Experience from start-ups of the first ANITA Mox Plants
M. Christensson1, S. Ekström1, A. Andersson Chan2, E. Le Vaillant1 and R. Lemaire3
1
AnoxKaldnes AB, Klosterängsvägen 11A, 226 47 Lund, Sweden ([email protected];
[email protected]; [email protected]). 2 Tekniska Förvaltningen, Box 1222 351, Växjö, Sweden
([email protected]) 3Veolia Water Technical Department, 1 rue Giovanni Battista Pirelli, 94417
St-Maurice, France ([email protected])
Abstract
ANITA Mox is a new 1-stage deammonification MBBR developed for partial nitrification to
nitrite and autotrophic N-removal from N-rich effluents. This deammonification process offers
many advantages such as dramatically reduced oxygen requirements, no COD requirement, lower
sludge production, no pre-treatment or requirement of chemicals and thereby being an energy and
cost efficient nitrogen removal process. An innovative seeding strategy, the “BioFarm concept” has
been developed in order to decrease the start-up time of new ANITA Mox installations. New ANITA
Mox installations are started with typically 3-15% of the added carriers being from the “BioFarm”,
with already established anammox biofilm, the rest being new carriers. The first ANITA Mox plant,
started up in 2010 at Sjölunda WWTP in Malmö, Sweden, proved this seeding concept, reaching an
ammonium removal rate of 1.2 kgN/m3.d and approximately 90% ammonia removal within 4
months from start-up. This first ANITA Mox plant is also the BioFarm used for forthcoming
installations. Typical features of this first installation were low energy consumption, 1.5 kW/NH 4Nremoved, low N2O emissions, <1% of the reduced nitrogen and a very stable and robust process
towards variations in loads and process conditions. The second ANITA Mox plant, started up at
Sundets WWTP in Växjö, Sweden, reached full capacity with more than 90% ammonia removal
within 2 months from start-up. By applying a nitrogen loading strategy to the reactor that matches
the capacity of the seeding carriers, more than 80% nitrogen removal could be obtained throughout
the start-up period.
Keywords
Anammox, ANITA Mox, Deammonification, MBBR, Sidestream treatment
INTRODUCTION
With the raising awareness on energy savings, anaerobic sludge digestion is now one of the key
processes leading towards energy-neutral or even energy producing WWTP. The drawback is the
increased ammonium concentration in the sidestream reject water that is recycled to the inlet of the
WWTP, which may constitute up to 20-30% of the overall incoming N load. In order to prevent
costly extension of the aeration capacity and the size of aerobic/anoxic tanks, together with
increased consumption of expensive external carbon sources, dedicated sidestream treatment
processes using autotrophic N-removal through anaerobic ammonium oxidation (anammox) have
gained increased popularity seen as a long-term solution, especially when considering energy- and
cost-effectiveness. Compared to conventional nitrification/denitrification processes, autotrophic Nremoval processes (i.e. anammox-type processes) offer many advantages such as (i) 60% reduction
of O2 demand, (ii) no COD requirement and (iii) lower sludge production.
Due to the very slow growth rate of anammox bacteria, different biofilm N-removal processes, with
or without support material, have been developed as 2-stage systems, such as combined SHARON /
Anammox-granular process (Abma, 2007) and 1-stage systems, also referred to as
“Deammonification” process, such as granular Sequencing Batch Reactors (SBR) (Wett, 2007;
Vlaeminck, 2008; Vazquez-Padin, 2009) or Moving-Bed Biofilm Reactors (MBBR) (Rosenwinkel,
2005; Cema, 2009).
The ANITA Mox process is a 1-stage MBBR deammonification process where partial nitrification
to nitrite by ammonia oxidizing bacteria (AOB) and autotrophic N-removal by anammox bacteria
occur simultaneously within the aerobic and anoxic zones of the biofilm resulting from oxygen
mass transfer limitation under limited dissolved oxygen (DO) condition. The biological processes
taking place inside the biofilm developed on the plastic carriers are illustrated in Figure 1.
NH4+
Liquid
N2
Nitritation
AOB
NO2
Anammox
Biofilm
Carrier
-
O2
Aerobic
Anoxic
Figure 1. Schematic of 1-stage nitritation/anammox biological processes occurring inside a
carrier’s biofilm. AOB oxidize NH4 to NO2 in the aerobic zone of the biofilm (i.e. outer part) while
anammox bacteria located in the anoxic zone of the biofilm (i.e. inner part) consume NO2 produced
by AOB together with the excess NH4.
The very slow growth of anammox bacteria and sensitivity towards high concentrations of oxygen,
nitrite/free nitrous acid and free ammonia during the start-up phase has been widely reported and
therefore limit a widespread application of anammox type processes. To shorten the start-up phase,
new ANITA Mox processes are seeded with a small fraction of colonized carriers, which reduce the
time required for the development of a mature deammonification biofilm on the brand new carriers.
The concept of seeding has proven to dramatically decrease the start-up time from up to a year
down to 2-3 months depending on the amount of seeding (Christensson et al., 2011, Lemaire et al.,
2011) and are in contradiction with studies by Schneider et al. (2009) reporting that seeding with
fully functional deammonification biofilm was not an efficient start-up strategy for MBBR
deammonification systems. To meet the request for seeding carriers, the first ANITA Mox full-scale
plant at Sjölunda WWTP is used as a nursery for anammox bacteria growing on suspended carriers,
referred to as the “BioFarm”.
An add-on feature to the ANITA Mox is the real-time DO control strategy, where nitrite production
is maximised without nitrite being further oxidized to nitrate. A control loop continuously calculates
the amount of NO3-Nprod/NH4-Nrem. based on signals from online sensors in the inlet and outlet
of the reactor. If the NO3-Nprod/NH4-Nrem.-ratio is higher than 11% it indicates that there is
oxygen available in excess favouring nitrite oxidizing bacteria, why DO set-point should be
decreased, and vice-versa, if the NO3-Nprod/NH4-Nrem.-ratio is lower than 11%, DO set-point
should be increased to further increase NH4-removal.
This paper presents experiences from start-up and operation of the first ANITA Mox implemented
in full-scale at Sjölunda WWTP in Malmö and from the second installation at Sundet WWTP in
Växjö.
The Sjölunda WWTP (550,000 PE) has a total capacity of 40 tBOD7/d and a maximum hydraulic
capacity of 4.4 m3/s to the biological treatment (Mases et al. 2010). Approximately 60% of the
reject water is treated by an existing SBR (1,920m3) operated in nitritation mode only and the
remaining 40% representing around 200 kgN/d is today treated by a full-scale ANITA Mox plant
(200m3).
The Sundet WWTP (80 000 PE) has a total capacity of 5.5 tBOD7/d and a maximum hydraulic
capacity of 0.8 m3/s to the biological treatment. Today, only primary sludge and biological sludge
are digested in two mesophilic anaerobic digesters, but in a near future external substrates (food
residuals) will be co-digested, hygienizated and digested a second time in a newly built digester.
After dewatering, all reject water is treated in the ANITA Mox process.
METHODS
Design parameters of full-scale ANITA Mox plant at Sjölunda WWTP and Sundet WWTP
The full-scale plant at Sjölunda WWTP consists of 4 MBBR with a total volume of 200m3. The
individual 50m3 reactors are 6m high, made of fibre glass, fully covered and insulated and can be
operated independently or in series. BiofilmChip™M K3 and Anox™ K5 carriers (AnoxKaldnes,
Sweden) have so far been used in these MBBRs. NH4+, NO3-, pH and TSS are measured on-line
(WTW, Germany) in the reject water feed and inside each reactor. Airflow and dissolved oxygen
(DO) are also measured and controlled in each reactor. An off-gas online analyser (Horiba, Japan)
was used for oxygen transfer trials and determination of potential N2O emission.
The main design and operating parameters of the ANITA Mox reactors are presented in Table 1 and
the reject water composition and sludge dewatering operation are detailed in Table 2.
Table 1 Design and operating parameters of full-scale ANITA Mox MBBRs
Parameters
ANITA Mox Sjölunda
ANITA Mox Sundet
Volume
4 x 50m3 (=200m3)
350m3
Design N-load
200 kgN/d
320 kgN/d (first phase)
430 kgN/d (second phase)
Carrier type
BiofilmChip™ M / K3 /
Anox™ K5
Anox™ K5
Temperature
22-33 °C (no control)
22-33 °C (no control)
pH
6.7-8.0 (no control)
6.7-7.5 (no control)
DO
0.5-1.5 mgO2/L (advanced
0.5-1.5 mgO2/L (advanced DO control
DO control strategy)
strategy)
Start-up
August 2010
November 2011
The full-scale plant at Sundet consists of a single 350m3 covered reactor. It is a refurbished SBR,
using the existing fine-bubble aeration system and with two 4 kW STAMO mixers. 43% Anox™ K5
carriers was added to meet the first design phase. NH4+ and pH are measured on-line (WTW,
Germany) in the inlet reject water feed and NH4+, NO3-, pH, DO inside the reactor.
Table 2. Reject Water characteristics at Sjölunda WWTP and Sundet WWTP
Characteristics
Units
Sjölunda Reject Water
Sundet Reject Water
Type of sludge digested
-primary + secondary sludge
primary, secondary and 20% external
sludge
Total flow
m3/d
650 (160 ; 126)*
144 1)
*
NH4
mgN/L
855 (148 ; 75)
1,043 1)
*
CODs
mg/L
257 (67 ; 22)
BOD7
mg/L
147 (73 ; 16)*
HCO3
mg/L
4,570 (130 ; 111)*
4,6502)
TSS
mgSS/L
330**
355 2)
*
mean (SD ; sample number); ** mean (from operator since 2006) 1) mean 2009-2011, 2) mean 2011,
Analysis
Mixed liquor samples were filtered at 1.6 µm before analysis for NH4+, NO3-, NO2- and soluble
COD using Dr Lange kits (Hach Lange, Germany). TSS were analysed according to the standard
methods (APHA, 1995).
RESULTS AND DISCUSSION
Due to limitations of seeding media available and in order to optimize the start up procedure, the 4
ANITA Mox reactors at Sjölunda WWTP were started up at different time. In 2010 the first 50m3
ANITA Mox full-scale reactor (MBBR1) was filled with 40% BiofilmChip M (BCM) carriers and
seeded with 0.9m3 pre-colonised carriers with anammox activity corresponding to 3% of the overall
volume. After 3 months of operation, effluent NH4 was typically below 100 mgN-NH4/L except
after occasional incidents of power failure (e.g. day 280), see Figure 2. The nitrate level in the outlet
was also on average below 100 mgN-NO3/L and the ratio NO3-produced : NH4-removed was in the
range of 8-22% (Figure 2c) which is close to the stoechiometric ratio (i.e. 11%) if nitrate was only
produced by anammox bacteria with no further oxidation of nitrite into nitrate by the nitrite
oxidizing bacteria (NOB) and no reduction of nitrate by heterotrophic denitrifiers. The real-time DO
control strategy developed for the ANITA Mox process was very successful in limiting the activity
of NOB in the system and therefore ensuring that most of the nitrite produced by the AOB is
actually used by the anammox bacteria. The nitrite concentration in the outlet was always very low
(i.e. < 5 mgN-NO2/L) although transitory accumulations of nitrite up to 120 mgN-NO2/L were
observed during the start-up period with no real impact on the MBBR performances. The fact that
nitrite was barely present in the MBBR indicates that the system was likely NO2 limited and that the
anammox activity was limited by the supply of NO2 from the AOB.
The N-loading and NH4-removal rate was increased steadily from the start-up as shown in Figure
2b. After only 4 months operation and with only 3% of seeding material, NH4-removal rate reached
1.2 kgN/m3react.d (Figure 2b) with approximately 90% NH4 removal efficiency (Figure 2c). These
performances were obtained (i) without any pre-treatment of the reject water, (ii) without any
chemical addition of methanol, acid or base solution, (iii) with no need for mechanical mixers in the
MBBR (i.e. continuous aeration strategy sufficient for mixing carriers) and (iv) without any heating
system even during the winter months in Sweden. Between operation day 280-450 the N-loading
and removal rate decreased to 0.70 – 1.1 kgN/m3day and then increased back to its initial capacity
of 1.2 kgNH4-N/m3d, Figure 2b. The period with lower N-loading and removal rates coincides with
issues with reject water supply but also with the warmer summer months where the process
temperature increases from around 27-29 °C up to 33 °C. It seems like the better performances that
are observed from operation day 450 are also related to the process temperature returning to below
30 °C from mid November 2011 (operation day 450). In spite of variation in N-load the NH4removal stayed high, typical > 90% for the entire operation period.
Long term operation of the MBBR1 at Sjölunda WWTP has proven the ANITA Mox to be a robust
and stable process towards variations in loads of reject water supply, variations in suspended solids,
power failure and pump issues typical encountered at treatment plants and still maintaining high
ammonium removal capacity. High energy efficiency, with a power consumption of 1.45-1.75
kWh/kgNH4-N, and low carbon footprint, with average N2O emissions in the range of 0.2 – 0.9% of
the reduced nitrogen, were obtained in MBBR1 making the ANITA Mox a sustainable solution
(Christensson et al., 2011).
50
100
150
200
250
300
350
400
450
500
550
600
1,2
30
1,0
25
0,8
20
0,6
15
0,4
NH4-N load
NH4-N removal
Temp
0,2
0,0
NO3-N, NO2-N out (mgN/L)
5
0
80
600 45
40
70
35
90 0
% NH4-N & TN removal
10
50
100
150
200
250
300
350
400
450
60
500
550
% NH4-N removal
% TN removal
% NO3-N produced
50
40
30
25
20
30
15
20
10
10
5
0
% NO3-N produced
N-load & removal rate (kgN/m3.d)
0
260
240
220
200
180
160
140
120
100
80
60
40
20
0
35
Temp C
NH4-N in
NH4-N out
NO3-N out
NO2-N out
NH4-N in & out (mgN/L)
1300
1200
1100
1000
900
800
700
600
500
400
300
200
100
0
1,4
0
0
50
100
150
200
250
300
Days
350
400
450
500
550
600
1,4
35
1,2
30
1,0
25
0,8
20
NH4-N load
NH4-N removal
Temp
0,6
15
0,4
10
0,2
5
0,0
Temp C
N-load & removal rate (kgN/m3.d)
Figure 2. Performances of the first 50m3 MBBR filled with 40% BCM since start-up: NH4, NO3
and NO2 measured in inlet and outlet (a); NH4 loading and removal rates (kgN-NH4/m3reactor.d) and
T°C in the MBBR (b); % TN and NH4 removal and ratio NO3-produced : NH4-removed in % (c).
0
0
20
40
60
80
100
120
140
160
180
Days
Figure 3. Temperature, NH4-N load and removal rates in MBBR4 from the BioFarm during restart
after 20m3 of colonised K5 carriers were taken out to seed the ANITA Mox process at Sundet
WWTP.
During 2010-2011 the remaining ANITA Mox reactors at Sjölunda were taken into operation using
different carriers and seeding percentage. Figure 3 shows the N-loading and removal rate after the
second start-up of MBBR 4 on Day 308 when 20m3 of Anox K5 carriers (i.e. corresponding to 80%
of the total media amount in the reactor) were taken out in order to seed the reactor at Sundet
WWTP for fast start-up. When 20m3 of new K5 carriers were added to the remaining 5m3 of precolonized anammox carriers left in the reactor, removal rate reached back to 1 kgN/m3d after only 2
months of operation and is still increasing to more than 1.2 kgN/m3d (Figure 3).
NH4-N in
NH4-N out
NO3-N out
NO2-N out
40
60
80
100
120
140
160
180 35
1,2
30
1,0
25
0,8
20
NH4-N load
0,6
15
0,4
10
0,2
5
90
0
20
40
60
80
100
120
140
160
0
180 45
80
40
70
35
60
% NH4-N removal
% TN removal
% NO3-N produced
50
40
Temp C
20
0,0
% NH4-N & TN removal
260
240
220
200
180
160
140
120
100
80
60
40
20
0
30
25
20
30
15
20
10
10
5
0
% NO3-N produced
N-load & removal rate (kgN/m3.d)
NH4-N in & out (mgN/L)
1300
1200
1100
1000
900
800
700
600
500
400
300
200
100
0
1,4 0
NO3-N, NO2-N out (mgN/L)
Performance of ANITA™ Mox full-scale plant at Sundet WWTP
The ANITA Mox process was started up at Sundet WWTP by seeding with 20m3 K5 carriers from
the BioFarm on the 6th of December 2011. The seeding media corresponds to 13% of total carrier
filling degree. Since the process was started up during the coldest winter months, the temperature
was kept around 28°C by batch wise heating of water when needed during the first month of
operation. To promote growth of anammox bacteria the reactor was fed with an amount of reject
water corresponding to an ammonium load slightly higher than the capacity of the seeding media.
The influent flow was continuously increased and the total amount of reject water was treated in the
ANITA Mox process after only 2 months of operation. Influent ammonium concentration to the
ANITA Mox reactor varied between 820-1100 mgNH4-N/L.From the first day of operation the
effluent NH4-N concentration has always been below 220 mgNH4-N/L with an average NH4-N
removal of 88% (Figure 4a&c). Nitrate concentrations in the MBBR outlet have been in the range
of 16-88 mgNO3-N/L which corresponds to a ratio 2-8.2% of produced NO3-N to NH4-N removed,
see (Figure 4a&c). One nitrite peak, corresponding to 60 mgNO2-N/L has been observed during the
operation period, otherwise the effluent NO2-N have an average concentration of 2.9 mgNO2-N/L
(Figure 4a).
0
0
20
40
60
80
100
120
140
160
180
Days
Figure 4. Operation results ANITA Mox Sundets WWTP, NH4, NO3 and NO2 measured in inlet
and outlet (a); NH4 loading and removal rates (kgN-NH4/m3reactor.d) and T°C in the MBBR (b); %
TN and NH4 removal and ratio NO3-produced : NH4-removed in % (c).
The nitrogen load on the process has increased from 0.01 to 0.83 kgN/m3d, which corresponds to an
increasing nitrogen removal from 0 to 0.63 kgN/m3d, (Figure 4b). An overall nitrogen removal over
84% has been maintained since the process start-up (Figure 4c). Relatively low loading and removal
rates in kgN/m3d at Sundet WWTP are explained with present shortage of reject water supply at the
WWTP compared to the design load and not due to poor process performance as could be seen at
the restart of MBBR4 where the seeded carriers original came from (Figure 3).
N2O measurements performed during 10 days in May 2012 shows that the N2O emissions from the
ANITA Mox unit at Sundet WWTP corresponds to an average of 0.75% of nitrogen removed. The
highest N2O emission recorded during the period equals to 3.1% of nitrogen removed and N2O in
the off gas varied between 0-163 ppm. The study shows that the process tends to emit least N2O
during stable operation without any disturbances and a DO concentration > 1 mgO2/L. Daily
averages shows that operation at a DO concentration < 1 mgO2/L can give rise to N2O emissions
corresponding up to 2.15% of nitrogen removed while stable operation with a DO concentration > 1
mgO2/L can allow low daily averages of N2O emissions down to 0.06% of reduced nitrogen. Figure
5 illustrates emitted N2O in percentage of nitrogen removed together with effluent NH4-N, NO3-N
and DO concentrations in mg/L during two 24 hour periods from the 23rd to 24th and from the 29th
to 30th of May. The graphs of DO and emitted N2O from the first 24 hour period indicates that the
aeration of the process is disturbed and that the disturbance occurs more frequently between
midnight of the 23rd and noon on the 24th (four times) where N2O emissions above 1% of reduced
nitrogen are measured directly after the unaerated period. The disturbance in aeration occurs less
frequently after noon until midnight of the 24th (two times) and Figure 5a shows that the N2O
emissions tend to decrease as the process is operated during longer periods without disturbances,
the N2O emissions measured after noon on the 24th have peaks corresponding to 0.8% of nitrogen
removed and low concentrations of less then 0.1% is recorded through longer periods. Figure 5b
shows a 24 hour period where aeration of the process is undisturbed and the N2O emissions are low
in the range of 0.04-0.16% of nitrogen removed.
80
NH4-N out
NO3-N out
% NO2-N produced
DO conc.
2
60
50
40
1
30
20
% N₂O-N prod, DO conc.
NH4-N & NO3-N out
70
10
2012-05-23 12:00
Days
2012-05-23 18:00
0
2012-05-24 00:00
2
NH4-N & NO3-N out
130
120
110
100
90
80
70
60
50
40
30
20
10
0
2012-05-29 18:28
2012-05-23 06:00
1
2012-05-30 00:28
2012-05-30 06:28
Days
2012-05-30 12:28
% N₂O-N prod, DO conc.
0
2012-05-23 00:00
0
2012-05-30 18:28
Figure 5. Typical concentrations of effluent NH4-N, NO3-N, DO and emitted N2O given in
percentage of reduced nitrogen during two 24 hour periods from the ANITA Mox plant at Sundet
WWTP.
CONCLUSION
• The ANITA Mox process is a very robust MBBR deammonification technology to address
issues of increasing N-load on WWTP coming from advanced sludge treatment processes while
reducing the overall carbon footprint and tending towards energy-sufficiency.
• The specific start-up strategy, where carriers with already colonized anammox biofilm were
seeded to the ANITA Mox process, shortened start-up time to 2 months at Sundet WWTP and 4
months at Sjölunda WWTP, reaching a N-removal rate of 1.2 kgN-NH4/m3reactor.d with 90%
NH4 removal efficiency
• Carriers with very large protected surface area, such as BiofilmChip M and Anox K5, are well
suited for deammonification processes as it creates a dense and compact biofilm with high
anammox activity allowing for smaller reactor design.
• The average N2O-emissions from the ANITA Mox process (i.e. 0.2-0.9% of N-removed) is in
the low range of what has been previously reported from different full-scale sidestream
anammox processes.
ACKNOLEDGMENTS
The authors would like to acknowledge all the personnel at Sjölunda WWTP (Malmö, Sweden) and
Sundet WWTP for their kindly assistance with the ANITA Mox operation.
REFERENCES
APHA (1995). Standard methods for the examination of water and wastewater. 19th edn, American
Public Health Association, Washington DC, USA.
Abma W., Schultz C., Mulder J., van der Star W., Strous M., Tokutomi T. and van Loosdrecht M.
(2007). Fullscale granular sludge Anammox process. Wat. Sci. Tech., 55(8/9), 27-33.
Cema G. (2009). Comparative study on different Anammox systems. PhD Thesis, KTH Stockholm.
Christensson M., Ekström S., Lemaire R., Le Vaillant E., Bundgaard E., Chauzy J., Stålhandske L.,
Hong Z., and Ekenberg M. (2011). ANITA Mox -A BioFarm solution for fast start-up of
deammonifying MBBRs. Conference Proceedings WEFTEC 2011, Los Angeles.
Lemaire R., Liviano I., Ekström S., Roselius C., Chauzy J., Thornberg D., Thirsing S., and Deleris
S (2011). 1-stage Deammonification MBBR process for reject water sidestream treatment:
investigation of start-up strategy and carriers design. Conference Proceedings WEF Nutrient
Recovery and Management 2011, Miami.
Mases M., Dimitrova I., Nyberg U., Gruvberg C. and Anderson B. (2010). Experiences from
MBBR Post-Denitrification Process in Long-term Operation at 2 WWTPs. IWA/WEF
Biofilm reactor technology conference, Portland, Oregon, USA, 15-18 Aug 2010.
Rosenwinkel K. and Cornelius A. (2005). Deammonification in the Moving-Bed process for the
treatment of wastewater with high ammonia content. Chem. Eng. Tech., 28(1).
Schneider Y., Beier M. and Rosenwinkel K. (2009). Impact of seeding on the start-up of the
deammonification process with different sludge systems. IWA conference on Nutrient
Removal. Krakow, Poland, 6-9 Sept 2009.
Vazquez-Padin J., Pozo M., Jarpa M., Figueroa M., Franco A., Mosqueral-Corral A., Campos J.L.
and Mendez R. (2009). Treatment of anaerobic sludge digesters effluent by the Canon
process in an air pulsing SBR. J. of Haz. Mat., 166, 336-341.
Vlaeminck S., Cloetens L., Carballa M., Boon N. and Verstraete W. (2008). Granular biomass
capable of partial nitritation and anammox. Wat. Sci. Tech., 58(5), 1113-1120.
Wett B. (2007). Development and implementation of a robust Deammonification process. Wat. Sci.
Tech., 56(7), 81-88.