Classification machine learning to detect de facto reuse and cyanobacteria at a drinking water intake.

Harmful algal blooms (HABs) or higher levels of de facto water reuse (DFR) can increase the levels of certain contaminants at drinking water intakes. Therefore, the goal of this study was to use multi-class supervised machine learning (SML) classification with data collected from six online instruments measuring fourteen total water quality parameters to detect cyanobacteria (corresponding to approximately 950 cells/mL, 2900 cells/mL, and 8600 cells/mL) or DFR (0.5, 1 and 2 % of wastewater effluent) events in the raw water entering an intake. Among 56 screened models from the caret package in R, four (mda, LogitBoost, bagFDAGCV, and xgbTree) were selected for optimization. mda had the greatest testing set accuracy, 98.09 %, after optimization with 7 false alerts. Some of the most important water parameters for the different models were phycocyanin-like fluorescence, UVA254, and pH. SML could detect algae blending events (estimated <9000 cells/mL) due in part to the phycocyanin-like fluorescence sensor. UVA254 helped identify higher concentrations of DFR. These results show that multi-class SML classification could be used at drinking water intakes in conjunction with online instrumentation to detect and differentiate HABs and DFR events. This could be used to create alert systems for the water utilities at the intake, rather than the finished water, so any adjustment to the treatment process could be implemented.

Quantitative Microbial Risk Assessment Framework Incorporating Water Ages with Legionella pneumophila Growth Rates.

Water age in drinking water systems is often used as a proxy for water quality but is rarely used as a direct input in assessing microbial risk. This study directly linked water ages in a premise plumbing system to concentrations of Legionella pneumophila via a growth model. In turn, the L. pneumophila concentrations were used for a quantitative microbial risk assessment to calculate the associated probabilities of infection (Pinf) and clinically severe illness (Pcsi) due to showering. Risk reductions achieved by purging devices, which reduce water age, were also quantified. The median annual Pinf exceeded the commonly used 1 in 10,000 (10-4) risk benchmark in all scenarios, but the median annual Pcsi was always 1-3 orders of magnitude below 10-4. The median annual Pcsi was lower in homes with two occupants (4.7 × 10-7) than with one occupant (7.5 × 10-7) due to more frequent use of water fixtures, which reduced water ages. The median annual Pcsi for homes with one occupant was reduced by 39-43% with scheduled purging 1-2 times per day. Smart purging devices, which purge only after a certain period of nonuse, maintained these lower annual Pcsi values while reducing additional water consumption by 45-62%.

Effects of temperature on nitrifying membrane-aerated biofilms: An experimental and modeling study.

Temperature is known to have an important effect on the morphology and removal fluxes of conventional, co-diffusional biofilms. However, much less is known about the effects of temperature on membrane-aerated biofilm reactors (MABRs). Experiments and modeling were used to determine the effects of temperature on the removal fluxes, biofilm thickness and morphology, and biofilm microbial community structure of nitrifying MABRs. Steady state tests were carried out at 10 °C and 30 °C. MABRs grown at 30 °C had higher ammonium removal fluxes (5.5 ± 0.9 g-N/m2/day at 20 mgN/L) than those grown at 10 °C (3.4 ± 0.2 g-N/m2/day at 20 mgN/L). The 30 °C biofilms were thinner and rougher, with a lower protein to polysaccharides ratio (PN/PS) in their extracellular polymeric substance (EPS) matrix and greater amounts of biofilm detachment. Based on fluorescent in-situ hybridization (FISH), there was a higher relative abundance of nitrifying bacteria at 30 °C than at 10 °C, and the ratio of AOB to total nitrifiers (AOB + NOB) was higher at 30 °C (95.1 ± 2.3%) than at 10 °C (77.2 ± 8.6 %). Anammox bacteria were more abundant at 30 °C (16.6 ± 3.7 %) than at 10 °C (6.5 ± 2.4 %). Modeling suggested that higher temperatures increase ammonium oxidation fluxes when the biofilm is limited by ammonium. However, fluxes decrease when oxygen becomes limited, i.e., when the bulk ammonium concentrations are high, due to decreased oxygen solubility. Consistent with the experimental results, the model predicted that the percentage of AOB to total nitrifiers at 30 °C was higher than at 10 °C. To investigate the effects of temperature on biofilm diffusivity and O2 solubility, without longer-term changes in the microbial community, MABR biofilms were grown to steady state at 20 °C, then the temperature changed to 10 °C or 30 °C overnight. Higher ammonium oxidation fluxes were obtained at higher temperatures: 1.91 ± 0.24 g-N/m2/day at 10 °C and 3.19 ± 0.40 g-N/m2/day at 30 °C. Overall, this work provides detailed insights into the effect of temperature on nitrifying MABRs, which can be used to better understand MABR behavior and manage MABR reactors.

Impact of fixture purging on water age and excess water usage, considering stochastic water demands.

Higher water ages are linked with water quality decline as chlorine dissipates, temperatures become more favorable for microbial growth, and metals and organic matter leach from the pipes. Water fixtures with automated purging devices can limit water age in premise plumbing systems, but also increase water use. To develop purging strategies that lower age while also minimizing water use, the stochastic nature of water demands must be considered. In this research, a hydraulic plumbing network model, with stochastic demands at fixtures, was used to compare water age and water use for five purging conditions: purging at regular intervals, "smart" purging (considering the time of last use), purging with different volumes of water, purging at different fixtures, and the purging with different levels of home occupancy. Higher purging frequency and volume resulted in lower water ages, but higher water use. Purging greatly reduced the variability in water ages, avoiding extreme ages entirely. Water age was minimized by scheduling the purging around occupancy behavior, such as before the occupants wake up or return from work. Scheduled purging used more water than smart purging. Purging after 12 h of nonuse used only 55% of the additional water required for purging every 12 h. Purging after 24 h of nonuse at the kitchen tap and shower used only 38% of the additional water required for purging every 24 h, while maintaining lower water ages and removing the variability in water ages. While larger purging volumes had a greater impact on water age, there were diminishing returns. Purging has a larger impact on low-occupancy homes because fixtures have less frequent use. Overall, this research provides a methodology to compare purging strategies that minimize both water age and water use. While the numerical results presented here are only valid for the specific layout and usage habits, they provide insights and trends applicable to other cases.

Effect of membrane type on the behavior of nitrifying membrane aerated biofilms: silicone membranes vs. micromembrane cords.

There is increasing interest in membrane-aerated biofilm reactors (MABRs), due to their energy efficiency and ability to intensify wastewater treatment. While MABR membranes play a key role, supporting biofilms and transferring O2, little research has addressed how membrane types impact MABR performance. This research compared two types of membranes used in commercial MABRs: a silicone hollow-fibre membrane and a 'micromembrane cord,' consisting of an inert cord surrounded by fine proprietary polymeric membranes. We used single-membrane MABRs to determine the oxygen mass transfer coefficient, Km, and explore biofilm development. The silicone membrane had a measured Km of 2.6 m/d, and the micromembrane cord had an apparent Km of 1 m/d. Pure MABR bundles (only biofilm) were operated with synthetic wastewater, and hybrid MABRs (suspended biomass and biofilm) with real wastewater, to explore behaviour for a wide range of conditions. The maximum ammonium oxidation fluxes with synthetic wastewater were 7.8 gN/m2d for the silicone membrane and 4.3 gN/m2d for the micromembrane cord. However, at bulk NH4+ concentrations below 5 mgN/L, the ammonium oxidation fluxes were similar. A previously published MABR model effectively captured the behaviour of each membrane. Nitrification fluxes with real wastewater were lower than synthetic wastewater, likely because of the presence of chemical oxygen demand (COD). Although the ammonium oxidation fluxes were higher for the silicone membranes for a given air supply pressure, the fluxes for the micromembrane cord could be increased using higher intramembrane air pressures. Overall, this research helped understand the impact of membrane types on nitrification fluxes.