Project IC18CT97-0136 (INCO-DC)

This project aimed to identify situations in continuous and intermittent water supplies in which regrowth of bacteria occurred, and to develop predictive models of these systems to evaluate various control scenarios to limit the regrowth problem. Continuous distribution systems in the UK and Portugal, and intermittent systems in Jordan, Lebanon and Palestine have been intensively sampled for physical, chemical, bacteriological and hydraulic parameters over a period of 1-2 years. The selected systems have been simulated using the public-domain Epanet model, and the output files have been interfaced with a new event-driven model (Enhanced Performance Q), developed by our Portuguese partners, which incorporates prediction of bacterial growth and death, in suspension or as biofilm.

The coefficients of growth and death of distribution system isolates and other organisms known to be present in distribution, have been determined during the project, and the kinetics of chlorine decay, both in the bulk water and due to wall (biofilm) effects have been evaluated so that appropriate values can be used in the bacterial model.

Biofilm development has been studied in actual distribution systems and in experimental rigs, where biofilm interactions with chlorine and effects of drying and/or velocity on bacterial detachment have been measured.

The continuous distribution system in the temperate UK climate was shown to suffer bacterial regrowth problems especially where river-derived supplies were subject to long retention times (long mains, service reservoirs), low chlorine concentrations and relatively warm (above 15^oC) summer temperatures.

The intermittent water supplies of the Mediterranean countries demonstrated several scenarios which led to poor bacterial water quality. Household water storage tanks are an essential component of systems subject to intermittent supplies, and our studies showed that the bacterial quality of the water deteriorates, with increases of up to 5 orders of magnitude of bacteria, when water is stored for up to seven days at temperatures of 20^oC and above.

A second potentially serious situation has been identified when water pumping restarts after a period of no flow in an intermittent supply. When the pipes are in an unpressurised state, infiltration of contaminated water is possible; partially-dried biofilm has also been shown to detach as water velocities increase. Both events lead to poor bacterial water quality in the first flush, as water pumping starts.

Throughout these studies bacterial isolates have been made, organisms identified using a variety of techniques, and these results will be presented.

In addition to Enhanced Performance Q, a simple plug flow model has been developed to simulate bacterial growth, death and detachment, chlorine decay and other chemical changes in water quality, and this has been validated both for the distribution system data and for the household storage tank data. Both models are being used to predict regrowth in a variety of situations and to evaluate alternative control scenarios, e.g. most appropriate locations in the systems for rechlorination, and what chorine concentrations will be effective.

Comparison of the two models indicates that the plug flow models work well both in continuous and intermittent supplies, but that Enhanced Performance Q is most useful in the continuous supply situation.