LeakG3PD: A Python Generator and Simulated Water Distribution System Dataset

Abstract

The United Nations’ (UN) 2030 Agenda for Sustainable Development 6th goal is of special interest here: “Ensure availability and sustainable management of water and sanitation for all”. To manage consumption water, it’s surely fundamental to be able to measure, diagnose and control the whole path from it’s origin in natural reservoirs till the end customers, requiring a net of sensors along the Water Distribution System (WDS). Such a network of sensors, will generate data over time constituting a dataset on which computer intelligence algorithms can work to detect abnormal behaviour, especially leakages. Unfortunately in Brazil the initiatives in monitoring WDS through sensors are very incipient. Since we don’t have a real dataset with sensor information over time yet, in this work we aim to provide a solution to this part of the problem through simulation. In 2018, LeakDB [1] was presented as a “benchmark dataset for leakage diagnosis in water distribution systems”, in fact a very important and interesting work. However it has some drawbacks, especially inconsistent data, which prevents its use as a solid basis. In this paper we improve LeakDB implementing LeakG3PD, opening the path to analyze the viability of different solutions in water management, with special focus on leakages.