Nutrient conditions that contribute to bacterial regrowth in drinking water distribution systems were investigated. The major limiting nutrient for the experimental program was biodegradable organic matter (BOM), which is a surrogate for the amount of available organic carbon. A suite of organic compounds consisting of carboxylic acids. aldehydes, and amino acids was spiked into an existing drinking water source. The spiked tap water was pumped into an annular reactor (AR), which was used to simulate conditions of a section of a distribution system. Bulle (suspended) cell numbers and biofilm cell numbers, as measured by heterotrophic plate counts (HPCs) and total direct cell counts, were used as response variables. A method for removing biofilm cells from a polycarbonate coupon, which is flush-mounted to the side of an AR, was developed as part of the research.

The major experimental factors were:​ BOM concentration, the mass fraction of amino acids, and hydraulic retention time. It was found that high BOM loading rates (i.e. high BOM concentrations and low retention times) significantly increased the number of biofilm HPCs. The presence of amino acids also increased the number of biofilm HPCs. The net bulk HPCs were found to increase at longer retention times. A high BOM concentration significantly decreased the net bulk HPCs. The amount of amino acids did not affect the net bulk HPCs. ln general, total cell counts were less responsive to the three experimental factors than HPC numbers.

Based on the influent organic components dosed to the ARs, a composite BOM was calculated ( on a mass carbon basis) for each experimental trial. It was found that external and internal mass transfer to and within the biofilm did not limit BOM removal. Based on eight experimental trials at various BO M loading rates, a first-order rate expression for the removal of BOM was calculated. Pseudo-first-order rate constants for each BOM component were also calculated. The rate constants provide an indication of the relative rate of biodegradability. From this analysis it was shown that formate has the highest rate of biodegradability and glyoxal has the lowest.

The per cent removal of the added amino acids (i.e., aspartic acid, glutamic acid, and serine) was essentially complete. High or near complete removals for formate and formaldehyde were also calculated. Acetate and glyoxal had much lower fractions removed.

Additional experiments were performed whereby three ARs were connected in series and free chlorine was dosed into the first AR. It was found that chlorine was able to mitigate bulk and biofilm numbers in the first AR, however, chlorine had less of an effect on HPC numbers in the subsequent ARs. BOM components were removed by greater than 80 % after the first AR regardless of the influent residual chlorine concentration (i.e. either 0.0 or 0.2 mg/L). Refractory organic matter, as measured by UV absorbance at 285 nm, was found to increase through the three ARs.

An approach for relating bench-scale experiments to full-scale pipes was developed. The critical parameter for this relationship is the ratio between the specific surface area of the bench-scale reactor (e.g. annular reactor) and the specific surface area of the pipe diameter. Based on this approach an AR having a I h retention time would approximately correspond to a 7 h retention time in a 100 mm (4 in.) pipe.