full-scale application of mabr technology ......reach the attached-growth biomass from the bulk....

Environmental Engineering and Management Journal October 2020, Vol. 19, No. 10, 1781-1789

http://www.eemj.icpm.tuiasi.ro/; http://www.eemj.eu

“Gheorghe Asachi” Technical University of Iasi, Romania

FULL-SCALE APPLICATION OF MABR TECHNOLOGY FOR UPGRADING AND RETROFITTING AN EXISTING WWTP:

PERFORMANCES AND PROCESS MODELLING

Giuseppe Guglielmi1*, Daniel Coutts1, Dwight Houweling2, Jeff Peeters2

1Suez Water Technologies and Solutions Italy Srl, Via Benigno Crespi, 57, 20159 Milano, Italy 2Suez Water Technologies and Solutions, 3239 Dundas Street West, Oakville, ON, Canada

Abstract MABR (Membrane Aerated Biofilm Reactor) consists in an attached-growth process in which a gas permeable membrane is used as a biofilm carrier, and bubble-less oxygen transfer allows for a counter-diffusional mechanism where the electron donor (NH4+, biodegradable organics) and electron acceptor (O2) reach the biofilm from opposing sides. The paper reports results form the largest MABR installation in the world, demonstrating the beneficial impact of the technology in terms of improved process resilience under different operational conditions, through the key performance indicator oxygen transfer rate (OTR), which ranged between 7 and 16 gO2/m2/d. The OTR monitoring over the considered operational period shows (i) a weak impact of sludge temperature on the oxygen transfer within the considered range (9-22 degC) and also (ii) proves the reliability of OTR as an indicator of the improved resilience of the entire wastewater plant in the new MABR configuration, under peak loading conditions. Finally, a mathematical model was developed in the BioWin software platform, to evaluate its use as support tool for model-aided process optimization. The model was calibrated in terms of hydrodynamic behaviour of the MABR zone and seeding effect on the nitrifying biomass, which was expressed as solids retention time (SRT) of the attached growth biomass; the calibrated mathematical tool was then successfully validated in terms of prediction of ammonia concentration in the MABR zone. Keywords: modelling, Membrane Aerated Biofilm Reactor (MABR), resilience, upgrading Received: February, 2020; Revised final: June, 2020; Accepted: July, 2020; Published in final edited form: October, 2020 1. Introduction

Climate change and continuing population growth pose new challenges to wastewater management and make it necessary to adopt new technological solutions to achieve sustainability while guaranteeing reliable performance in a variety of operating conditions, which are more significant than in the past. With the aim of improving efficiency in water, land and energy utilization in a framework of circular economy, the need for smarter technological solutions in wastewater engineering is widely recognized as a top-priority challenge (EEA, 2019). The overall social-economical impact of these aspects have been pointed out by Daigger (2007) who *Author to whom all correspondence should be addressed: e-mail: [email protected]; Phone: +39 342 1287963

suggests a holistic view of water management in which water supply, wastewater collection and treatment and storm water management are interconnected components of broader picture.

Drivers such as (i) energy efficiency, (ii) performance reliability, (iii) optimized used of existing infrastractures and (iv) operational resilience become therefore crucial criteria when developing new technologies and optimizing their application at full-scale. In such context Membrane Aerated Biofilm Reactor (MABR) technology offers unique advantages which are intrinsically related to the technology working principle.

The great potential of coupling a gas-permeable membrane with a biofilm process in a so-

Guglielmi et al./Environmental Engineering and Management Journal 19 (2020), 10, 1781-1789

1782

called Membrane Biofilm Reactor (MBfR) has been deeply discussed by Rittmann (2007), who introduces possible applications of the technology with either O2 or H2 as gaseous substrate, for both drinking- and waste-water applications. In the MABR technology a gas-permeable membrane is used to transfer the electron acceptor (O2) to a biofilm naturally growing on the membrane. The complementary substrates (i.e. ammonia NH4+, biodegradable organics) typically diffuse from the bulk into the biofilm. This counter-diffusional mechanism results in significant differences compared to the traditional biofilm processes in which both electron donor and acceptor reach the attached-growth biomass from the bulk. Nerenberg (2016) reports a thorough description of the advantages of MABR, including.

development of unique microbial community with most favourable conditions for nitrifying microrganisms, which are also protected against possible inhibitors since they mainly growth in the inner layer of biofilm;

lower susceptibility to liquid diffusion layer resistance compared to traditional biofilm processes. In a counter-diffusional process, the layer actually tends to contrast the loss of internal substrate from the biofilm to the liquid bulk, whereas in a co-diffusional mechanism a high bulk concentration is necessary to overcome the liquid diffusion layer resistance;

much higher energy efficiency compared to conventional technologies, thanks to the bubble-less molecular diffusion of O2. The above-mentioned pros result in an overall more stable, efficient and environmental-friendly solution for wastewater treatment.

More specifically, the process intensification and improved process stability of the MABR technology under a wide range of operational conditions have been discussed in the literature, with a special focus on the hybrid configuration. A hybrid MABR-activated sludge (MABR-AS) system is where the attached-growth MABR is coupled with a conventional suspended-growth process. Starting from early studies (inter alia Brindle and Stephenson, 1996; Brindle et al., 1998), the achievement of consecutive nitrification and denitrification in the biofilm was demonstrated over a 300-days trial by Shin et al (2008) on a bench-scale CSTR system fed with synthetic wastewater, with a maximum nitrification rate of 2.06 gN-NH4+/m2/d and a denitrification rate of 1.76 gN-NO3-/m2/d. Hu et al. (2008) reported steadily high removal efficiencies for both ammonia nitrogen, total nitrogen and COD at hydraulic retention time varying between 12 and 20 hours. The same research group investigated the technology potential in achieving simultaneous nitrification denitrification (SND) under different COD/N ratios. The application of FISH (Fluorenscence In-Situ Hybridization) and confocal laser scanning microscopy techniques to biofilm samples collected from two bench -scale reactors revealed a strong prevalence of Nitrosomonas and Nitrospira, mainly distributed in the inner layer of

biofilm (Liu et al., 2010), thus confirming the beneficial effect of the counter-diffusional mechanism on the growth of nitrifying organisms over heterotrophs. Performances of a bench-scale MABR in a long-term continuous operation study have been recently presented by Lin et al. (2016), who demonstrate stable nitrification and total nitrogen removal efficiency in a COD/N ratio in the influent ranging between 3 and 5; also in this case, the application of microbiological techniques such as PCR-DGGE prove the impact of the counter-diffusional mechanism on the microbial diversity in the biofilm. The beneficial impact of MABR for retrofitting existing operational systems, by reducing the sludge age even below the washout SRT value for nitrifying biomass, has been demonstrated by Houweling et al. (2018) on demo-scale plants fed with real wastewater.

Using bubble-less oxygen transfer to the biofilm also allows for a better control of the electron acceptor for unconventional nitrogen removal pathways. Early investigations on the utilization of MABR to control nitration have been reported Lackner et al. (2008; 2010). A bench-scale study carried out by Gilmore et al. (2013) proves that a biofilm grown ex-novo with continuous aeration can combine nitratation and anaerobic ammonia oxidation in a single-stage MABR. A synthetic wastewater free of organic carbon was fed to the system, using relative loading ratios of oxygen and nitrogen as control tool for the suppression of NOB (Nitrite Oxidizing Bacteria). Removal rate of 1.7 gN/m2/d were demonstrated, the total removal efficiency of nitrogen being up to 85% through anaerobic ammonium oxidation. Experimental studies have proven that MABR can increase flexibility, be an energy efficient upgrade, and can be retrofitted in existing conventional activated sludge systems, thus making the technology a reliable alternate for the current and future wastewater treatment challenges. Sunner et al. (2018) have published results from a one-year study on a pilot unit fed with real sewage, reporting high efficiencies for BOD5 removal and nitrification (90% and 96% respectively), improved sludge settleability and aeration efficiency of 4 kgO2/kWh.

Despite the great potential that the technology has shown at both bench- and pilot/demo-scale level, there is a lack of reporting from full-scale installations proving the actual reliability of the process. This paper focuses on more than two years of operations of the site which is, based on authors’ knowledge, the largest hybrid MABR currently running worldwide. More specifically, the work focuses on:

1. the Key Performance Indicator (KPI) Oxygen Transfer Rate (OTR) and its capability to describe the process development and behaviour;

2. the benefits derived from the implementation of MABR in terms of overall stability and resilience of the nitrification process;

3. the use of mathematical modelling as advanced tool for optimization of process design and operation.

Full-scale application of MABR technology for upgrading and retrofitting an existing WWTP

1783

2. Material and methods

2.1. The Yorkville MABR The wastewater treatment plant of Yorkville-

Bristol Sanitary District (YBSD) was originally built on 1957, and it has been progressively modified and upgraded to handle the loading increase and the changes in environmental regulation. Until 2016 the YBSD facility consisted of: mechanical pre-treatment (coarse screening 6 mm, fine screening 1 mm wedge wire); fully aerobic biological process, in two lines which can be operated either in series or in parallel depending on the weather conditions. Each line is configured as an in-series scheme with five plug-flow reactors, equipped with fine bubble diffuser. The overall installed volume for the biological process is 5122 m3; three secondary clarifiers; UV disinfection.

In 2017, YBSD projected increased influent loading due to new industrial streams entering the facility and learned of the introduction of a new discharge limit for total phosphorus. The utility decided to implement a biological phosphorus removal process, by converting one of the two treatment lines into a hybrid MABR-AS system. Therefore, an anaerobic compartment was created within the volume of Line 1 and, in order to retain the overall nitrification capacity of the system under the new loading conditions, twelve ZeeLung cassettes were installed in an anoxic compartment immediately downstream the anaerobic zone. Therefore, the modification implemented for Line 1 from the original configuration to the upgraded one is shown in Fig. 1.

ZeeLung is the MABR solution developed by Suez Water Technologies Solutions. Each cassette consists of forty-eight modules, which in their turn are formed by cords. The cord structure consists of a braided polyester support, surrounded by a number of filaments (lumens). Each lumen is coated with a dense gas-permeable membrane. Air, at pressure higher than the hydrostatic pressure of the water level, is supplied to the module top header. The air passes inside of the lumens and oxygen diffuses through the membrane

where it is consumed by bacteria that have formed a biofilm on the outer wall of the lumen. The braided polyester structure of the cord makes it virtually unbreakable and suitable for installation in activated sludge.

Each membrane module accounts for 40 m2 of media available for biofilm growth, thus resulting in an overall media surface of 1920 m2 per cassette and 23,040 m2 in total. The process air flowrate ranges between 96 and 192 Nm3/h (8-16 Nm3/h/cassette) with a median value of 9.2 Nm3/h/cassette over the considered monitoring period. The pressure value at the air inlet of each cassette is approximately 480 mbar, whereas the outlet pressure is 270 mbar. The pressure loss across the media is partially due to headloss and to the loss of oxygen that has been transferred to the biofilm. Exhaust air at the bottom of each cassette is collected and sent to a sensor measuring oxygen concentration and then is expelled to the environment. An ejector is installed to allow for removal of condensate from each cassette. An addition coarse bubble aeration grid, designed around SUEZ’s membrane bioreactor LEAP aeration, is installed beneath the cassette and is fed with air that is scalped from the YBSD aerobic blower system. This installed coarse bubble aeration grid is used for substrate renewal within the cassette and biofilm thickness management.

Average data about influent feed water characteristics are summarised in Table 1.

Over the whole monitoring period considered for this study, MLSS concentration in the biological process tanks and overall SRT in the systems were 3000 ± 500 gMLSS/m3and 12 days respectively; dissolved oxygen concentration in the aerobic section of the plant was kept at the set-point of 2 gO2/m3.

2.2. Key Performance Indicators for the MABR technology

The typical KPI’s for the MABR process are

those parameters which summarize biofilm activity through the consumption of oxygen and the overall media efficiency in delivering oxygen. The Oxygen Transfer Efficiency (OTE) can be calculated as per Eq. (1):

Fig. 1. Original configuration (on the left) and upgraded configuration (on the right) of biological process scheme in Line 1 at the YBSD wastewater treatment plant


1784

Table 1. Average flowrate and feedwater composition during the considered monitoring period

Parameter Unit Value Average Daily Flowrate m3/d 10400 COD g/m3 422 BOD5 g/m3 190 TSS g/m3 164 TKN g/m3 32 N-NH4 g/m3 22.5 Total P g/m3 4.8

(1) The Oxygen Transfer Rate (OTR) represents

the mass of O2 delivered per unit of biofilm surface per unit of time. It is dimensionally expressed as gO2/m2/d, and can be calculated as per Eq. (2):

(2)

where: Qair is the process air flow rate fed to the ZeeLung cassettes; Abiofilm is the biofilm surface covering the membrane media, which is assumed to be equal to the installed membrane surface area.

2.3. Mathematical modelling of the full-scale MABR

Modelling is a tool in the wastewater industry

to test the hypothesis of new processes, understand the current operation of full-scale processes, and effectively communicate to a variety of audiences. At YBSD modelling was used as a communication and education tool to empower the operations team with a deeper understanding of the installed MABR-Activated Sludge hybrid system. A calibrated model would demonstrate how specific decisions made in the field would impact the overall process. Also, modelling would provide the operations team at YBSD a platform to probe operational approaches to various situations and see how the hybrid MABR-Activated Sludge system would react.

Modelling a biofilm process has two major phases for analysis: a solids phase and a liquid phase. There is no constant transport of materials in the solids phase and the liquid phase transfers soluble material into the biofilm where reaction occurs (Tartakovsky and Guiot, 2004); also, the rate of soluble material transfer to the biofilm is a function of diffusion characteristics of each soluble component (sCOD, sTKN, sNH4-N, etc) (Dzianach et al., 2019). There is a transfer of particulate components from the biofilm, which is modelled as a wasting process from the biofilm into the suspended growth system (Houweling and Daigger, 2019). The model of biofilm brings together the physical chemistry and the biological reaction components of the process to understand

impact the biofilm has on the hybrid MABR-AS system (Tartakovsky and Guiot, 2004).

Modelling the MABR in BioWin requires a combination of a few process modules to achieve the transport of materials through the liquid phase and restrict the transfer of materials in the solids phase. The conceptual framework of MABR modelling in BioWin eliminates the solids transfer from the bulk solution to a side-stream reactor, which represents the biomass attached to the membrane media. The model of the MABR process at YBSD was calibrated based on two main parameters: i) the hydrodynamic behaviour of the cassette, which determined the substrate renewal within the cassette; and ii) The biofilm solids retention time, which determined the treatment rate of the MABR biofilm.

In BioWin the transport of materials from the bulk solution to the “virtual” side-stream reactor is achieved through a “splitter tool”. The BioWin splitter represents the MABR mixing system, which is analogous to an air lift system. To understand the rate of exchange between the bulk solution and the MABR cassette volume a tracer study was completed onsite and a mathematic model calibrated to determine the hydraulic residence time (HRT) of the cassette. The results of the tracer study concluded that the rate of exchange between the MABR cassette and bulk solution was 24.9 m3/h, which meant that the MABR cassette had a 19-minute retention time (Underwood et al., 2018). The hydrodynamics of the anoxic zone at YBSD were also experimentally assessed through tracer tests. It was determined that the MABR anoxic zone was described as three continuous stirred tank reactors (CSTRs), with four cassettes in each CSTR. The BioWin splitter flow was calibrated based on the output of the hydrodynamic tests, thus providing an accurate model of the MABR operation within the anoxic tank.

When modelling a MABR process in BioWin, the seeding effect is modelled by wasting solids from the side-stream biofilm reactor into the bulk suspended growth system. The calibration of the seeding effect occurs by comparing the measured and modelled values of the aerobic bulk specific nitrification rate (SNR) for the suspended-growth biomass. Three biofilm SRT’s were calibrated: 25 days, 30 days, 35 days. The sum of squared differences (SSD) was used to determine which biofilm SRT best suited the measured SNR data of 2.57 gN/kgVSS/h, which was determined by completing a nitrogen mass balance around the aerobic system. A 30-days SRT resulted as best fitting value for the actual SNR, and

𝑂𝑇𝐸 20.9 % 𝑂2%𝑒𝑥ℎ𝑎𝑢𝑠𝑡 𝑎𝑖𝑟 ∙ 1 20.9 %1 𝑂2%𝑒𝑥ℎ𝑎𝑢𝑠𝑡 𝑎𝑖𝑟

20.9%

𝑂𝑇𝑅 𝑔𝑂2 𝑚2 ∙ 𝑑

𝑂𝑇𝐸 ∙ 𝑄𝑎𝑖𝑟 𝐿𝑎𝑖𝑟ℎ ∙ 20.9%

𝑚𝑜𝑙𝑂2𝑚𝑜𝑙𝑎𝑖𝑟 ∙ 32𝑔𝑂2𝑚𝑜𝑙𝑂2 ∙ 24

ℎ𝑑

22.4 𝐿𝑚𝑜𝑙𝑎𝑖𝑟 ∙ 𝐴𝑏𝑖𝑜𝑓𝑖𝑙𝑚 𝑚2


1785

was carried forward to model validation, which is summarized in the Results section. 3. Results and discussion 3.1. Process KPIs: OTR, OTE and NR

Real-time calculations of OTE and OTR was carried out from the online measurement of oxygen concentration in the exhaust air and air flow from the process air blower. The trend of daily OTR and average daily temperature measured during the first year of operation are reported in Fig. 2. Over the entire considered period, daily OTR ranged between 7.1 and 15.9 gO2/m2/d, with an average value of 11.6 ± 1.5 gO2/m2/d. OTE measured during the same period was steadily higher than 25%, with peak values up to 50% during high loading conditions.

Two statistical tests were performed to determine the OTR dependency on temperature: Chi-squared and Cramer’s V (see Table 2). The results of the Chi-squared (216 on 365-data series of T and OTR) and Cramer’s V (0.28 on the same data series) showed a relatively weak relationship between the two variables was observed. This weak relationship between temperature and OTR suggests that other

factors (e.g. the bulk ammonia concentration) influence OTR more than temperature itself.

A dedicated monitoring campaign was also carried out to correlate OTR and nitrification rate (NR). Composite samples were collected at the inlet and the outlet of the MABR zone to calculate NR and, over the entire monitoring campaign the observed nitrification rate on the biofilm was 2.1 gN-NH4+/m2/d, with peaks up to 3.1 gN-NH4+/m2/d. The average value OTR/NR ratio was 5.4 gO2/gN-NH4nitrified, which is higher than the stoichiometric ratio of 4.57 gO2/gN-NH4nitrified, thus suggesting that yet a part of oxygen was depleted by heterotrophic micro-organisms for oxidation of biodegradable organics. As far as it concerns the ammonia removal efficiency onto the biofilm, an average daily removal of 52 kgN-NH4+/d was observed, approximately corresponding to 22% of the influent ammonia nitrogen loading.

The removed loading of soluble COD (after 0.45 micron filtration) around the MABR zone over the same monitoring period was 232 kgCOD/d, due to both reactions occurring onto the biofilm (aerobic oxidation and denitrification) and simultaneous denitrification in the bulk sludge of the anoxic tank.

Fig. 2. Trend of average daily values for OTR and temperature during the considered monitoring period

Table 2. Absolute frequency of OTR values for different temperature values over the considered monitoring period

Temperature (°C) 9-10 10-11 11-12 12-13 13-14 14-15 15-16 16-17 17-18 18-19 19-20 20-21 Total

OTR

(gO

2/m

2/d)

7-8 1 0 0 1 1 0 0 0 0 0 0 0 38-9 0 2 0 1 1 1 1 1 1 0 1 0 99-10 0 5 5 7 5 3 3 1 1 0 2 0 3210-11 0 5 17 21 6 8 14 7 3 7 6 2 9611-12 1 3 18 9 13 3 10 3 4 9 12 3 8812-13 2 6 12 0 1 1 4 2 6 7 7 9 5713-14 0 3 2 0 3 1 1 3 1 2 4 21 4114-15 0 2 2 3 2 1 0 1 2 1 2 18 3415-16 0 0 0 0 0 1 0 0 0 0 0 4 5

Total 4 26 56 42 32 19 33 18 18 26 34 57 365


1786

3.2. MABR resilience

In general terms, the concept of resilience refers to the ability to overcome upsets and return to previous performance levels. When applied to wastewater engineering processes, resilience is synonymous to robustness, flexibility and ability of a system to promptly react to conditional changes. Fig. 3 demonstrates the resiliency of MABR with respect to changes in ammonia concentration within the bulk solution. In fact, Fig. 3 shows a counter-phase relationship between the MABR exhaust oxygen and the bulk ammonia concentration, thus proving that the biofilm is more active, and using more oxygen, when subject to higher ammonia concentrations in the bulk solution. The adaptive biofilm self-regulates its activity based on the demand created by the bulk ammonia concentration. Likewise, what is typically observed in respirometric tests for the assessment of nitrifying biomass activity through Ammonia Utilization Rate (inter alia Spanjers and Vanrolleghem, 2016), OTR reflects the “breathing” capability of the biofilm.

Fig. 3 also demonstrates a biofilm that is subject to an ammonia limited environment. A biofilm that is not ammonia limited will demonstrate no correlation or a weak correlation between bulk ammonia concentration and exhaust MABR oxygen concentration. By employing the Chi-squared (438.9 in a 288 data series) and Cramer’s V (0.87 in the same data series) tests, a strong correlation was found between bulk ammonia concentration and MABR exhaust oxygen concentration, thus supporting the YBSD MABR is an ammonia limited biofilm. The strong correlation between bulk ammonia and MABR exhaust oxygen concentration provides a more reliable alternative for performance and operation monitoring with minimal maintenance. The gas-phase MABR

exhaust oxygen sensor requires less maintenance, replacement parts, and calibration than the liquid-phase ammonia probes. The liquid sensors are more susceptible to drift and foreign material affecting the data. Therefore, by establishing a relationship of OTR and nitrogen removal, the MABR can use the oxygen balance to gain insights into the performance and operation of the system with minimal maintenance.

Fig. 4 demonstrates how the self-regulating biofilm activity in an ammonia limited biofilm is beneficial to the downstream aerobic suspended growth system. The introduction of the MABR at YBSD provides a self-regulating biofilm upstream of the aerobic suspended growth biomass, reducing the fluctuations in ammonia load on the downstream aerobic tanks. Therefore, the counter-phase relationship shown in Fig. 3 develops into the reduced amplitude in ammonia load taxing the downstream aerobic tanks, as shown in Fig. 4. The ability of the MABR biofilm to reduce the amplitude of the diurnal ammonia concentration pattern on the aerobic suspended growth makes the overall hybrid system more resilient to influent streams that have highly variant ammonia concentrations.

The YBSD MABR has also demonstrated hydraulic peak loading resilience. Peak hydraulic loads are known to cause biomass washout and, consequently, drop in performance of conventional activated sludge systems. Fig. 5 reports a wet weather event that occurred at YBSD in February 2019. The OTR trend observed before the wet weather event shows a regular behaviour, with an average OTR of (8.7 ± 1.1) gO2/m2/d. Wet weather in the area results in an influent flow increase of 2.4 times the average daily flow for 24 hours. During the time of the wet weather event the OTR maintains an average of (7.5 ± 1.3) gO2/m2/d, greater than 85% of the trend leading up to the wet weather event.

Fig. 3. Typical daily pattern of oxygen concentration in the exhaust air and ammonia concentration in the mixed liquor in the MABR zone


1787

Fig. 4. Calculated OTR and influent and effluent ammonia loading to/from the MABR zone

Fig. 5. Impact of rainy weather conditions on the OTR. The peak event (flow up to 2.4 times the average flow value) in the light grey area, the ammonia dilution effect in the MABR zone in the dark grey area

There is a clear impact of the dilution of

ammonia bulk concentration in the MABR zone reflected in the OTR trend during the 48 hours following the peak event. Leading up to the wet weather event, the ammonia ISE sensor showed a bulk ammonia concentration of (10.4 ± 3.9) gN-NH4+/m3, which drops to (4.5 ± 1.3) gN-NH4+/m3 in the next two days after the rainfall. During this same timeframe, the OTR decreases with the ammonia concentration down to (5.6 ± 0.9) gO2/m2/d. Nonetheless, once the ammonia concentration is restored to a normal operating range the biofilm promptly returns to the performance seen before the wet weather event, with OTR values back to (7.9 ± 1.4) gO2/m2/d.

3.3. Process modelling

A model calibration for YBSD was completed around data collected in May of 2018. This model calibration was later validated with data from August through September 2018. Online data provided the diurnal flow, RAS flow, wasting flow, and diurnal ammonia concentration calibration. A daily inlet composite sample was used to get the daily average of the inlet total COD, soluble COD, total BOD, TSS, TKN, total phosphorus, ortho-phosphate and ammonia. Composite samples were used to get the ammonia removal across the MABR zone. Grab samples were used to check the MLSS concentration.


1788

The influent model used COD, TKN, total phosphorus, and ISS as inputs and based on the fractionation of species the BOD, ammonia, ortho-phosphate and TSS were calculated. Based on MLSS concentration from collected samples the sludge production of the model was matched to achieve a representative model. Input values for wastewater characterisation and main biokinetic and stoichiometric parameters are summarised in Table 3. The ammonia removal of the biofilm was validated based on the composite samples that were collected around the MABR zone (Fig. 6).

Model capability to fit the experimental data was determined by means of absolute and relative error, as per Eq. (3) and Eq. (4) respectively.

(3)

(4)

While the absolute error calculated over the modelled period was 6.6%, the relative error was 1.1%. The difference in between the absolute error and relative error points to times of overestimation and underestimation. That said, both the absolute and relative error are less than 10% suggesting that the model is a good fit for the data collected from site,

which is subject to noise due to sample collection and processing.

The model developed a relationship of oxygen transfer rate to the biofilm, which was compared to the empirical data collected from online measurement of air flow to the media and exhaust gas concentration. The model does accurately represent the activity of the biofilm at a bulk ammonia concentration between 10-15 g/m3 but overestimates the OTR at higher bulk ammonia concentrations and underestimates the OTR at lower bulk ammonia concentrations. The discrepancy between the two values could be explained by the lack of modelling around the diffusion characteristics and the impact that has on the biofilm composition when employed in an ammonia limiting environment. One possible solution to overcome this would be the implementation of an additional model able to describe the physical properties on the diffusion rates of various soluble materials.

This BioWin model approach still does provide a platform for communication for how the biofilm system would respond to changes in the process. The trend is similar: as the ammonia concentration in the bulk goes up the OTR to the biofilm increase, and the opposite is true too. With this information, operators understand the impact of major wet weather events, impacts of how they manage the RAS flow, and the impact of acute increase in ammonia concentration from centrate returns from the sludge dewatering section.

Table 3. Input data used in the BioWin model for the validation period

Parameter Unit Value Average Daily Flowrate m3/d 7220 ± 1710 COD g/m3 395 ± 58 BOD5 g/m3 170 ± 25 TSS g/m3 166 ± 40 TKN g/m3 30.2 ± 3.6 N-NH4 g/m3 24 ± 2.9 Total P g/m3 4.8 ± 1 Ortho-phosphate g/m3 2.9 ± 0.6

Fig. 6. Validation of the YBSD MABR model in Biowin on influent and effluent ammonia concentration in the MABR zone

% 𝑒𝑟𝑟.𝑎𝑏𝑠𝑜𝑙𝑢𝑡𝑒 ∑ 𝑥𝑖,𝑚𝑜𝑑𝑒𝑙 𝑥𝑖,𝑚𝑒𝑎𝑠𝑢𝑟𝑒𝑑𝑛𝑖 1

∑ 𝑥𝑖,𝑚𝑒𝑎𝑠𝑢𝑟𝑒𝑑𝑛𝑖 1

% 𝑒𝑟𝑟.𝑟𝑒𝑙𝑎𝑡𝑖𝑣𝑒 ∑ 𝑥𝑖,𝑚𝑜𝑑𝑒𝑙 𝑥𝑖,𝑚𝑒𝑎𝑠𝑢𝑟𝑒𝑑𝑛𝑖 1

∑ 𝑥𝑖,𝑚𝑒𝑎𝑠𝑢𝑟𝑒𝑑𝑛𝑖 1


1789

This model also confirms the MABR peak trimming effect discussed in paragraph 3.2, where the large fluctuation of ammonia concentration is passively dampened by the MABR response to the ammonia concentration in the bulk. This phenomenon requires no control but provides an extra level of safety factor from breakthrough at the daily peak ammonia concentrations.

4. Conclusions

Field data from a full-scale membrane aerated

biofilm reactor in hybrid configuration have been presented and discussed, to demonstrate the reliability of MABR to intensify the nitrification process in existing wastewater treatment plants. The main outcomes from the collected data can be summarized as follows:

1. real-time monitoring of the MABR exhaust oxygen concentration provides a reliable, repeatable and affordable assessment of the overall process performance, through the calculated key performance indicator OTR;

2. OTR well correlates with ammonia removal rate on the biofilm, thus proving the favourable conditions for nitrifying micro-organisms in MABR;

3. impact of temperature on OTR shows a weak correlation, thus confirming the great potential of the technology under the most challenging conditions for nitrification (cold climates);

4. MABR technology has demonstrated overall process resilience, to handle both nitrogen and hydraulic peak events. The OTR trend observed during influent N peak loading events demonstrates that MABR has a self-regulating biofilm that reduces fluctuations in nitrogen loading to the downstream aerobic system. Similarly, OTR data collected during a wet weather event proves the technology ability to recover performance when regular conditions are restored;

5. a mathematical model built and calibrated in Biowin, was able to adequately describe the full-scale performance in terms of ammonia removal in the MABR zone and can now be used as support tool for process optimization. References Brindle K., Stephenson T., (1996), Nitrification in a bubble-

less oxygen mass transfer membrane bioreactor, Water Science and Technology, 34, 261-267.

Brindle K., Stephenson T., Semmens M., (1998), Nitrification and oxygen utilisation in a membrane aeration bioreactor, Journal of Membrane Science, 144, 197-209.

Daigger G.T., (2007), Wastewater Management in the 21st Century, Journal of Environmental Engineering, 133.

Dzianach P.A., Dykes Gary A., Strachan Norval J.C., Forbes Ken J., Pérez-Reche Francisco J., (2019), Challenges of biofilm control and utilization: lessons from mathematical modelling, Journal of the Royal Society

Interface, 16. EEA, (2019), European Environment Agency, Urban waste

water treatment for 21st century challenges, Briefing No 5/2019, On line at: https://www.eea.europa.eu/publications/urban-waste-water-treatment-for.

Gilmore K.R., Terada A., Smets B.F., Love N.G., Garland J.L., (2013), Autotrophic nitrogen removal in a membrane-aerated biofilm reactor under continuous aeration: a demonstration, Environmental Engineering Science, 30, 38-45.

Houweling D., Daigger G.T., (2019), Intensifying Activated Sludge Using Media-Supported Biofilms, CRC Press, Taylor and Francis Group.

Houweling D., Long Z., Peeters J., Adams N., Côté P., Daigger G.T., Snowling S., (2018), Nitrifying below the “washout” SRT: experimental and modelling results for a hybrid MABR/Activated Sludge Process, Proceedings of the Water Environment Federation, 16, 1250-1263.

Hu S., Yang F., Sun C., Zhang J., Wang T., (2008), Simultaneous removal of COD and nitrogen using a novel carbon-membrane aerated biofilm reactor, Journal of Environmental Sciences, 20, 142-148.

Lackner S., Terada A., Smets B.F., (2008), Heterotrophic activity compromises autotrophic nitrogen removal in membrane-aerated biofilms: results of a modeling study, Water Research, 42, 1102-1112.

Lackner S., Terada A., Horn H., Henze M., Smets B.F., (2010), Nitratation performance in membrane-aerated bifilm reactors differs from conventional biofilm systems, Water Research, 44, 6073-6084.

Lin J., Zhang P., Li G., Li J., Zhao X., (2016), Effect of COD/N ration on nitrogen removal in a membrane-aerated biofilm reactor, International Biodeterioration and Biodegradation, 113, 74-79.

Liu H., Yang F., Shi S., Liu X., (2010), Effect of substrate COD/N ratio on performance and microbial community structure of a membrane aerated biofilm reactor, Journal of Environmental Sciences, 20, 540-546.

Nerenberg R., (2016), The Membrane-Biofilm Reactor (MBfR) as a Counter-Diffusional Biofilm Process, Current Opinion in Biotechnology, 38, 131-136.

Rittmann B.E., (2007), The membrane biofilm reactor is a versatile platform for water and wastewater treatment, Environmental Engineering Research, 12, 157-175.

Shin J.-H., Sang B.-I., Chung Y.-C., Choung Y.-K., (2008), A novel CSTR-type of hollow fiber membrane biofilm reactor for consecutive nitrification and denitrification, Desalination, 221, 526-533.

Sunner N., Long Z., Houweling D., Monti A., Peeters J., (2018), MABR as a low-energy compact solution for nutrient removal upgrades – results form a demonstration in the UK, Proceedings of the Water Environment Federation, doi: 10.2175/193864718825137908.

Underwood A., Mc Mains C., Coutts D., Peeters J., Ireland J., Houweling D., (2018), Design and start-up of the first full-scale Membrane Aerated Biofilm Reactor in the United States, Proceedings of the Water Environment Federation, 16, 1282-1296.

Tartakovsky B., Guiot S.R., (2004), Biofilm Modelling, In: Fundamentals of Cell Immobilisation Biotechnology, (Focus on Biotechnology), Nedović V., Willaert R. (Eds.), vol. 8, Springer, Dordrecht.

/ColorImageDict > /JPEG2000ColorACSImageDict > /JPEG2000ColorImageDict > /AntiAliasGrayImages false /CropGrayImages true /GrayImageMinResolution 300 /GrayImageMinResolutionPolicy /OK /DownsampleGrayImages true /GrayImageDownsampleType /Bicubic /GrayImageResolution 300 /GrayImageDepth -1 /GrayImageMinDownsampleDepth 2 /GrayImageDownsampleThreshold 1.50000 /EncodeGrayImages true /GrayImageFilter /DCTEncode /AutoFilterGrayImages true /GrayImageAutoFilterStrategy /JPEG /GrayACSImageDict > /GrayImageDict > /JPEG2000GrayACSImageDict > /JPEG2000GrayImageDict > /AntiAliasMonoImages false /CropMonoImages true /MonoImageMinResolution 1200 /MonoImageMinResolutionPolicy /OK /DownsampleMonoImages true /MonoImageDownsampleType /Bicubic /MonoImageResolution 1200 /MonoImageDepth -1 /MonoImageDownsampleThreshold 1.50000 /EncodeMonoImages true /MonoImageFilter /CCITTFaxEncode /MonoImageDict > /AllowPSXObjects false /CheckCompliance [ /None ] /PDFX1aCheck false /PDFX3Check false /PDFXCompliantPDFOnly false /PDFXNoTrimBoxError true /PDFXTrimBoxToMediaBoxOffset [ 0.00000 0.00000 0.00000 0.00000 ] /PDFXSetBleedBoxToMediaBox true /PDFXBleedBoxToTrimBoxOffset [ 0.00000 0.00000 0.00000 0.00000 ] /PDFXOutputIntentProfile () /PDFXOutputConditionIdentifier () /PDFXOutputCondition () /PDFXRegistryName () /PDFXTrapped /False

/CreateJDFFile false /Description > /Namespace [ (Adobe) (Common) (1.0) ] /OtherNamespaces [ > /FormElements false /GenerateStructure false /IncludeBookmarks false /IncludeHyperlinks false /IncludeInteractive false /IncludeLayers false /IncludeProfiles false /MultimediaHandling /UseObjectSettings /Namespace [ (Adobe) (CreativeSuite) (2.0) ] /PDFXOutputIntentProfileSelector /DocumentCMYK /PreserveEditing true /UntaggedCMYKHandling /LeaveUntagged /UntaggedRGBHandling /UseDocumentProfile /UseDocumentBleed false >> ]>> setdistillerparams> setpagedevice

full-scale application of mabr technology ......reach the attached-growth biomass from the bulk....

Documents