A framework for assessing uncertainty of drinking water quality in distribution networks with application to monochloramine decay

Abstract

To analyze changes in water quality conditions in water distribution systems (WDS), such as disinfectant byproduct formation, chlorine residual, and biofilm growth, water utilities can use coupled hydraulic and chemical models. Model-based operational decisions are complicated by uncertainties in model input parameters, which propagate through the model resulting in uncertain model predictions. Previous works have focused on conducting sensitivity analyses of input parameters in various water quality reactions and the interactions with hydraulic factors, but have overlooked the specific contribution of hydraulic and chemical parameter uncertainty. In this work, a framework for assessing the hydraulic and chemical uncertainty in water quality models is presented and demonstrated using monochloramine decay in WDSs supplied by gravity and pumping systems. In the first step, a sensitivity analysis is conducted to determine the influential and non-influential chemical parameters which govern the rate of monochloramine decay. In the second step, Monte Carlo simulations are used to explore the individual and combined effects of uncertainty in hydraulic and chemical parameters. Results show that uncertainty in modeled monochloramine concentration increases with water age and higher reaction rates, varies throughout the course of a day, and heavily depends on hydraulic variability, emphasizing the need to account for input uncertainty. The computational tool developed in this work can be extended to other reaction mechanisms, water quality parameters, and distribution systems for case-specific conditions, and used to evaluate system-wide effects of uncertainty on water quality.