Analysis of Water Pressure Control With Machine Learning in Sectorized Distribution Network

This paper addresses the problem of water pressure control in distribution networks. Lack of adequate control can lead to unnecessarily high pressures, which in turn can cause breakages, leaks and significant water losses, as well as deteriorate and damage equipment such as pipes and flow regulators in the supply network. The objective of this research is to analyze the behavior of water pressure in a distribution network, using an experimental laboratory prototype controlled by a Proportional Integral Derivative (PID) system. The analysis is carried out using a machine learning algorithm based on a regression model, which allows identifying patterns of variation in pressure within the network. The data used for this study were generated by the experimental prototype for one week. A descriptive statistical analysis of the data was performed, using central tendency measures, and the linear regression algorithm was applied to recognize behavior patterns. The results obtained offer a detailed analysis of the PID control applied to water pressure, contributing to the evaluation of its effectiveness in a sectorized water network. The use of machine learning algorithms allowed a better understanding of the relationship between the water pressure variable and the PID control implemented in the distribution network.