Full-scale water distribution networks modelling for disinfection by-products estimation

(English) The quality of the water that reaches our homes depends on multiple factors: the origin of the water, the treatment it undergoes, and the distribution network that delivers it. To ensure water safety not only at the treatment plant’s outlet but also at the points of consumption, chemical d...

ver descrição completa

Detalhes bibliográficos
Autor: Vinardell Magre, Laura
Formato: tesis doctoral
Fecha de publicación:2026
País:España
Recursos:Universitat Politècnica de Catalunya (UPC)
Repositorio:UPCommons. Portal del coneixement obert de la UPC
Idioma:catalán
OAI Identifier:oai:upcommons.upc.edu:2117/457573
Acesso em linha:https://hdl.handle.net/2117/457573
https://dx.doi.org/10.5821/dissertation-2117-457573
Access Level:acceso abierto
Palavra-chave:disinfection by-products
trihalomethanes
haloacetic acids
chlorate
EPANET
hydraulic simulations
modelling
68 - Indústries oficis i comerç d'articles acabats. Tecnologia cibernètica i automàtica
628 - Enginyeria sanitària. Aigua. Sanejament. Enginyeria de la il·luminació
546 - Química inorgànica
Àrees temàtiques de la UPC::Informàtica
Àrees temàtiques de la UPC::Desenvolupament humà i sostenible
Àrees temàtiques de la UPC::Enginyeria química
Descrição
Resumo:(English) The quality of the water that reaches our homes depends on multiple factors: the origin of the water, the treatment it undergoes, and the distribution network that delivers it. To ensure water safety not only at the treatment plant’s outlet but also at the points of consumption, chemical disinfectants are often added. Chlorine is one of the most widely used disinfectants worldwide due to its low cost, versatility, and its ability to maintain residual disinfectant power throughout the network. However, despite these advantages, chlorine can form potentially harmful by-products under certain conditions. For this reason, the latest European Union directive (2023) mandates the regulation of several disinfection by-products: trihalomethanes (THMs), haloacetic acids (HAAs), chlorate, chlorite, and bromate. This thesis presents the development of several models to estimate the concentration of by-products resulting from the use of hypochlorite as a disinfectant. Two case studies have been examined: one high-pressure distribution network and another with both high and low-pressure distribution. Over the course of a year, multiple sampling campaigns were carried out to collect sufficient data to calibrate and validate the estimation models. Prior to this, hydraulic models were adjusted to provide input for the water quality models. The study focuses on three types of disinfection by-products regulated by European standards: the THM family, the HAA family, and chlorate. It was observed that the speciation of THMs and HAAs is strongly influenced by the water source. Moreover, the evolution of the concentration of these compounds varies depending on network operation and the number of rechlorination stations. For chlorate, its formation was studied in relation to the use and storage of concentrated hypochlorite, both in primary disinfection and in rechlorination stations. Two modelling approaches were applied for THMs and HAAs: mechanistic models and data-driven models. The mechanistic models are based on the relationship between compound formation and chlorine decay, whereas data-driven models identify patterns among variables without explicitly relying on the underlying physical or chemical processes. In the latter case, specific models were developed for each case study, as well as a general model applicable to both networks, despite their differences in size, water origin, and operation. All models developed for THMs and HAAs were applied both to individual compounds and to the regulated totals (TTHMs and THAAs). The study assessed which strategy yields better estimations: summing individual models or directly modelling the total concentrations. In this study, direct estimation of totals gave slightly better results. However, modelling individual compounds can provide added value, both from a risk management perspective (since each compound may have a different impact) and for a deeper understanding of their formation mechanisms. Finally, for chlorate, a mechanistic model was developed based on the known kinetics of hypochlorite degradation. The formation of chlorate in concentrated hypochlorite storage tanks was integrated into a hydraulic model simulating the behavior of the water tank receiving chlorination, enabling estimation of the chlorate concentration at the outlet. The model provided accurate predictions and can be applied to other distribution networks.