D. Gatel

Effect of adding phosphate to drinking water on bacterial growth in slightly and highly corroded pipes

The effect of phosphate addition in drinking water was tested under static conditions as batch tests and under dynamic conditions using continuously fed reactors. Phosphate supplements in batch tests from 0.1 to 2 mg P-PO4 L(-1) did not show any relationship between bacterial growth and phosphate concentration. Dynamic tests in slightly corroded reactor (stainless steel) treated at 1 mg P-PO4 L(-1) showed only a moderate improvement in the growth of microorganisms. On the contrary, phosphate treatment applied to the highly corroded reactor (unlined cast iron) led to an immediate, drastic drop in iron oxide release and bacterial production. Phosphate uptake by the reactor wall was less than 14% with the stainless-steel reactor and 70-90% with the corroded cast iron reactor. Moreover, about 5% of the phosphate associated to corroded iron pipe walls was released for 20 days after the end of treatment.

Fate of Escherichia coli experimentally injected in a drinking water distribution pilot system

Abstract Detection of coliforms and E. coli in drinking water distribution systems can be explained by accidental contaminations and/or growth of these bacteria in the network. This last point still remaining debatable, an experiment was carried out with two E. coli strains ( E. coli O126:B16 and one isolated from a drinking water distribution system) separately injected in a drinking water distribution pilot plant continuously fed with drinking water. The work aimed to study (i) the partition of the injected bacteria between the water phase and the indigenous biofilm phase and (ii) the kinetic of disappearance or growth of these bacteria. Depending on the strain, 1–50% of injected bacteria adsorbed within a few hours to the indigenous bacterial biofilm. During the first 5–7 days, both E. coli populations behaved similarly: the total number of injected bacteria decreased more quickly than the theoretical washout (death, lysis, predation, ...). Then the number of these bacteria slowly increased and at day 9–12, the number of recoverable injected bacteria became higher than the predictable number calculated from the theoretical washout. This result clearly proves that both E. coli strains are able to grow at 20°C in the absence of residual chlorine in a distribution network system largely colonized with an autochthonous population. However, colonization of the network by E. coli strains was only partial and transient.

Microbial characteristics of a distribution system fed with nanofiltered drinking water

Abstract This work was designed to study the transition from water treated with O 3 and granular activated carbon to nanofiltered water with special regard to biofilm sloughing and turbidity, and to measure the extent of biofouling in an “old” network fed with non-chlorinated nanofiltered water. The experimental drinking water distribution system contained six loops in series, with a total hydraulic residence time (HRT) of 72 h. It was first fed with biologically treated water (GAC water) having organic carbon (1.6 mg-C litre −1 , of which 22% was biodegradable) and microorganisms (1.12 × 10 5 cells ml −1 , of which 7% was culturable after 15 days of incubation). It was then fed with nanofiltered water, which was contaminated with bacteria in transit from the nanofiltration unit to the network. This nanofiltered water 1 contained 0.23 mg-C litre −1 (biodegradable fraction −1 ) and 3.0 × 10 4 cells ml −1 (22% was culturable after 15 days). Finally, the system was fed with nanofiltered water 2 that contained few microorganisms (50 cells ml −1 ) and few nutrients (0.21 mg-C litre −1 , biodegradable fraction −1 ). The changes in the water and biofilm quality in the experimental network fed successively with the three waters, and the water quality during transport through the network, were determined. The bacterial population in the system showed a gradual drop in the total number of cells and their culturability with the water changes. The total number of cells and the colony forming units (CFU) after 3–15 days decreased respectively about 10-fold and 100-fold after 1 year of feeding with nanofiltered water 2 compared to GAC water to reach 1.3 × 10 4 cells ml −1 , 2 CFU ml −1 (after 3 days of incubation at 22°C) and 121 CFU ml −1 (after 15 days of incubation at 22°C) in the first loop of the experimental network (HRT 12 h). No massive biofilm sloughing was observed. The quality of nanofiltered water 2 changed during its transport through the network (total HRT of the water 72 h) with immediate bacterial contamination. But the total number of cells increased slowly and reached 1.4 × 10 5 cells ml −1 at the end of the system. The number of culturable bacteria decreased slowly and regularly as a function of the HRT to reach 8 CFU (3 days) ml −1 and 260 CFU (15 days) ml −1 at the end of the network.

Dynamic modeling of bacteria in a pilot drinking-water distribution system

Computer modeling can be a useful tool in understanding the dynamics of bacterial population growth. Yet, the variability and complexity of biological systems pose unique challenges in model building and adjustment. Recent tools from Bayesian statistical inference can be brought together to solve these problems. As an example, the authors modeled the development of biofilm in an industrial pilot drinking-water network. The relationship between chlorine disinfectant, organic carbon, and bacteria concentrations was described by differential equations. Using a Bayesian approach, they derived statistical distributions for the model parameters, on the basis of experimental data. The model was found to adequately fit both prior biological information and the data, particularly at chlorine concentrations between 0.1 and 2 mg/liter. Bacteria were found to have different characteristics in the different parts of the network. The model was used to analyze the effects of various scenarios of water quality at the inlet of the network. The biofilm appears to be very resistant to chlorine and confers a large inertia to the system. Free bacteria are efficiently inactivated by chlorine, particularly at low concentrations of dissolved organic carbon.

Relationship between coliform culturability and organic matter in low nutritive waters.

The objective of this work was to investigate the behaviour of coliform bacteria in specific low nutritive waters conveying organic fractions from different origin of which an unknown part is likely to pass through the treatment barrier. For this purpose, we studied the growth (microscopic counting) and the culturability (count on nutritive medium) of ten coliform bacteria species as a function of the amount of organic matter in a river water collected after a period of heavy rain and in an algal bloom water. Assays were carried out in the presence of autochthonous heterotrophic bacteria from the Nancy (France) drinking water, with variable concentrations of dissolved organic carbon (DOC) representative of drinking waters (0.5-1.5 mg l(-1) for diluted river water samples and 1.3-2.5 mg l(-1) for diluted algal bloom water samples). Bacterial growth was measured in the two types of water, regardless of the initial concentration of DOC. We found that coliform bacteria lost their culturability in both sample series, and that the lower the initial DOC concentration the more rapidly the culturability was lost. The quantity of DOC consumed by the bacteria in the two water types (0.03-0.13 mg l(-1) in river water and 0.77-1.29 mg l(-1) in algal bloom water) and the resulting consequences on bacterial behaviour suggested that bloom water contains algal organic compounds that are antagonistic to the growth and/or the culturability of coliform bacteria. Organic matter thresholds, beyond which coliform bacteria are unlikely to keep their culturability, have not been determined experimentally. Indeed, at the end of the assays some culturable coliform bacteria were systematically detected in both types of water. Enterobacter cloacae was the predominant species. Thus, during these adverse events the probability of coliform occurrence can be considered as high in treated water.

Release of organic matter in a discontinuously chlorinated drinking water network

The effects of discontinuous chlorination on the characteristics of the water in a pilot drinking water distribution network were investigated. The release or consumption of organic matter (as dissolved organic carbon, DOC) following chlorination and non-chlorination periods were estimated, as were changes in bacterial cell production. In each unchlorinated network 0.3 mg DOCl(-1) was consumed and the average cell production was approximately 1.3 x 10(5) cells ml(-1). In discontinously chlorinated networks (chlorine treatment: 3.3 mg Cl2l(-1), chlorine residual: 0.1 mg Cl2l(-1)) the DOC release (DOCout-DOCin) was between 0.1 and 0.2 mg Cl(-1). Biomass production (cells(out)-cells(in)) during this chlorination period was lower (approximately 2 x 10(4) cells ml(-1)). The delay before DOC was released in chlorinated networks appeared to be less than 24 h, which corresponds to one hydraulic residence time. Likewise, when chlorination was stopped, 24 h or less were required before an efficient DOC removal was resumed. When chlorination was prolonged the observed release of DOC was progressively reduced from 0.2 mg l(-1) to zero, thus after 6 weeks of continuous chlorination the DOCin was equivalent to the DOCout.

Monitoring and modeling of disinfection by-products (DBPs)

In the United States, the newly promulgateddisinfectant/disinfection by-product (D/DBP) regulationsforce water treatment utilities to be more concerned withfinished and distributed water qualities. In this study,monitoring of DBP formation was conducted from three Frenchwater treatment plants trying to assess DBP variationsthrough time and space. Compared to the in-plant totaltrihalomethanes (TTHM) levels, TTHM levels in thedistribution system increased from less than 150% to morethan 300%. Significant variations for TTHM and bromate(BrO3-) levels throughout the seasons were alsoobserved; generally higher levels in the summer and lowerlevels in the winter. Combining chemical DBP models(empirical power functional models) and hydraulicsimulations, DBPs including TTHM and BrO3- weresuccessfully simulated from the full-scale monitoring data,indicating that empirical DBP model can be a potential toolto access DBP formation in actual plants. This study alsoprovides the protocols to assess DBP simulations in thewater treatment systems.

Comparative Assessment of Bromate Control Options

Abstract Various bromate control options were assessed through an intensive pilot testing program, performed with a four-cell bubble contactor, and focused on intermediate ozonation of conventionally-treated sand-filtered water. Both acid addition and ammonia addition independently provided good bromate reduction, with their combinational addition providing no further reduction. On a CT basis, the use of a static mixer did not increase bromate formation, while staged ozonation enhanced bromate formation over single stage application.

Controlling the corrosiveness of blended waters

Abstract Corrosion can occur in pipelines, especially in the consumers plumbing. Corrosion products can result in dirty and rusty water, leading to customer complaints. It has also been suggested that corrosion favours bacterial proliferation due to rapid chlorine decay and coliform regrowth. Finally, corrosive water can lead to excessive lead and copper concentrations. Corrosion problems can result from mixing two different waters, often the case when using new water sources, and consequently corrosion control strategies are a concern for most utilities. One of the possibilities is to add soda or lime, although the latter has an undesirable impact on water hardness. The suburbs of Paris Water Utility is faced with the situation of having to blend nanofiltered, thus decarbonated water (37 mgd), with conventionally treated hard-surface water (53 mgd). A 1-year research project has been conducted with the following objectives: (1) selection of a model to calculate the suitable soda dosage; (2) determination of contact time; (3) testing the performances of the concept at pilot scale, especially during changes in flow regimes and water blending proportions; and (4) document possible interaction with the chlorine dosage. Experiments were conducted in the Mery-sur-Oise WTP from January through December 1996. The two different waters used were nanofiltered water with a calcium concentration of 60 mg/L-pH 6.5, and a “standard” water with a calcium concentration of 120 mg/L, pH 7.5. The water temperatures vary from 1 to 25°C. The soda dosage is calculated using the Hallopeau-Dubin equations. These equations are used in a software package developed for automatic control systems. Experiments show that to obtain a stable pH, the necessary contact time can be as high as 7 min in cold-water conditions. In addition, it is also necessary to anticipate the hydraulic residence time. As a result, the software had to be modified to ensure that the saturation index remains positive in all situations. On-line records indicate that this goal was achieved on the pilot unit during 6 months of successful operation. Other experiments have shown that there might also be an interaction with the chlorine dosage, resulting in potential pH underestimation.

Comparison of three antiscalants, as applied to the treatment of water from the River Oise

Abstract Nanofiltration, used in the Mery-sur-Oise water treatment plant (90 mgd) provides a softened water with a low organic matter content from which micropollutants and microorganisms have been removed. The experience acquired on the site since 1992 has revealed that the use of a crystal formation inhibitor is essential to slow down the clogging of the nanofiltration membranes, essentially on the third stage. Given the complexity of the action of the antiscalants (threshold effect, metal ion sequestration capacity, particle dispersion capacity), it is difficult to theoretically define the type of product to be used and its optimal dosing level. In addition to laboratories or modelling studies, pilot experiments are needed to assess antiscalants and optimise their doses. Nanofiltration cycles being long (1–2 months), it is nevertheless difficult to compare results of cycles, and the objective of the project was to develop a methodology for sequestering agent testing. In order to carry out strictly comparative studies of commercially available products as well as studies using the same product but at several different dosage levels, a pilot unit comprising three parallel nanofiltration trains was build. The pilot unit is fed with 2nd stage concentrate, produced without sequestering agent, and each trains has its own separate antiscalant injection point. The first cycles aimed to provide answers to the following: (i) choice of the antiscalant best adapted to the treatment of water from the river Oise, (ii) identification of the clogging agents, (iii) the efficiency of the chemical regeneration procedure. On completion of comparative tests on three commercially available products, only one gave redhibitory water permeability and pressure drop results. The two other products, applied at the dose levels recommended by the manufacturers, gave results that were equivalent. Analysing the cleaning waters provided further information on the type of clogging agents deposited on the membrane. Although, the overall nature of the clogging is identical, essentially composed of organic matter and mineral elements (calcium, phosphorus, aluminium, etc.). The similitude of the water permeability and salt rejection at the beginning of two consecutive cycles tends to demonstrate that the chemical regeneration procedure (soda/citric acid) is efficient.