the linear model expressing MNC as a function of the yearly average user outflow fits data better than the traditional model, at increasing aggregation levels.