Prediction of the maximum demand in a DMA of the water distribution network of Guanajuato City, Mexico

Abstract

This article presents a model to simulate and predict the maximum weekly water demand of a DMA belonging to the distribution network of Guanajuato city. The model is made up of two sub-models, one deterministic and one stochastic, which can be calibrated with historical demand data. The Fourier series describes the deterministic sub-model, while the stochastic sub-model involves the quantile function of the log-logistic distribution. Data from a holistic telemetry system were used to calibrate and validate the model to test the model's goodness in describing the demand behavior. Such a system consists of a network of stations for the measurement and recording of data and a system for storing and processing the information.