Modeling of drinking water distribution networks using deterministic and stochastic approaches

Abstract

Home water consumption is one of the most difficult parameters to estimate while modeling drinking water distribution networks. It has been shown that this parameter follows a stochastic process, and that it is possible to characterize it with rectangular pulses—having a certain intensity, duration, and adherence to an arrival frequency—using stochastic models, such as the Neyman-Scott Rectangular Pulses Model (NSRPM). This model is based on the solution of a nonlinear optimization problem involving theoretical moments that represent the synthetic (equiprobable) series, and observed moments (in-field measurements) that statistically establish the measured series. The methodology and the initial validation of the process applied to home consumption have been published; however its validation in an actual distribution network has not been made, nor has it been compared with conventional methods.This paper compares the results obtained by using simulations of synthetic series which are stochastic in nature, product of the NSRPM model applied to the determination of pressures, flows, and leaks, with those obtained with the traditional method that uses demand time-variation curves and with pressure and flow measurements made in the Humaya sector, in Culiacán, Sinaloa, Mexico.