Quality zones automatically identified in water distribution networks by applying data clustering methods to conductivity measurements

This paper presents a clustering study showing how conductivity measured every five minutes by 215 probes over four years can be used to determine specific quality zones for a large Water Distribution Network (WDN): 8500 km of pipes, 4.6 M customers. Conductivity time-series are compared using Dynamic Time Warping. Then, probes are ordered using a density-based method, and probe clusters are extracted automatically. The clusters are a sound representation of water quality in the WDN, both in terms of water origin and water residence time. More specifically, zones directly impacted by plants or by external water imports, mixing zones and zones influenced by tanks, can be isolated and analyzed. Globally, 82% of the probes were found to be clustered, consistent with expert knowledge on the WDN operation; 13% were unclassified; 3% were erroneously clustered; and 1% seemed to be reasonably clustered, without any physical understanding yet. Besides providing users with an increased understanding of water quality in WDNs, conductivity-based clusters offer an interesting prior tool for contamination warning systems.

Mixture of Joint Nonhomogeneous Markov Chains to Cluster and Model Water Consumption Behavior Sequences

The emergence of smart meters has fostered the collection of massive data that support a better understanding of consumer behaviors and better management of water resources and networks. The main focus of this article is to analyze consumption behavior over time; thus, we first identify the main weekly consumption patterns. This approach allows each meter to be represented by a categorical series, where each category corresponds to a weekly consumption behavior. By considering the resulting consumption behavior sequences, we propose a new methodology based on a mixture of nonhomogeneous Markov models to cluster these categorical time series. Using this method, the meters are described by the Markovian dynamics of their cluster. The latent variable that controls cluster membership is estimated alongside the parameters of the Markov model using a novel classification expectation maximization algorithm. A specific entropy measure is formulated to evaluate the quality of the estimated partition by considering the joint Markovian dynamics. The proposed clustering model can also be used to predict future consumption behaviors within each cluster. Numerical experiments using real water consumption data provided by a water utility in France and gathered over 19 months are conducted to evaluate the performance of the proposed approach in terms of both clustering and prediction. The results demonstrate the effectiveness of the proposed method.

Enrichment of free-living amoebae in biofilms developed at upper water levels in drinking water storage towers: An inter- and intra-seasonal study

Free-living amoebae (FLA) are ubiquitous organisms present in various natural and artificial environments, such as drinking water storage towers (DWST). Some FLA, such as Acanthamoeba sp., Naegleria fowleri, and Balamuthia mandrillaris, can cause severe infections at ocular or cerebral level in addition to being potential reservoirs of other pathogens. In this work, the abundance and diversity of FLA was evaluated in two sampling campaigns: one performed over five seasons in three DWST at three different levels (surface, middle and bottom) in water and biofilm using microscopy and PCR, and one based on the kinetics analysis in phase contrast and confocal microscopy of biofilm samples collected every two weeks during a 3-month period at the surface and at the bottom of a DWST. In the seasonal study, the FLA were detected in each DWST water in densities of ~20 to 25amoebaeL^-1. A seasonal variation of amoeba distribution was observed in water samples, with maximal densities in summer at ~30amoebaeL^-1 and minimal densities in winter at ~16amoebaeL^-1. The FLA belonging to the genus Acanthamoeba were detected in two spring sampling campaigns, suggesting a possible seasonal appearance of this potentially pathogenic amoeba. Interestingly, a 1 log increase of amoebae density was observed in biofilm samples collected at the surface of all DWST compared to the middle and the bottom where FLA were at 0.1-0.2amoebae/cm². In the kinetics study, an increase of amoebae density, total cell density, and biofilm thickness was observed as a function of time at the surface of the DWST, but not at the bottom. To our knowledge, this study describes for the first time a marked higher FLA density in biofilms collected at upper water levels in DWST, constituting a potential source of pathogenic micro-organisms.

Pilot-scale evaluation of UV reactors' efficacy against in vitro infectivity of Cryptosporidium parvum oocysts

An experimental protocol was developed to assess the efficacy of two UV reactors (medium-pressure UVaster), and a low-pressure reactor) on the infectivity of Cryptosporidium parvum oocysts under conditions mimicking small- or medium-size water distribution units. The protocol included purification of large amounts of viable oocysts from experimentally infected calf feces, pilot spiking, sample concentration and purification after UV radiation, oocyst quantification and in vitro evaluation of oocyst infectivity on HCT-8 cells. Water samples were collected at intervals upstream and downstream from the UV reactor after spiking. Oocysts were concentrated by centrifugation, purified by immunomagnetic capture and quantified using laser-scanning cytometry. An enhanced in vitro infectivity test on HCT-8 cells was developed, where oocysts were pretreated in order to obtain maximized in vitro infectivity, and infectious foci were enumerated after immunofluorescence staining after 3 days of culture. This method was superior to viability measured by excystation for assessing oocyst infectivity. The infectivity rate of untreated oocysts ranged between 9% and 30% in replicate experiments. The method allowed us to determine inactivation rates >4.92 (log) with UVaster and >4.82 with the LP reactor after exposition of oocysts to an effective dose of 400 J m(-2) at flow rates of 15 and 42 m(3) h(-1), respectively.

Detection of Escherichia coli in biofilms from pipe samples and coupons in drinking water distribution networks

Fluorescence in situ hybridization (FISH) was used for direct detection of Escherichia coli on pipe surfaces and coupons in drinking water distribution networks. Old cast iron main pipes were removed from water distribution networks in France, England, Portugal, and Latvia, and E. coli was analyzed in the biofilm. In addition, 44 flat coupons made of cast iron, polyvinyl chloride, or stainless steel were placed into and continuously exposed to water on 15 locations of 6 distribution networks in France and Latvia and examined after 1 to 6 months exposure to the drinking water. In order to increase the signal intensity, a peptide nucleic acid (PNA) 15-mer probe was used in the FISH screening for the presence or absence of E. coli on the surface of pipes and coupons, thus reducing occasional problems of autofluorescence and low fluorescence of the labeled bacteria. For comparison, cells were removed from the surfaces and examined with culture-based or enzymatic (detection of beta-d-glucuronidase) methods. An additional verification was made by using PCR. Culture method indicated presence of E. coli in one of five pipes, whereas all pipes were positive with the FISH methods. E. coli was detected in 56% of the coupons using PNA FISH, but no E. coli was detected using culture or enzymatic methods. PCR analyses confirmed the presence of E. coli in samples that were negative according to culture-based and enzymatic methods. The viability of E. coli cells in the samples was demonstrated by the cell elongation after resuscitation in low-nutrient medium supplemented with pipemidic acid, suggesting that the cells were present in an active but nonculturable state, unable to grow on agar media. E. coli contributed to ca. 0.001 to 0.1% of the total bacterial number in the samples. The presence and number of E. coli did not correlate with any of physical and/or chemical characteristic of the drinking water (e.g., temperature, chlorine, or biodegradable organic matter concentration). We show here that E. coli is present in the biofilms of drinking water networks in Europe. Some of the cells are metabolically active but are often not detected due to limitations of traditionally used culture-based methods, indicating that biofilm should be considered as a reservoir that must be investigated further in order to evaluate the risk for human health.

Investigation of by-product formation during the irradiation of drinking water with a medium pressure lamp

The goal of this study was to determine the effect of UV irradiation with a polychromatic spectrum on natural organic matter. Several irradiation tests were carried out with or without cut-off of wavelengths lower than 240 nm on water samples coming from different drinking water plants. DOC, BDOC, SEC analyses, chlorine demand, nitrate and nitrite concentration measurements were made. Changes were noticed as regarding SEC and chlorine demand analyses. Indeed a consistent trend of breaking-up NOM molecules into smaller fragments was observed. Moreover, the chlorine demand increased with the dose when the cut-off filter was not applied, whereas a maximum value resulted when the cut-off filter was applied. Most of these results were obtained at high UV doses (40000 J x m(-2)), suggesting that UV irradiation would not have a noticeable effect on the water samples tested at doses usually used for drinking water treatment.

Coliforms and other microbial indicators occurrence in water and biofilm in full-scale distribution systems

Biofilm and microbial water quality were studied in four middle size full-scale distribution systems (DS) in France serving 5,000-30,000 inhabitants (maximum residence time 23-160h) through three sampling campaigns over 1 year. Three of these DSs were chosen because of a quite high occurrence of bacterial indicators (i.e. total coliforms), the last DS was considered as a reference. Biofilm was studied on cast iron coupons incubated for more than 1 month in devices continuously fed with water from the DS in conditions imitating those met in DS. The devices were located at different points (4-6) along each DS. The abundance of bacteria in biofilm was estimated by heterotrophic plate counts (HPC) after detachment of the biofilm from the support by sonication. Microbiological water quality was estimated in parallel; analysis of total coliforms, E. coli, enterococci and anaerobic sulphide-reducing bacteria spores (ASRB spores) was carried out in biofilm and water. Over the period of the study, 171 water samples and 57 biofilm samples were collected. Over these 171 waters, 19 (11%) were positive for at least one of the measured indicators while two biofilm samples were positive (3.5%). Significant differences were observed in the levels of contamination between the DSs. High residence time in the DS, low disinfectant residual and high temperature increased the risk of indicator occurrence in the water phase. Due to the low number of biofilm samples positive for bacterial indicators, the data collected in the present study did not allow observation of a direct association between biofilm and water contaminations, even if the occurrence of indicators in water appeared on DSs with the highest density of biofilm (HPC).

Use of hydrogen peroxide in scrubbing towers for odor removal in wastewater treatment plants

The aim of this work was to replace sodium hypochlorite (NaCIO) with hydrogen peroxide (H202) in chemical scrubbing towers, in order to avoid the formation of chlorinated species, harmful for human health. Some previous studies have already shown the ability of H2O2 to treat the hydrogen sulfide (H2S) pollution. However, an important decomposition of the oxidant was observed in the scrubbing solution (carbonates, transition metal and high pH are responsible for this decomposition) leading to high reactant consumption. Consequently, this study first focused on research into a compound able to reduce the hydrogen peroxide degradation. Experiments were conducted on a pilot unit (3,000 m3 h(-1)) in a wastewater treatment plant. The sodium silicate (Na2SiO3) proved to be a good scrubbing solution stabilizer. A very good removal of hydrogen sulfide (up to 98%) was also obtained. Finally, the study resulted in the determination of the best operating conditions to achieve both an efficient and economical process.