Impact of hydraulic behavior and modeling uncertainty on the optimization of water quality degradation monitoring in distribution networks

The optimization of water quality monitoring in distribution networks is challenging due to the spatial and temporal variability of contaminants and hydraulic behavior, which changes with demand. Hydraulic models include uncertainties related to parameters and input variables. This study developed and compared scenarios to examine the impact of hydraulic behavior and modeling uncertainty on the optimal location of monitoring points. Scenarios included different hydraulic time scales (intra-day, daily, seasonal), modeling parameters, and variable uncertainties (demand, consumption patterns, population served). A sub-network of the Quebec City distribution system was used as case study. Optimization was performed for free residual chlorine (FRC) and trihalomethanes (THMs), targeting global representativeness and at-risk population monitoring. Results show that optimization is most influenced by seasonal variability, particularly for FRC, requiring a differential seasonal strategy. Under seasonal demand variability, optimal FRC monitoring nodes may increase by up to 300% in summer compared to winter. Optimization was also sensitive to the hydraulic time step, especially for identifying at-risk population locations. For THMs, although variations in population at risk (≤73%) and monitoring requirements (≤25%) were observed, optimal node locations remained unchanged, requiring fewer adjustments. Model uncertainties had low impact on hydraulic behavior, allowing accurate identification of representative monitoring nodes.