Experimental appraisal and numerical modelling of chlorine demand and decay in a typical drinking water distribution network in South Africa.

Safe drinking water requires sound monitoring and maintenance of residual chlorine within drinking water distribution networks (DWDNs) to suppress possible microbial regrowth. However, in the developing world, DWDNs face unique challenges, including aging infrastructure, water pipes laid near or even aboveground thus exposing water to high temperature fluctuations, and relatively high organic loads. Therefore, safely maintaining sustainable residual chlorine levels and restricting the problems of hazardous disinfection by-products (DBPs) formation and of antimicrobial resistance (AMR), both tracing back to extensive chlorination, is a difficult exercise in those settings. Here, the temperature dependent bulk chlorine decay, i.e., the rate at which chlorine residual is consumed, was estimated for a typical DWDN system in South Africa. To this end, experimental assays were performed and a mathematical model was developed to predict chlorine levels within the DWDN under study. A direct relationship (R = 0.99) between bulk chlorine decay and initial chlorine dosage was identified, with bulk chlorine decay following the first-order decay model. The bulk chlorine decay rate coefficient (Kb) and the reaction constant with the pipe walls (Kw) were experimentally estimated, with the first being the main chlorine consumer and the latter only slightly contributing to chlorine decay. EPANET was used to simulate the chlorine concentrations within the examined DWDN, while residual chlorine concentrations were modelled using COMSOL. The software programs were calibrated and validated using experimental results. The optimum liquid chlorine dosage was 5 mg L, and this could maintain residual levels at 0.5 mg L for 3500 min in the water distribution tanks. Yet, the residual chlorine levels at the distal end of the DWDN were below the recommended safety limits, suggesting the need for chlorine booster stations to supplement residual chlorine rather than further increasing chlorine initial dosages which will lead to unsafe chlorine levels at the proximal points and inevitably will increase DBPs in drinking water. This relatively high chlorine dosage reflects the overall poor quality of the raw water that feeds the drinking water treatment plant under study, which is consistent with the poor water quality of surface water in South Africa. Overall, this methodology can be replicated in DWDNs in South Africa and across the developing world, where similar challenges persist and ensure safe drinking water but not at the expense of DBPs formation and possibly AMR spreading.