Robustness of profile sampling in detecting dissolved lead in houshold drinking water

Abstract

[EN] The norm for dissolved lead in drinking water is lowered in the Netherlands per 2023, from 10 to 5 µg/L. The effect is that: (a) homeowners and the Dutch water utilities want to find problematic lead components and remove them; and (b) the utilities want to know if copper service mains with lead soldering are expected to lead a norm exceedance. This calls for an improved sampling protocol (collected at the tap) to trace dissolved lead. We help solve this problem using hydraulic and water quality simulations in EPANET. With short time steps, low flow rates, and short pipe lengths, a water quality time step of 1 second is not short enough. As EPANET cannot work with timesteps smaller than 1 second, we devised a work-around to ensure that the numerical solution does not result in a large error. A household is defined as a system of pipes with various locations where water is used. The water demand patterns at each tap are generated using SIMDEUM. A few (sections of the) pipes are assigned to contain lead wherefrom dissolution occurs. The benchmark scenario involves a two-person household with average Dutch water consumption and with solely the service line containing lead. Next, fourteen scenarios are considered wherein either the geometry of the system, the water consumption, or location/extent of the lead releasing surface(s) are varied individually. Of the sampling protocols considered, the results show that “profile sampling” is the most promising. In such a protocol, twenty consecutive samples of 300 ml are collected following a minimum of six hours of prolonged stagnation. Irrespective of the scenario (with a lead component), at least one of the twenty samples is guaranteed to possess a high concentration of lead, owing to our choice of a fine sample volume. Just two days of profile sampling can potentially unveil: (1) the location of the lead releasing component, (2) the volume of the lead releasing region, and (3) the saturation concentration of lead dissolution. The key conclusions (for the current assumptions) are as follows. A lead service line potentially leads to an exceedance of the new norm, necessitating immediate action. The results also provide evidence that one week of proportional sampling is insufficient. Presently, it is considered to be the gold standard for measuring weekly intake of dissolved lead, however, it is susceptible to the stochastic nature of water demand leading to variations exceeding 50%. Lastly, a profile sampling protocol possesses the best opportunity to detect dissolved lead in the household drinking water network in a robust manner. In the future, a bespoke experimental facility (“PilotCity indoor installation”) will be used to validate the assumptions surrounding the lead dissolution and advection/diffusion model.