Biostability in drinking water distribution systems study at pilot-scale

Biostability and related issues (e.g. nitrification) were investigated for 18 months in 18 pilot distribution systems, under various water quality scenarios. This study specifically investigated the impact of steady-state water changes on HPC levels in chlorinated and chloraminated distribution systems. Chlorination was more effective than chloramination in reducing HPC levels (1--2 log difference). There was a rapid increase in HPC corresponding to the change in steady-state water quality, which was observed in all PDS. Modeling effort demonstrated that HPC levels reached a maximum within five days after water quality change and return to initial level ten days after the change. Since alkalinity was used as a tracer of the steady-state water quality change, time to reach maximum HPC was related to a mixing model using alkalinity as a surrogate that confirmed alkalinity transition was complete in approximately eight days.; Biostability was assessed by HPC levels, since no coliform were ever detected. It was observed that HPC levels would be above four logs if residual droped below 0.1--0.2 mg/L as Cl2, which is below the regulatory minimum of 0.6 mg/L as Cl2. Therefore bacterial proliferation is more likely to be controlled in distribution systems as long as residual regulatory requirements are met. An empirical modeling effort showed that residual, pipe material and temperature were the most important parameters in controlling HPC levels in distribution systems, residual being the only parameter that can be practically used by utilities to control biological stability in their distribution systems. Use of less reactive (i.e. with less chlorine demand) pipes is recommended in order to prevent residual depletion and subsequent bacterial proliferation.; This study is investigated biofilm growth simultaneously with suspended growth under a wide range of water quality scenarios and pipe materials. It was found that increasing the degree of treatment led to reduction of biofilm density, except for reverse osmosis treated groundwater, which exerted the highest biofilm density of all waters. Biofilm densities on corrodible, highly reactive materials (e.g. unlined cast iron and galvanized steel) were significantly greater than on PVC and lined cast iron. Biofilm modeling showed that attached bacteria were most affected by temperature and much less by HRT, bulk HPC and residual. The model predicts biofilms will always be active for environments common to drinking water distribution systems. As American utilities do not control biofilms with extensive and costly AOC reduction, American utilities must maintain a strong residual to maintain biological integrity and stability in drinking water distribution systems.; Nitrite and nitrate were considered the most suitable indicators for utilities to predict onset of a nitrification episode in the distribution system bulk liquid. DO and ammonia were correlated to production of nitrite and nitrate and therefore could be related to nitrification. However since ammonia and DO consumptions can be caused by other phenomena than nitrification (e.g. oxidation by disinfectant to nitrite and reduction at the pipe wall, respectively), these parameters are not considered indicators of nitrification. (Abstract shortened by UMI.)...