Calculating Chlorine Concentration Bounds in Water Distribution Networks: A Backtracking Uncertainty Bounding Approach

Abstract

The most recent EU directive on the quality of water intended for human consumption (European Commission, 2018) emphasizes the importance of water-quality monitoring for reducing the risks to health from drinking water, with chlorine by-products being among the quality parameters recommended for monitoring. Chlorine is commonly used by water utilities as a disinfectant in Water Distribution Networks (WDN). According to the World Health Organization (WHO), a chlorine residual needs to be sustained throughout the drinking water network which should be sufficient to deactivate waterborne pathogens and, at the same time, small enough to reduce the formation of harmful chlorine by-products (Mouly et al., 2010). Additionally, due to the fact that certain contaminants will affect chlorine residuals in a specific way (e.g., a bacterial toxin may decrease the concentration of free chlorine due to its reaction dynamics) (Helbling & Van-Briesen, 2009), chlorine residuals can be used as key indicators of contamination events (Hall et al., 2007).The installation of online chlorine sensors can enhance monitoring capabilities of chlorine residuals and enable more timely control actions. For contamination event detection, water-quality sensors may be installed which can monitor multiple parameters (e.g., pH, free chlorine, total organic carbon) and detect changes in water-quality. The placement of such sensors in WDN in order to detect a contamination event has been a subject of research for years (Eliades & Polycarpou, 2010;Hart & Murray, 2010). The problem is posed as a multiobjective optimization, with one of the objectives being the minimization of the number of sensors due to their high capital and maintenance costs (Zeng et al., 2016), while addressing the uncertainty present in these real-world systems is also a crucial component of these methodologies (Sankary & Ostfeld, 2018). Single-parameter sensors (such as free chlorine concentration sensors) are cheaper than multiparameter sensors, and using them not only for chlorine residual monitoring but also for contamination event detection, could significantly reduce the health risks from contaminated drinking water. However, in order to determine whether a sensor reading is abnormal (e.g., due to a contamination), it needs to be compared with an estimate of the expected concentration at the sensor location. The estimated concentration can be calculated using water-quality models (Rossman & Boulos, 1996). These are mathematical representations

Contamination event diagnosis in drinking water networks: A review

“…The task of estimating the concentration of these substances throughout the network, in real-time, is challenging due to the highly uncertain available models and the small number of measurements (Vrachimis et al, 2021). Mathematical models that estimate disinfectant concentration in the water at different parts of the network are typically complex and inaccurate, especially in large-scale networks.…”

“…The researchers go a step further by proposing an interval observer structure for chlorine residual estimation. In Vrachimis et al (2021) a methodology to calculate chlorine concentration bounds at nodes of WDS is proposed, using an input-output modeling approach, hydraulicstate estimates as input and incorporates water-quality measurements in a parameter estimation scheme. Feedback control of disinfectant concentration states aims to keep them within pre-defined levels and reduce disinfection by-products health impacts in the long term (e.g., in the case of Trihalomethanes).…”

La importancia de la temperatura del agua en las redes de abastecimiento

“…Esta combinación será suficiente para explicar la complejidad del proceso de decaimiento y los potenciales efectos de la temperatura, y es una asunción habitual en la literatura (e.g. Vasconcelos et al, 1997;Monteiro et al, 2014, Vrachimis et al, 2021.…”

Leakage Management Influence on Water Age of Water Distribution Networks

“…The change of daily water age due to leakage reduction is assessed using standard tools developed during the last century, mainly based on EPANET (Rossman, 2000), sometimes with some modifications and improvements. The need of understanding the water quality behavior throughout WDNs motivated some recent works to support design and planning for WDN management (e.g., Farmani et al., 2006; Quintiliani et al., 2019; Vrachimis et al., 2021). The EPANET (Rossman, 2000) water quality analysis toolkit has been used by several authors for various purposes and some contributions were proposed to improve it over time.…”