Optimization of disinfectant dosage for simultaneous control of lead and disinfection-byproducts in water distribution networks

Drinking water utilities are increasingly facing the challenges of maintaining water quality, and simultaneously complying with conflicting regulatory standards. One such challenge is the dosage of chlorine-based disinfectants which is typically regulated to prevent microbial growth in the water distribution systems, while limiting disinfection by-products (DBPs). On the other hand, chlorine residuals also influence the dissolution of lead into drinking water from corrosion scales in the pipe internals, as has been shown by previous studies. Hence, it is important to consider water lead levels (WLLs) while determining the appropriate chlorine dosage. This study proposes a multi-objective optimization framework to understand and balance the trade-offs between (i) minimizing total disinfectant dosage to reduce DBPs formation potential, and (ii) maximizing the volume of safe drinking water supply with respect to both WLLs and residual free chlorine concentrations. The approach comprises of the development of a dynamic simulation-optimization framework to account for the impact of spatial and temporal variations of network hydraulics and water chemistry on WLLs and residual free chlorine. The framework couples dynamic multi-species water quality simulations (pH, residual free chlorine, dissolved inorganic carbon, and orthophosphate concentration) with a multi-objective genetic optimization algorithm in an integrated MATLAB-EPANET-MSX platform. The application of the optimization formulation is demonstrated on a real-world distribution network case study. The resulting optimal solution set on pareto-plots are discussed for sensitivity of the trade-offs between the two objective functions, under various water chemistry conditions that have been suggested in earlier studies for minimizing lead release from plattnerite (PbO2) corrosion scales in lead service lines. The resulting optimal disinfectant dosage schedule, for pumping and booster station nodes in each case, provided insights on maintaining disinfectant residuals throughout the distribution system so as to prevent microbial growth as well as lead contamination events while limiting DBPs formation. Further, environmental implications related to the use of the proposed framework are also discussed.