141 Effect of Temperature on Membrane Aerated Biofilms
E. Clements*, P. Pérez-Calleja*, R. Nerenberg*
* University of Notre Dame, Department of Civil and Environmental Engineering and Earth Sciences 156 Fitzpatrick Hall, Notre Dame, Indiana 46556, USA
Keywords: MABR; Temperature Gradients; COD removal Summary of key findings
Temperature is an important variable in biological treatment processes, affecting biofilm processes in multiple ways. Key changes include the microbial kinetics, gas solubilities, and substrate diffusivities.
This research used modelling and experiments to assess the effects of temperature on membrane aerated biofilm reactors (MABRs), as compared to conventional, co-diffusional biofilms.
In MABRs, increasing the temperature has opposing effects: higher substrate utilization rates and substrate diffusivities on one hand, and lower O2 solubility on the other. This research explored the impact of temperature, biofilm thickness, bulk substrate concentrations and air supply pressure (for MABRs) on fluxes. Key insights include the relative importance of the temperature dependent parameters and a comparison of how conventional and MABR biofilms are affected by temperature.
Background and relevance
Temperature influences biofilm processes in multiple ways, including the microbial kinetics, oxygen solubility, and substrate diffusivity. While the effects of temperature are less significant for biofilms than for suspended growth, due to the greater role of diffusion in biofilm processes, it still can have important impacts (Zhu and Chen, 2002; Zhang et al., 2014; Wijffels et al., 1995).
The MABR is a wastewater treatment technology that can intensify treatment and significantly reduce energy requirements (Aybar et al., 2014). It is based on gas-transferring, hollow fiber membranes that deliver oxygen to a biofilm growing on the membrane surface. MABR biofilms are counter-
diffusional, where O2 comes from the base of the biofilm while the electron donor comes from the bulk liquid. Past research has shown that MABR biofilms display unique behaviour, distinct from conventional biofilms (Nerenberg 2016). However, the impact of temperature on MABRs has not been studied, and may also differ from co-diffusional biofilms.
An increase in temperature typically has opposing effects: (1) it increases microbial kinetics and substrate diffusivities, which increase substrate removal fluxes; and (2) it decreases the gas solubility, which results in lower O2 concentrations in MABR biofilms. The decrease in solubility has different effects on conventional and MABR biofilms. For conventional biofilms, the solubility does not affect contaminant removal fluxes if the bulk O2 concentration is maintained at a set point, say 2 mg/L. For MABR biofilms, in contrast, if the air supply pressure is maintained unchanged, it reduces the O2
concentration at membrane-biofilm interface, which typically decreases substrate fluxes.
This research compares the effects of temperature in conventional biofilm systems and MABRs. This work focuses on heterotrophic biofilms as a model system.
Results
Models for an MABR and conventional biofilm were developed in AQUASIM 2.1 (Reichert 1994).
The processes and stoichiometry were based on previous models (Shanahan and Semmens 2015; Ni et al. 2013). Preliminary bench scale experiments were performed to confirm trends shown by the model, and further experiments are ongoing.
To assess the individual impact of the microbial kinetics, diffusivities, and O2 solubility with temperature, the substrate removal fluxes for MABRs were compared with only one parameter changing at a time. The baseline condition was 20°C, a 500 µm biofilm with an air supply pressure of 6 psig, which had a substrate removal flux of 28.1 g COD/m2/day with a bulk COD concentration of
142
50 mg/L. Table 1 shows how changing any one parameter to its value at 10 or 30°C, while leaving the other parameters at their 20°C values, affects the substrate removal flux.
Table 1. Sensitivity analysis on the substrate removal fluxes in MABRs [g COD/m2/day] for a 500 µm biofilm 20°C with one parameter adjusted to its value at 10 or 30°C (bulk concentration=50 mg COD/L, air supply pressure=6 psig).
Temperature
[°C] µ
(Max spec. growth) H
(Henry’s constant) DO2
(O2 diffusivity) Ds (COD diffusivity) T effect
on µ µ effect
on J T effect
on H H effect
on J T effect on DO2
DO2 effect
on J T effect
on Ds DO2 effect on J
10 -49% -10% 23% 7% -26% -5% -26% -18%
30 97% 1% -17% -6% 30% 6% 30% 15%
The changing COD diffusivity had the largest impact on the substrate removal fluxes, though the temperature dependence of the diffusivity of oxygen, the solubility, and the specific growth rate also contribute to substantial changes in the flux. Therefore, when considering the impact of temperature on biofilm systems, none of these factors should be neglected.
For MABRs, a lower O2 concentration at the membrane-biofilm interface can greatly reduce the O2
penetration into the biofilm, reducing fluxes (Figure 1). For conventional systems at different temperatures, the aeration rate would likely be adjusted to maintain the bulk oxygen concentration.
Though this would require more energy, it would eliminate the negative effect of higher temperatures on O2 solubility. In an analogous fashion, MABRs could be operated with higher intramembrane oxygen supply pressures to maintain the same O2 concentration at the membrane-biofilm interface and thus increase fluxes.
Figure 1. Substrate profiles for 500 µm biofilms with 50 mg COD/L in the bulk liquid, at 5 °C and 30 °C for a conventional (left) and MABR (right) biofilm (LDL of 100 µm, 4 mg O2/L in the bulk liquid for the
conventional biofilm, 6 psig intramembrane pressure for the MABR)
For the conditions shown in Figure 1, the substrate removal flux for the MABR was 11.1 g
COD/m2/day at 5°C and 16.9 g COD/m2/day at 30°C. However, if the intramembrane air pressure was increased for the 30°C condition such that the O2 at the membrane surface was the same as it was at 5°C, the flux would increase from 16.9 g to 22.5 g COD/m2/day.
Conventional and MABR biofilms differ in the impact of biofilm thickness and bulk COD concentration on fluxes. In conventional systems, after the biofilm is no longer biomass limited, a thicker biofilm does not further increase fluxes. Similarly, the electron donor concentration does not increase fluxes once it is limited by O2. Since the activities are the same at different thicknesses and COD concentrations, the impact of temperature is the same as well (Figure 2). In MABRs, thick biofilms become mass transfer limited, so thicker biofilms have lower microbial activity. Since the fluxes can be higher in MABRs, and since COD has to diffuse deeper into the biofilm, the bulk COD concentration has a greater effect on fluxes. So, the biofilm thickness and bulk COD concentration determines how much of an impact different temperatures have (Figure 2). When the biofilm is limited by both oxygen and substrate, the temperature has less of an impact because there the microbial activity already is low.
143
Figure 2. Removal flux for conventional biofilms (left) and MABR biofilms (right) as a function of thickness for 10 and 30 °C. (LDL of 100 µm, 4 mg O2/L in the bulk liquid for the conventional biofilm, 6 psig intramembrane pressure for the MABR)
The effect of temperature gradients within the biofilm is also of interest. In a treatment plant, if hot air from the blowers is used to supply the membranes, for example, the temperature at the membrane side of the biofilm may be greater than the bulk liquid. Preliminary experiments, with and without
temperature gradients, were performed on a bench scale reactor and compared to the model (Table 2).
These also were simulated in the model by including a temperature gradient in the biofilm.
Table 2. Comparison of experimental data to the model predictions
Temperature at the membrane [°C]
Temperature of the bulk liquid [°C]
Oxygen
Supply Bulk S
concentration [mg COD/L]
Predicted Substrate Removal Flux [g/m2/day]
Experimental Substrate Removal Flux [g/m2/day]
11 11 1 psig in the
membrane 186 21.8 16.6
8 11 6.5 mg/L at
the membrane 143 12.7 16.8
36 11 6.5 mg/L at
the membrane 178 27.7 21.9
Though the model did not match the experimental results precisely, it captured the trends. When the temperature increased, even if the bulk water temperature remained constant, the removal rates for the substrate increased. With more experimental data, the model will be calibrated. Further study will be done on the impact of temperature gradients on COD removal fluxes and more experiments will be performed to further validate the model.
Discussion
This model explores how temperature affects MABRs and conventional biofilms. As temperature increases, the removal fluxes also increase, though the magnitude depends on a variety of factors including the biofilm thickness and the bulk COD concentration. These factors affect MABRs more than conventional biofilm systems because of the counter-diffusional geometry and the impact of oxygen solubility.
The decrease in oxygen solubility with increasing temperature reduces the removal rates. This has a larger impact on membrane aerated biofilms than suspended growth, because the dissolved oxygen for suspended growth would likely remain constant, though the increased bubbling rate would require more energy. However, this can be overcome in MABRs by increasing the air supply pressure. The increased substrates diffusivities increase the removal rates, as do the increased microbial rates.
Understanding the impact of temperature on biofilm processes is important for designing and
implementing biofilm processes. Future work will include more experimentation to validate the model
144
and studying how nitrifying biofilms are affected by temperature.
References
Songming Zhu and Shulin Chen. (2002). The impact of temperature on nitrification rate in fixed film biofilters. Aquacultural Engineering. 26, 4, 221-237.
Zhang S., Wang Y., He W., Wu M., Xing M., Yang J., Gao N., Pan M. (2014). Impacts of temperature and nitrifying community on nitrification kinetics in a moving-bed biofilm reactor treating polluted raw water. Chemical Engineering Journal. 236, 242-250.
Wijffels R.H, Englund G., Hunik J.H., Leenen E.J.T.M., Bakketun A., Gunther A., Obon de Castro J.M., Tramper J. (1995).
Effects of diffusion limitation on immobilized nitrifying microorganisms at low temperatures. Biotechnology Bioengineering.
45, 1-9.
Aybar, M., Pizarro, G., Boltz, J.P., Downing, L., Nerenberg, R., (2014). Energy-efficient wastewater treatment via the air- based, hybrid membrane biofilm reactor (hybrid MfBR). Water Science Technology 69, 1735–1741.
Nerenberg, R.. (2016). The membrane-biofilm reactor (MBfR) as a counter-diffusional biofilm process. Current Opinion in Biotechnology. 38, 131–136.
Reichert, P. (1994). Aquasim - a Tool for Simulation and Data Analysis of Aquatic Systems. Water Science and Technology;
London 30, 21–30.
Shanahan, J.W., Semmens, M.J. (2015). Alkalinity and pH effects on nitrification in a membrane aerated bioreactor: An experimental and model analysis. Water Research 74, 10–22.
Ni, B.-J., Joss, A., Yuan, Z., (2014). Modeling nitrogen removal with partial nitritation and anammox in one floc-based sequencing batch reactor. Water Research 67, 321–329.
Presenting Author
Emily Clements PhD Student
University of Notre Dame
Is the presenting author an IWA Young Water Professional? Yes Bio: Emily is a second year PhD student studying the effect temperature has on biofilm performance and properties.
145
MABR Enables Nitrification Below “Washout” SRT and Enhanced TIN removal
Reeve, M.*, Hacking, R.* Astrand, N.*, Mohammed, A.**, Heidary-Monfared, S**, Dr. Liu, Y.***
*SUEZ Water Technologies and Solutions, 3239 Dundas St W, Oakville, ON, Canada, L8P 1V9
**EPCOR Water Services Inc; 10977 50 St NW, Edmonton, AB T6A 2E9
***University of Alberta; 116 St & 85 Ave, Edmonton, AB T6G 2R3
Keywords: MABR; washout SRT; intensification, Summary of key findings
A pilot study at Gold Bar Wastewater Treatment Plant in Edmonton, Alberta investigated the intensification capability of membrane aerated biofilm reactor (MABR) technology on nitrification and denitrification of a biological nutrient removal process. An MABR equipped pilot train
demonstrated nitrification performance that exceeded both a control train performance and the performance that is predicted based on activated sludge design equations, when operated at low aerobic solids retention times (SRT). The MABR train also demonstrated effluent total inorganic nitrogen (TIN) concentrations that were 35% lower than the similarly operated control train. The results showed the ability of MABR to allow an activated sludge system to operate at lower aerobic SRT while achieving better TIN removal performance due to nitrification occurring in the biofilm and seeding of nitrifiers to the bulk mixed liquor.
Background and relevance
EPCOR Water Services Inc. (EPCOR) is evaluating membrane aerated biofilm reactor (MABR) technology for secondary treatment capacity expansion of the Gold Bar Wastewater Treatment Plant (Gold Bar) in Edmonton, Alberta. The plant is rated for 310 MLD average and 420 MLD maximum flow and is expected to increase with population growth of 1.5% annually. Gold Bar operates a biological nutrient removal (BNR) process with 11 trains in modified Johannesburg configuration consisting of pre-anoxic, anaerobic, anoxic, and aerobic zones, with respective volumes shown in Table 1.
Gold Bar is nearing capacity and is limited by the existing secondary biological tank volume and the solids handling capacity of the secondary clarifiers. There is little room at Gold Bar to expand
conventionally – adding additional treatment trains – due to the North Saskatchewan river to the north and Gold Bar park surrounding the balance. To treat additional load the plant must intensify the biological treatment capacity, for example, by operating at lower mixed liquor solids retention time (SRT) and relying on the seeding effect from an MABR to maintain nitrification.. The limits for ammonia are expected to become more stringent in the future and a limit for total nitrogen is expected to be added. As such, Gold Bar needs to maintain complete nitrification and maintain or improve total nitrogen removal while increasing capacity.
The objective of this study is to compare a hybrid MABR activated sludge (AS) process train with a standard BNR (conventional) train in a pilot plant designed to mimic the full-scale secondary process at Gold Bar. The specific objectives are to show that MABR 1) maintains complete nitrification at SRTs lower than the conventional train, 2) maintains complete or partial nitrification below the
“washout” SRT for nitrifiers, and 3) enables increased TN removal. This work is a demonstration of the “seeding” effect of MABR in AS processes discussed by Houweling et al. (2018). The seeding effect allows activated sludge processes to a) fully nitrify below the design SRT and b) partially nitrify
146
below the washout SRT due to a % removal of ammonia by the biofilm and observed yield of nitrifiers from the biofilm to the bulk MLSS. Ultimately, the results of this study are intended to be used to inform the full-scale design and to show that the MABR process can enable increased capacity at Gold Bar.
The pilot facility consists of two parallel bioreactors, each of which are roughly 1/600th the volume of a full-scale treatment train. The conventional train has the same modified Johannesburg configuration (pre-anoxic, anaerobic, anoxic, aerobic) as the full-scale plant and the MABR train is modified to accept ZeeLung MABR pilot scale cassettes and modules that were installed into the anoxic zone, as shown in Figure 1.1. The anoxic zone of the MABR train was extended to create a larger anoxic fraction and smaller aerobic fraction.
Results
The MABR train has shown better total inorganic nitrogen (TIN) removal compared to the conventional train as shown in Figure 1.2. The MABR train also had lower effluent ammonia
concentration when operated at low enough aerobic SRT and temperature that effluent ammonia began to increase, as shown in Figure 1.3. Effluent ammonia concentration and aerobic SRT are plotted in Figure 1.4 for the MABR train and the conventional train, along with the “washout” curve for nitrifiers at 14°C. Microbial analysis results from the biofilm and the bulk MLSS will be included in the final paper to quantify the nitrifier seeding effect. Results from qPCR analysis were not yet available at the time of abstract submission.
Discussion
The observed increased TIN removal is believed to be a result of 1) increased anoxic volume fraction 2) simultaneous nitrification and denitrification (SND) occurring in the MABR/anoxic zone, and 3) greater mass of nitrate introduced into the anoxic zone by the combination of SND and nitrate recycle pumping. These factors are enabled by the introduction of MABR modules into the anoxic zone of the activated sludge and could allow Gold Bar or other wastewater treatment plants to enhance total nitrogen removal performance.
The MABR modules installed in the pilot provided a seeding effect for nitrification by 1) removing between 30-40% of the influent ammonia in the biofilm and 2) sloughing nitrifiers into the mixed liquor to increase the nitrifying inventory of the activated sludge. These factors allowed the MABR pilot to have better nitrification than the control train while operating at a lower aerobic SRT. The ability to operate at a lower aerobic SRT while maintaining nitrification can allow wastewater treatment plants to increase their capacity without increasing the mixed liquor suspended solids concentration, avoiding the risk of overloading secondary clarifiers.
Figure 1.1 Pilot configuration
147
Figure 1.2 Final effluent total inorganic nitrogen for MABR and Conventional pilot trains.
Figure 1.3 Final effluent ammonia nitrogen for MABR and Conventional pilot trains and wastewater temperature
Figure 1.4 The theoretical “washout” SRT curve at 14°C and operating points at 14°C for the MABR train and conventional train.
25 20 15 10 5
0 0 50 100 150 200 250 300 350 400
Days of Operation
MABR Eff TIN Conventional Eff TIN
25 20 15 10 5
0 0 50 100 150 200 250 300 350 400
Days of Operation
MABR Eff NH3-N Conventional Eff NH3-N Bioreactor
Effluent Ammonia [mg/L]Effluent Ammonia NH3-N [mg/L], Temperature [°C]Effluent TIN, mg-N/L 30 25 20 15 10 5
0 0 1 2 3 4 5 6 7 8
Aerobic SRT [days]
Washout SRT MABR Conventional
148 References
Houweling, D., Long, Z., Peeters, J., Adams, N., Cote, P., Daigger, G., Snowling, S. (2018) “Nitrifying below the
“Washout” SRT: Experimental and Modelling Results for a Hybrid MABR/Activated Sludge Process”, Proceedings of the 2018 Water Environment Federation Technical Exhibition and Conference, New Orleans, Louisiana, September 29-October 3.
Presenting Author
Mr Matt Reeve Process Engineer
SUEZ Water Technologies and Solutions
Is the presenting author an IWA Young Water Professional? Y/N (i.e. an IWA member under 35 years of age)
Bio: Matt Reeve holds a degree in chemical process engineering from the University of British Columbia and is registered professional engineer in Ontario. He has experience in pilot testing, product development, and process engineering and is part of SUEZ Water Technologies & Solutions process engineering team, specializing in Membrane Aerated Biofilm Reactor technology. His focus is the commercialization of the ZeeLung MABR technology and is currently involved in several pilot and full-scale MABR projects.
149
Process Intensification Using MABR Technology: Subiaco WRRP Pilot Plant
I. Telles Silveira*, W. Bagg*, J. Peeters**, N. Oschmann**
*Water Corporation, Perth, WA, Australia
**SUEZ Water Technologies & Solutions, Oakville, ON, Canada
Keywords: Membrane Aerated Biofilm Reactor; Process Intensification; Pilot Plant
Summary of key findings
Following extensive investigation Water Corporation has identified Membrane Aerated Biofilm Reactor (MABR) as the preferred technology to increase secondary treatment capacity at Subiaco Water Resource Recovery Facility (WRRF), a large Perth Metro facility. When retrofitted into existing process volume, the innovative technology has the potential to intensify secondary treatment capability, reduce process aeration demand and greenhouse gas emissions. Started in July 2020, a two-year pilot trial is testing MABR suitability and will confirm location-specific system design requirements. The trial is the first of its kind in Australia and is supported by local and international experts.
Background and relevance
In 2016 a detailed review of potential secondary treatment upgrade and intensification options for Subiaco WRRF was carried out. The review considered the suitability of 14 different process options and, after a detailed comparison, including Multi-Criteria and NPV analysis, MABR technology was selected. Subiaco WRRF is a Modified Ludzack-Ettinger (MLE) system and has recently been upgraded to 67MLD, apart from the secondary treatment facility, which was excluded to minimize operational disruption. All tanks are covered for odour control, except the secondary clarifiers, and space on site is severely constrained. MABR technology has the advantage of simple installation. Retrofit of the technology directly to the aeration basins requires only minor changes to some odour control covers and provides the additional flexibility of being able to increase treatment capacity incrementally in step with future inflows.
To establish achievable nitrification rates under Western Australian temperature and sewage strengths, ZeeLung MABR technology has been trialled effectively using a side-stream pilot plant since July 2020. The 400 m2 footprint pilot plant has the capacity to treat 0.63MLD and is configured in a modular format, with eight tanks in series. Each of the tanks may be bypassed to test different arrangements. Tank 1 is a bioselector (not in use as per current configuration); Tanks 2 & 3 are designed as anoxic zones and are fitted with a commercial ZeeLung cassette each (Figure 1b); Tank 4 is fitted with fine bubble diffusers and a mixer, which allows the tank to operate under anoxic or aerobic conditions; Tanks 5 to 7 are fitted with fine bubble diffusers and Tank 8 holds a UF membrane, which is used to separate the MLSS from the treated water. Activated sludge is returned from Tank 8 to Tank 2. The objective of the treatment is to maximize ammonia removal, and assess system capability for nitrate removal by low dissolved oxygen denitrification.
Ammonia and nitrate levels are measured online throughout the treatment process: initially at the primary effluent, then at the inlet to the first ZeeLung tank; inside both ZeeLung tanks and finally at the treated water. Grab samples have also been collected and analysed three times a week to validate the online readings. COD of the primary effluent is also monitored online. As the air flows down the length of the ZeeLung filaments, oxygen is diffused through the media into the biofilm.
The oxygen and the diffused ammonia from the bulk liquid meet in the biofilm and nitrifying bacteria oxidize the ammonia to form nitrate. As an indication of the biofilm activity and the degree of treatment being carried out by
150
the ZeeLung zones, the inlet air flow rates, exhaust oxygen concentrations, temperature, pressure and volume of condensate are measured.
Figure 1 Pilot plant, overview of Tanks 1 to 4 on the left (a); MABR cassette, on the right (b)
Discussion
For the start-up, the reactors were seeded with activated sludge from Subiaco WRRF (sludge age of 10 days). The inflow of primary effluent was maintained at 0.30MLD during the initial couple of weeks. Subsequently, the plant has been fed continuously at 0.63MLD and total sludge age has been maintained at 4 days. The water temperature averaged 22 oC and ammonia concentration in the treated water reduced to approximately 1.0 mg/L after 1 month of the start-up (Figure 2). The spikes observed from 13 to 25 of September are attributed to mechanical problems with the aeration tankblower. Nitrate concentration has been maintained at 10 mg/L on the treated water.
Figure 2 Ammonia concentration in the treated wastewater
The nitrification rate (NR) by ZeeLung is calculated in terms of gNH4-N/d/m² (grams of ammonia per day per square meter of media) as shown in equation 1. Q and R are the plant inflow and recirculation flow, respectively; NH4N is the ammonia concentration entering and leaving the ZeeLung tank and; 1,920 m2 is the surface area of a ZeeLung cassette.
NR at each ZeeLung tank was assessed utilising grab sample test data (Figure 3). The rate of approximately 5 gNH4-N/d/m² was achieved in the first ZeeLung tank after 30 days of seeding the plant. In the second ZeeLung tank, the NR reached 3.3 gNH4-N/d/m². The variability observed in the results is a consequence of the diurnal profile of ammonia concentration in the primary effluent (ranges from 38 to 60 mg/L) and the respective sampling time.