innovations in wastewater treatment: the moving bed biofilm ... · the idea behind the development...
TRANSCRIPT
Innovations in wastewater treatment: the movingbed biofilm process
Hallvard Ødegaard
Norwegian University of Science and Technology (NTNU), Department of Hydraulic and Environmental
Engineering, NO-7491 Trondheim, Norway
Abstract This paper describes the moving bed biofilm reactor (MBBR) and presents applications of
wastewater treatment processes in which this reactor is used. The MBBR processes have been extensively
used for BOD/COD-removal, as well as for nitrification and denitrification in municipal and industrial
wastewater treatment. This paper focuses on the municipal applications. The most frequent process
combinations are presented and discussed. Basic design data obtained through research, as well as data
from practical operation of various plants, are presented. It is demonstrated that the MBBR may be used in
an extremely compact high-rate process (,1 h total HRT) for secondary treatment. Most European plants
require P-removal and performance data from plants combining MBBR and chemical precipitation is
presented. Likewise, data from plants in Italy and Switzerland that are implementing nitrification in addition
to secondary treatment are presented. The results from three Norwegian plants that are using the so-called
combined denitrification MBBR process are discussed. Nitrification rates as high as 1.2 g NH4-N/m2 d at
complete nitrification were demonstrated in practical operation at low temperatures (11 8C), while
denitrification rates were as high as 3.5 g NO3-Nequiv./m2.d. Depending on the extent of pretreatment, the
total HRT of the MBBR for N-removal will be in the range of 3 to 5 h.
Keywords Experiences; moving bed process; treatment; wastewater
Introduction
Biofilm processes are increasingly being favoured instead of activated sludge processes.
There are several reasons for this, such as:
† The treatment plant requires less space (an important cost factor).
† The final treatment result is less influenced by biomass separation since the biomass con-
centration to be separated is at least 10 times lower and there is greater flexibility with
respect to choice of biomass separation method (i.e. compact flotation or filtration).
† The attached biomass becomes more specialised (higher concentration of relevant
organisms) at a given point in the process train, because there is no biomass return.
There are already many different biofilm systems in use, such as trickling filters, rotat-
ing biological contactors (RBC), fixed media submerged biofilters, granular media biofil-
ters, fluidised bed reactors etc. They all have advantages and disadvantages. The trickling
filter is not volume-effective. Mechanical failures are often experienced with the RBCs.
It is difficult to get even distribution of the load on the whole carrier surface in fixed
media submerged biofilters. The granular media biofilters have to be operated discontinu-
ously because of the need for backwashing and the fluidised bed reactors show hydraulic
instability. For these reasons, the moving bed biofilm reactor process (Eur. pat. no.
0575314, US pat. no. 5,458,779) was developed in Norway in the late 1980s and early
1990s (Ødegaard et al., 1994; Ødegaard et al., 1999).
There are presently more than 300 treatment plants based on this process in operation or
under construction in 22 different countries all over the world. In addition, there are several
doi: 10.2166/wst.2006.284
Water
Science
&Techno
logyVol53No9pp17–33Q
IWAPub
lishing2006
17
Downloaded from https://iwaponline.com/wst/article-pdf/53/9/17/432865/17.pdfby gueston 08 November 2018
hundred on-site treatment units based on the MBBR – most of these in Germany. In
Europe, most of the larger plants are in operation in the Scandinavian countries, U.K., Italy
and Switzerland even though there are plants in 17 European countries. The MBBR plants
are used for municipal – as well as for various types of industrial wastewater treatment.
It has also been used for biological drinking water treatment as well as water treatment in
fish-farming. This paper will focus on municipal wastewater treatment applications.
Description of the moving bed biofilm reactor (MBBR)
The idea behind the development of the MBBR process was to adopt the best features of
the activated sludge process as well as those of the biofilter process without including the
worst. Contrary to most biofilm reactors, the MBBR utilises the whole tank volume for
Figure 1 Principle of the MBBR and shape of the original biofilm carrier (K1). (a) Aerobic reactor;
(b) Anoxic and anaerobic reactor; (c) The biofilm carrier (K1)
H.Ø
deg
aard
18
Downloaded from https://iwaponline.com/wst/article-pdf/53/9/17/432865/17.pdfby gueston 08 November 2018
biomass growth, as does also the activated sludge reactor. Contrary to the activated
sludge reactor, however, it does not need any sludge recycle, as is also the case in other
biofilm reactors. This is achieved by having the biomass grow on carriers that move
freely in the water volume of the reactor, kept within the reactor volume by a sieve
arrangement at the reactor outlet. Since no sludge recirculation takes place, only the
surplus biomass has to be separated – a considerable advantage over the activated sludge
process. The reactor may be used for both aerobic, anoxic or anaerobic processes, see
Figure 1.
In aerobic processes, the biofilm carrier movement is caused by the agitation set up by
the air, while in anoxic and anaerobic processes a mixer (normally a horizontal shaft
mounted banana mixer) keeps the carriers moving. A special coarse bubble aeration sys-
tem has been developed in the aerobic reactors. The carriers are kept within the reactor
by an outlet sieve. This may be vertically mounted, rectangular mesh sieves, but lately
the sieve is more often shaped as a cylindrical bar sieve, vertically or horizontally
mounted (Figure 2).
The original and mostly used biofilm carrier (K1) is made of high density polyethylene
(density 0.95 g/cm3) and shaped as a small cylinder with a cross on the inside of the cylin-
der and “fins” on the outside (see Figure 1(c) and Figure 3). The cylinder has a length of
7mm, and a diameter of 10mm (not including fins). Various carriers of other sizes and
shapes are since been introduced, but the K1 carrier is still the one mostly used.
One of the important advantages of the moving bed biofilm reactor is that the filling
fraction of carrier in the reactor may be subject to preferences. The standard filling frac-
tion is 67%, resulting in a total, specific carrier area of 465m2/m3 with the K1 carrier.
Since the biomass is growing primarily on the inside of the carrier (see Figure 3), one
is calculating with an effective specific surface area of 335m2/m3 at 67% filling fraction.
Figure 2 Photos of MBBR reactors under construction. (a) Aerobic reactor with aeration system and
vertically mounted cylindrical bar sieves; (b) Anoxic reactor with horizontally mounted shaft mixers and
rectangular mesh sieves
H.Ø
deg
aard
19
Downloaded from https://iwaponline.com/wst/article-pdf/53/9/17/432865/17.pdfby gueston 08 November 2018
In order to be able to move the carrier suspension freely, it is recommended that filling
fractions should be below 70% (corresponding to 350m2/m3 effective specific area for
K1). One may, however, use as much as needed below this. This is convenient, especially
when upgrading plants, – for instance, from activated sludge to moving bed reactors. It
has been demonstrated (Ødegaard et al., 1999) that the biofilm area is the key parameter
in design and therefore the design rate of the process is most correctly based on effective
carrier area (g=m2carrier aread).
As in every biofilm process, diffusion of compounds in and out of the biofilm plays a
key role. Because of the importance of diffusion, the thickness of the effective biofilm
(the depth of the biofilm to which the substrates have penetrated) is significant. Since this
depth of full substrate penetration is normally less than 100mm, the ideal biofilm in the
moving bed process is thin and evenly distributed over the surface of the carrier. In order
to obtain this, the turbulence in the reactor is of importance, both in order to transport the
substrates to the biofilm and to maintain a low thickness of the biofilm by shearing
forces. As demonstrated in Figure 3, much less biomass is growing on the outside of the
carriers than on the inside. This is probably caused by the fact that abrasion, caused by
carrier collisions, is limiting growth.
Various investigations have shown that the typical biomass concentration, when calcu-
lated on reactor volume, is in the order of 2–5 kg SS/m3 (Rusten et al., 1994, 1995a,
1998), approximately the same as in activated sludge reactors. However since the volu-
metric removal rate, has been demonstrated to be several times higher in the moving bed
process (Rusten et al., 1995a), the biomass of this process must be much more viable
than in similar activated sludge processes.
With an activated sludge plant, the process of developing an active biomass begins
with the growth of zoogleal bacterial floc which then becomes colonised by protozoa
which feed on the free swimming bacteria to produce a clarified effluent. In the moving
bed process, the order of colonisation seems to be reversed (Mosey, 1996). High loading
rates, around 30 g COD/m2d, produce compact bacterial biofilms, with protozoan popu-
lation either absent or limited to small free-swimming protozoa and Vorticella spp. Mod-
erate loading rates, around 10–15 g COD/m2d promote a more “fluffy” biofilm with a
rich variety of ciliated protozoa. Low loading rates (,5 g COD/m2d) promote very
“fluffy” biofilm generally dominated by stalked ciliates.
Figure 3 The moving bed biofilm reactor
H.Ø
deg
aard
20
Downloaded from https://iwaponline.com/wst/article-pdf/53/9/17/432865/17.pdfby gueston 08 November 2018
The MBBR processes
The moving bed biofilm reactor process has been used for many different applications.
It was developed at the time when nitrogen removal came into focus and most of the
scientific data has been gathered from this application. Later, however, organic matter
removal has been investigated further, including high-rate pretreatment for upgrading of
activated sludge plants. A process for high-rate secondary treatment based on the moving
bed reactor in combination with coagulation and flotation, resulting in a total residence
time of the process of less than 1 h, has also been developed (Ødegaard et al., 2004).
Research has also been conducted in order to develop a MBBR biological phosphate
Figure 4 Typical moving bed biofilm process flow diagrams for different applications
H.Ø
deg
aard
21
Downloaded from https://iwaponline.com/wst/article-pdf/53/9/17/432865/17.pdfby gueston 08 November 2018
removal process based on the sequencing batch reactor principle (Helness and Ødegaard,
2001). In Figure 4, the most common flow diagrams for various treatment purposes are
presented. These diagrams will be used as the basis for discussing results and experiences
with the moving bed process so far.
Presettling is normally used as pretreatment, primarily to avoid clogging of the
bio-reactor sieves. Presettling is not a prerequisite, but used to reduce load on the bio-
reactor and to add flexibility to the process by allowing chemically enhanced primary
treatment – if needed.
BOD/COD removal � possibly combined with P-removal (Figure 3(a)–(c))
Because of the compactness of the process, the residence time in moving bed reactors for
carbonaceous matter removal will be quite low (15–90min), depending on the organic
load and the strength of the wastewater. Biodegradable, soluble organic matter is quickly
degraded. Particulate organic matter is partly caught by the irregularities of the attached
biomass, hydrolysed and degraded, and, partly, it passes more or less unchanged through
the reactor. In order to evaluate degradation of organic matter independent of the biomass
separation step, one may look at the removal rate of soluble COD (SCOD). It is demon-
strated in Figure 5(a) that the maximum removal rate was found to be around 30 g
SCOD/m2d. SCOD is filtered (1mm) COD (Ødegaard et al., 2000). This does not, how-
ever, give the true picture since biodegradable, soluble organic matter is also produced in
the process by hydrolysis.
0
5
10
15
20
25
30
35
40
45
50(a)
(b)
0 20 40 60 80 100
Filtered COD loading rate (g SCOD/m2*d)
Filte
red
CO
D r
emov
al r
ate
(g S
CO
D/m
2 *d)
K1 K2 100%
K1 K2 100%
0
20
40
60
80
100
120
140
0 50 100 150 200
Total COD loading rate (g COD/m2*d)
Obt
aina
ble
rem
oval
rat
e(C
OD
in-S
CO
Dou
t) (g
/m2 *d
)
Figure 5 Soluble COD removal rate versus soluble COD loading rate (a) and “obtainable” COD removal
rate versus total COD loading rate (b)
H.Ø
deg
aard
22
Downloaded from https://iwaponline.com/wst/article-pdf/53/9/17/432865/17.pdfby gueston 08 November 2018
An alternative is to evaluate the so-called “obtainable” removal rate, meaning the
removal rate of total COD at 100% biomass separation. Figure 5(b), which gives the
“obtainable” removal rate versus total COD loading rate, demonstrates that high removal
efficiencies may be obtained even at extremely high loading rates if good biomass separ-
ation can be assured.
In highly loaded plants, clarification of the biomass does, however, represent a pro-
blem since the settleability of the biomass decreases with increasing organic loading
(Ødegaard et al., 2000). The key to solving this problem is either biomass flocculation by
coagulant addition (metal salts or cationic polymers) and/or use of another type of bio-
mass separation, such as flotation or filtration.
In most European countries, phosphate removal is required and carried out by chemi-
cal precipitation. It is very common, therefore, to use process 4 a,b when the target is
BOD/COD and P-removal. These plants will be equipped with adequately designed floc-
culation tanks and they may use alum, iron chloride or prepolymerised aluminium chlor-
ide as coagulants.
The 2000–2002 treatment results from four Norwegian plants based on this process
scheme are shown in Table 1. These are all normal rate plants consisting of two MBBRs
in series followed by coagulation/flocculation and sedimentation. The design organic
loading rates are in the range of 7–10 g BOD7/m2d at a temperature of 10 8C. The results
demonstrate that high removal efficiencies and low effluent concentrations were achieved
for the moving bed reactor used in this process scheme.
Upgrading from chemical treatment to biological–chemical treatment
For the TAU direct precipitation plant in Tonsberg, Norway, pilot tests and cost analysis
for upgrading to secondary treatment have been carried out (Rusten and Nedland, 2003,
2004). The plant will be required to remove more than 90% phosphorus in addition to
fulfilling the requirements for secondary treatment in the EU wastewater directive.
Upgrading with conventional activated sludge would require large aeration basins and
new settling tanks for activated sludge. For upgrading with the MBBR process, pilot tests
showed that the existing settling tanks are more than adequate, as the biological–chemi-
cal sludge settled well at overflow rates up to 2.9m/h. At the design temperature of 8 8C,
Table 1 2000–2002 results from four Norwegian plants for BOD/COD- and P-removal
WWTP BOD7 COD Tot P
In Out % In Out % In Out %
Steinsholt1
Average 398 10 97.4 833 46 94.4 7.1 0.30 95.8Max 1720 38 99.7 2760 130 98.4 12.0 0.72 98.8Min 120 5 93.5 190 30 93.5 4.0 0.12 92.6TrettenAverage 361 4 98.9 – – – 7.3 0.10 97.9Max 695 16 99.7 – – – 15.5 0.44 99.8Min 125 2 97.7 – – – 4.2 0.03 89.4SvarstadAverage – – – 403 44 89 5.1 0.25 89Max – – – 850 83 94 13.0 0.78 94Min – – – 230 30 78 2.0 0.10 78FryaAverage 181 5 97.7 – – – 8.6 0.21 97.6Max 290 20 99.0 – – – 12.0 0.53 99.5Min 85 2 93.1 – – – 6.0 0.06 95.1
1Data from 1996–97
H.Ø
deg
aard
23
Downloaded from https://iwaponline.com/wst/article-pdf/53/9/17/432865/17.pdfby gueston 08 November 2018
future treatment requirements were fulfilled at organic loads on the MBBR up to 25 g
COD/m2/d. This will limit the upgrade to a new pumping station and a compact MBBR
process between the sand traps and flocculation basins of the existing plant. The biologi-
cal process will also reduce the chemical costs by at least 30%, translating to significant
savings in O&M costs.
The design load for the plant is 130,000 PE. Costs for upgrading with MBBR technol-
ogy or with conventional activated sludge are shown in Table 2. The O&M costs show
the difference between the existing plant and an upgraded plant, and they are low due to
the savings in chemical costs for a biological-chemical plant.
The huge cost savings by using a MBBR process are typical for upgrading of chemical
treatment plants, where the MBBR process can use existing settling tanks, while an acti-
vated sludge process will require new settling tanks.
A high rate process for secondary treatment
Currently, a new process concept is being developed for secondary treatment based on
the combination of a high rate moving bed process combined with low dose coagulation
(Ødegaard et al., 2004). If combined with fine sieving for pretreatment and flotation for
biomass separation, this gives an extremely compact process with a total residence time
of less than 1 hour (Figure 6).
The process is based on the fact that a very considerable fraction of the organic
matter (65–85%) and a substantial fraction (35–55%) of the phosphorous in wastewater
appears as colloidal and suspended matter (Levine et al., 1985; Ødegaard, 1998; van
Nieuwenhuisen and Mels, 2002). Direct separation of suspended and colloidal particles
by chemical coagulation will normally, therefore, yield an organic matter removal in the
65–75% range (Ødegaard, 1992). There are two drawbacks with traditional coagulation
of wastewater in view of a secondary treatment standard, however; (a) true soluble
organic matter is not removed and (b) metal (Al or Fe) hydroxide precipitation leads to
high sludge production (Ødegaard, 1998). The process consists of pretreatment by a fine
Table 2 Relative costs for upgrading to biological–chemical treatment at the TAU WWTP. Investment costs
for MBBR upgrade set at 100
Relative costs
MBBR upgrade Activated sludge upgrade
Investment costs 100 378O&M costs 1.8 8.0Total annual costs 12.8 50.0
Pretreatmentfine sieve
MBBRHRT: 15–30 min
FlocculationHRT: 5 –10 min
FlotationHRT: 15 –25 min
Coagulant
Air
Figure 6 A high rate secondary treatment process based on moving bed, low dose coagulation
and flotation
H.Ø
deg
aard
24
Downloaded from https://iwaponline.com/wst/article-pdf/53/9/17/432865/17.pdfby gueston 08 November 2018
sieve (or existing primary settling) followed by highly loaded moving bed biofilm reactor
after which a polymer- and iron-coagulant is added at low dosage and the biomass/floc
aggregate is separated from the water by flocculation/ flotation.
The basic idea is that the polymer/iron coagulation/floc separation will take care of
the suspended and colloidal organic matter while the biofilm will take care of the low
molecular weight, easily biodegradable, soluble organic matter. A cationic polymer is
chosen as the primary coagulant in order to be able to operate with a very low dosage of
iron and thereby minimise hydroxide precipitation and, consequently, sludge production.
Flotation is chosen because a highly loaded bioreactor may result in poor settleability but
easily floatable sludge, and also because it is a low foot-print separation technique.
It has been demonstrated (Ødegaard et al., 2004) in a fairly large scale pilot-plant
(5m3/h) that the process could operate favourably at dosages as low as 5 g polymer/g SS
and 35 g Fe/g SS – corresponding to dosages of 1mg polymer/l and 7mg Fe/l at a typical
200mg SS/l in the bioreactor effluent. These dosages resulted in a sludge production
which was very close to the suspended solids removed from the wastewater, i.e. very
little additional sludge production caused by precipitation.
In a high rate process like this, the residence time distribution of the bioreactor will be
very important and the reactor should be divided in steps in order to prevent short-circuiting.
In order to be sure to reach secondary treatment standard, the total reactor loading should not
exceed 20–25g CODfiltered/m2biofilm aread (or 15–20g BOD5 filtered/m
2biofilm area d). This corres-
ponds, however, to loadings as high as 65–85 g CODtot/m2biofilm area d (45–60g BOD5tot/
m2biofilm area d) in typical wastewater, which is 2–4 times higher that the normal design load
for traditional biofilm reactors (i.e. RBCs) targeting secondary treatment standard.
Design of a high rate MBBR will be carried out by choosing the desired MBBR efflu-
ent FCOD concentration and calculating the required volume and biofilm growth area
that incorporates the effect of the hydraulics and inert filtered COD. The total suspended
solids concentration in the MBBR effluent can be calculated as the sum of influent SS
concentration and the biological sludge production (of 0.5 g SS/g FCODdegraded) in the
MBBR. The specific sludge production in the separation part of the process was found to
be 1.0 g DS/g SSremoved, indicating that there was minimal extra sludge production due to
precipitation. The total sludge production design for the process can, therefore, be based
on the influent suspended solids plus the MBBR sludge value for production, provided
that a low dose of metal coagulant is used.
Nitrification
The moving bed biofilm reactor process has been used for nitrification purposes in
various process schemes (see Figure 4 (d)–(f)). Traditionally, nitrification is obtained by
adding reactor volume to reactors that have already removed most of the organic matter
(Figure 4(d)). In contrast to an activated sludge system, the heterotrophs will dominate in
the start of the process (first reactor) and the nitrifiers in the end of the process (last reac-
tor). This makes it possible to optimise each of the processes independent of the other. In
some Scandinavian plants, P-removal and substantial BOD-removal is achieved by
chemically enhanced primary treatment, thus reducing the size of the bioreactor necessary
to obtain nitrification. In this case, the load of particles (biomass) on the nitrifying reac-
tors will be less, which results in higher nitrification rates than when precoagulation is
not used (Rusten et al., 1995b).
Three factors, the load of organic matter, the ammonium concentration and the oxygen
concentration, primarily determine the nitrification rate. The influence of these parameters
is schematically shown in Figure 7 (Hem et al., 1994). Figure 7(a) demonstrates that the
organic load controls nitrification and should be as low as possible. At an organic loading
H.Ø
deg
aard
25
Downloaded from https://iwaponline.com/wst/article-pdf/53/9/17/432865/17.pdfby gueston 08 November 2018
above approximately 4 g BOD7/m2d, a high oxygen concentration (.6mg O2/l) is
required in order for nitrification to take place at all. The figure also demonstrates the
linear relationship between nitrification rate and oxygen concentration with the actual
concentration range. As shown in Figure 7(b), the ammonium concentration is only limit-
ing the nitrification rate at low concentrations (,1–3mg NH4-N/l). Above a bulk
concentration of 3mg NH4-N/l, the nitrification rate is governed by the oxygen concen-
tration and the organic load.
Far more important is the influence of oxygen concentration that may limit the nitrifi-
cation rate, even at concentrations above 2–3mg O2/l. The nitrification rate is found to
be close to linearly dependent upon the oxygen concentration, up to more than 10mg
O2/l (Ødegaard et al., 1994; Æsøy et al., 1998). An advantage of the linear relationship
between oxygen concentration and nitrification rate is that it may very favourably be
used for process control.
High oxygen concentration is only required for nitrification. For carbonaceous
removal, 2–3mg O2/l has proved to be sufficient in practice.
Figure 8 shows the pilot data from the Givaudan WWTP in Switzerland that was
run at 10 8C and a filling fraction of 60%. They demonstrate that an effluent concen-
tration ,1mg NH4-N could be maintained at loadings up to 0.35 g NH4-N/m3d (1.17 g
NH4-N/m2d).
In 2003, another Swiss plant – Durrenroth WWTP, a retrofitted RBC-plant – demon-
strated close to complete (97%) nitrification on a yearly basis with mean ammonium
2.5
2.0
1.5
1.0
0.5
0.00 2 4 6 8 10
(a) (b)
Am
mon
ia r
emov
al r
ate
g N
H4-N
/m2 d
Oxygen concentration, mg O2/ l
2.5
2.0
1.5
1.0
0.5
0.00
Ammonium concentration, mg NH4-N/l
Nitr
ific
atio
n ra
te, g
NH
4-N/m
2 d
DO=9mg/l
DO=6mg/l
DO=3mg/lOrgan
ic load
= 0.0g BOD 7
/m2 d
1.0
2.0
3.04.0
5.06.0
7.0
1 2 3 4
0.4g BOD7/m2d
Figure 7 Influence of BOD7, oxygen and ammonium on nitrification rate
600
500
400
300
200
100
00 100 200 300 400 500 600 700 800
Volumetric ammonium load (g NH4-N/m3.d)
Nitr
ific
atio
n ra
te (
g N
H4-
N/m
3 .d)
NH4 Elimination NO3-Production Poly. (NH4 Elimination)
Loading limit foreffluent concentration
<1 mg/l/ NH4-N
Figure 8 Nitrification performance 10 8C at Givaudan WWTP (Tschui, 2004)
H.Ø
deg
aard
26
Downloaded from https://iwaponline.com/wst/article-pdf/53/9/17/432865/17.pdfby gueston 08 November 2018
concentration in the effluent of 0.2mg NH4-N/l and a 90% limit value of 0.6mg NH4-
N/l, even though the temperature over the year varied between 5 and 18 8C (Stolz, 2004).
In one Italian plant (Bergamo WWTP) the nitrifying bioreactor is placed down-stream
an activated sludge plant (see Figure 4(e)) and because of the low sludge production, no
biomass separation after the nitrifying reactor is used. If a stringent SS standard is to be
complied with, direct sand filtration could be used to remove the biomass. Due to the low
organic load on the MBBR, the observed nitrification rates in this plant are high, typically
1.2–1.5 g NH4-N/m2 d (at 15 8C and DO . 5mg O2/l).
The process scheme depicted in Figure 4(f), often called a hybrid nitrification process,
is not extensively used in Europe, but has become popular in the USA. In this process,
the last part of the aeration tank of an activated sludge plant is retrofitted into a moving
bed process. In this way, the nitrifying biomass of the system is increased and nitrifica-
tion may be achieved in a volume that otherwise could not comply with a nitrification
standard.
Nitrogen removal
Nitrogen removal in MBBR plants may be achieved by several process combinations, for
instance predenitrification (Figure 4(g)), post-denitrification (Figure 4(h)) or a combi-
nation of the two – the so-called combined denitrification process (Figure 4(i)). The
MBBR process may also be used in activated sludge plants as a post-denitrification
process (Figure 4(j)) or as hybrid activated sludge/biofilm process. The latter is unusual
in Europe but common in the USA.
The denitrification rate may be limited by the nitrate concentration, the biodegradable
organic matter concentration or by the oxygen concentration (or rather the presence of
oxygen). At NO3-N concentrations above approximately 3mg NO3-N/l, the denitrification
removal rate will be completely governed by the type and availability of easily biode-
gradable carbon source (Rusten et al., 1995b). If oxygen is supplied to the reactor with
the inlet water or recirculated water, biodegradable organic matter will be consumed for
oxygen respiration and thus reduce the available amount for denitrification.
The limitation of the predenitrification process (Figure 4(g)) stems from the fact that
oxygen-rich water from the nitrification step will have to be returned to the predenitrifica-
tion step. The raw water carbon source is very often not sufficient, and the denitrification
rate in predenitrification systems will normally be limited by the carbon source avail-
ability and consequently be rather low. In post-denitrification systems (Figure 4(h)–(j)),
one will add an easily biodegradable carbon source with a very high denitrification rate
as a result.
The post-denitrification mode has several advantages over the predenitrification mode.
It may lead to considerably lower total bioreactor volumes (40–50%) and it gives much
better process control. The major drawback is the need for carbon source addition. In
order to minimise the use of carbon source, the flow scheme of Figure 4(i), combined
denitrification, – has been preferred in several Norwegian plants. The combination pro-
cess offers greater flexibility with respect to operation of the plant. In periods when the
organic load is high or the water is very cold, one may reduce the organic load by using
precoagulation. The organic matter in the pretreated wastewater will partly consist of low
molecular weight, easily biodegradable organic matter that can denitrify a certain amount
of nitrate in the pre-denitrification at a moderate recirculation flow (0.25–0.5 times Q),
thus minimising oxygen recirculation. The rest of the nitrate is removed in the post-deni-
trification step where the measured effluent nitrate concentration controls the carbon
source addition.
H.Ø
deg
aard
27
Downloaded from https://iwaponline.com/wst/article-pdf/53/9/17/432865/17.pdfby gueston 08 November 2018
Because good documentation exists, we shall refer to the experiences from three Nor-
wegian treatment plants that are all designed according to the combined denitrification
scheme (see Figure 9). The design data for the plants are given in Table 3. Lillehammer
is operated slightly under design load, Nordre Follo at design load and Gardermoen
above the organic design load at times (under design load on nitrogen).
AN AN/AE
AE AE AE AN AE
Pre sedimentation Moving Bed Biofilm Process Flocculation FlotationChemCOD
Moving Bed Biofilm ProcessPre sedimentation
AN AN/AE
AE AE AE AN
COD
Flocculation
Chem
AEAE AN
(a)
(b)
Finalsedimentation
Figure 9 (a) Lillehammer WWTP, (b) Nordre Follo WWTP and Gardermoen WWTP
Table 3 Design data for the three Norwegian N-removing plants
Parameter Lillehammer Nordre Follo Gardermoen
Design valuesDesign flow (m3/h) 1,200 750 920Max.- flow (m3/h) 1,900 1,125 1,300
Design loads (kg/d)BOD7 2,900 2,660 2,680COD 5,925 5,900 5,480SS 2,900 4,390 3,290Tot N 755 460 704Tot P 107 – 101Temp. (8C) 3–14 6–14 4–14
Plant sizePre-settling area (m2) 600 696 420Tot. MBBR vol. (m3) 3,840 3,710 5,790Flocculation vol. (m3) 600 230 180Final sep. area (m2) 8601 1502 2152
Chemicals consumpt.Coagulantkg PAX/kg Premoved
26.8 17.0 17.5
Carbon sourcegCODadded/gNOx-Nequiv 4.25 (ethanol) 5.5 (methanol) 4.9 (glycol)Efficiency, 2002Average in-out conc. andtreatment efficiency
In%
Out In%
Out In%
Out
COD (mg/L) 386 32 452 30 583 3291 94 94
Tot N (mg/L) 34.6 4.5 30 8.0 51 1085 73 81
Tot P (mg/L) 4.38 0.10 4.86 0.27 7.01 0.1797 94 98
1Sedimentation2Flotation
H.Ø
deg
aard
28
Downloaded from https://iwaponline.com/wst/article-pdf/53/9/17/432865/17.pdfby gueston 08 November 2018
The Lillehammer plant was built in 1994 in connection with the winter Olympics.
In many ways it was a prototype for the recommended combined dentrification concept
of the moving bed process. Prior to 1994 the plant was a secondary precipitation plant
with the main target of removing phosphorous to prevent eutrophication of Norway’s lar-
gest lake (Mjoesa). In 1994 it was retrofitted to become a full nutrient removal plant with
nitrogen removal. The plant receives a typical town’s wastewater.
The Nordre Follo plant was originally a primary treatment plant that was upgraded to a
secondary plant based on chemical treatment in 1985 and retrofitted in 1997 with a biological
step that included nitrogen removal based on the moving bed process. The plant receives
wastewater from both residential and industrial areas that is known to be difficult to treat.
This plant was one of the early ones to use flotation for biomass separation (since 1985).
Since it was experienced that flotation went very well together with the moving bed
process, this separation process was also selected for the newest plant of the three, Garder-
moen WWTP, which is treating wastewater from residential areas as well as from Oslo
international airport Gardermoen. This includes waste de-icing liquid (monopropylene
glycol) at times, which is partly the reason for the much more concentrated wastewater at
this plant (see Table 3). Obviously, waste glycol is used as the external carbon source at
this plant while ethanol is used at Lillehammer and methanol at Nordre Follo. The effluent
standard at Gardermoen WWTP requires disinfection and the plant is equipped with UV
disinfection in addition to those unit processes that are depicted in Figure 9.
All the plants are divided into two lines and have pretreatment consisting of screening,
grit and fat removal and presettling (design overflow rate of 2.5m/h). The plants may be
operated with chemically enhanced primary treatment, if necessary, to reduce the organic
loading on the bioreactor and increase nitrification capacity at extraordinary low tempera-
tures or high organic loading (as during the winter Olympics at Lillehammer). The MBBR
of the oldest plant (Lillehammer WWTP) consists of nine reactors while Nordre Follo
WWTP and Gardermoen WWTP have seven reactors in series. The first reactor (two first
at Lillehammer) is anoxic, having mixer and no aeration. The next reactor is equipped
both with mixer and aerators. This gives operation flexibility. When more nitrification
volume is needed (in winter during low temperatures), the reactor will be operated with
aeration – less predenitrification volume will be available – and hence more carbon
source will have to be added in the post-denitrification step. The two larger reactors in the
middle are both aerobic (nitrification reactors) and equipped with aerators. Then follows
another combined reactor equipped with both aerator and mixer. The first one may be run
aerobically, however with mixing and no aeration (or very little aeration). This is to
consume as much oxygen as possible since this reactor is the one from which a fraction
(0.25–1.0, – depending on availability of carbon source in bioreactor inflow) of nitrified
water is returned to the predenitrification tank. If needed, for nitrification purposes, this
reactor can also be run with higher oxygen concentration but this will then increase the
demand for external carbon source in the post-denitrification step that follows in the
reactor train. This step consists of one reactor (two at Lillehammer) equipped with mixer
only. All the plants have a small, aerated reactor at the end of the reactor train for oxygen-
ation of the water as well as for removal of any residual biodegradable matter.
Especially, the Lillehammer plant has been extensively followed up and data from this
plant will be used to demonstrate the performance of the process. In Table 4 are shown
results from the guarantee test of the plant when operated in post-DN mode. Only one of
the two trains was operated resulting in loads close to or above the design load (Ødegaard
et al., 1999).
Figure 10(a) shows the total N removal efficiency for each day in the guarantee period
plotted versus total bioreactor residence time. The day to day removal of total N is
H.Ø
deg
aard
29
Downloaded from https://iwaponline.com/wst/article-pdf/53/9/17/432865/17.pdfby gueston 08 November 2018
shown in Figure 10(b), together with actual flow, actual total N load and the correspond-
ing design values.
Table 5 shows treatment results from the same plant, now operated in combined
denitrification mode at 60% of the design load but at very low temperatures. In 1999,
intensive performance tests were carried out at the plant in order to determine nitrifica-
tion- and denitrification capacities. Figure 11 shows the dependency of the nitrification
rates in reactor 4 and 5 on ammonium-concentration.
Very high nitrification rates were demonstrated, even at a tempeturere as low as 11 8C.
When calculated on reactor 4 alone, the maximum rate was around 1.5 g NH4-N/m2. d.
1 2 3 4 5
60
40
20
0
Actual TNremoval
100
80
60
40
20
0
Tot
al N
rem
oval
, %
Total empty bed HRT, hours
Design HRT: 3.5 hoursDesign TN-removal: 70%
TN
-load
(kg
N/d
)
Date, 1997
1300
1100
900
700
500
300
100
18/7 1/8 1/9 1/10 17/10
TN
rem
oval
(%
)
100
80
Design avg.TN removal
Design TNload
Actual TNload
(a)
(b)
Figure 10 Treatment results from the Lillehammer treatment plant at and above design load – when
operated in the post-denitrification mode. (a) Treatment efficiency versus total bioreactor residence time
(based on empty reactor); (b) Day to day removal efficiency of tot N at actual load compared to design load
Table 4 Total N removal at Lillehammer treatment plant operated at design load (Plant operated in post-
denitrification mode without precoagulation)
Bioreactor
residence time (h)
Influent tot
N conc. mg N/l
Effluent tot
N conc. mg N/l
Removal% Ethanol consumpt.
kg /kg Nrem.
Temperature 8C
Average 3.3 27 6.0 76 1.48 (3.1)1 13.7Minimum 1.6 16 2.9 48 0.55 (1.2)1 11.6Maximun 4.5 48 12.7 91 2.58 (5.4)1 15.7
1kg COD/kg Nremoved
H.Ø
deg
aard
30
Downloaded from https://iwaponline.com/wst/article-pdf/53/9/17/432865/17.pdfby gueston 08 November 2018
When calculated on both reactor 4 and 5 it was around 1.2 g NH4-N/m2 d, up to which
nitrification was close to complete.
Figure 12 plots the denitrification rate versus nitrate-load in the denitrification reactors
(7 and 8) during the same test. Even though this figure has only three points in it, they
give a fairly good demonstration of what is going on.
In the lower point, at a load of around 2 g NO3-Nequiv./m2d, denitrification is close to
complete because there is an excess of biodegradable, soluble COD. The second point
Lillehammer, train 1, Nov. 1999
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
2.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
Rem
oval
rat
e, g
NH
4-N/m
2 /d
R4
R4 + R5
Temp. 11˚C
Ammonium load, g NH4-N/m2/d
Figure 11 Nitrification rate versus ammonium load in reactor 4 and 5
Table 5 Treatment results from the Lillehammer plant when operated in the combination-DN mode at very
low temperatures
Temperature 8C Average influent inorg.
N conc., mg N/l
Average effluent inorg.
N conc., mg N/l
Removal
inorg. N %
Fraction of denitrification
in pre-DN, %
Average 6.3 17.2 3.1 92.0 16Minimum 6.0 16.1 2.2 74.5 15Maximum 6.5 17.7 4.1 87.6 17
0
1
2
3
4
0 1 2 3 4 5
Added C/N, g BSCOD/g NO3-Neq
DaytimeaveragesR7+R8
5.149
11.1
3.458
11.2
3.132
10.2
Rem
oval
rat
e, g
NO
3-Neq
/m2 /d
Load, g NO3-Neq/m2/d
Effl. FCOD, mg/LTemp.,˚C
Figure 12 Denitrification rate versus nitrate equivalents load in reactor 7 and 8
H.Ø
deg
aard
31
Downloaded from https://iwaponline.com/wst/article-pdf/53/9/17/432865/17.pdfby gueston 08 November 2018
demonstrates a situation where an insufficient amount of biodegradable, soluble COD
was added. This has obviously led to BSCOD limitation and reduced (around 60%)
denitrification. In the third point situation, even though the C/N-ratio is only slightly
higher, a very high denitrification rate (3.5 g NO3-Nequiv./m2d) was achieved, probably
because no BSCOD limitation prevailed.
Different external carbon sources give different denitrification rates. This is demon-
strated in Figure 13 in which denitrification rates versus temperature with various car-
bon sources are shown. These tests were carried out at Gardermoen WWTP (Rusten
et al., 1996).
Conclusions
The moving bed biofilm reactor (MBBR) has established itself as a well-proven, robust
and compact reactor for wastewater treatment. The efficiency of the reactor has been
demonstrated in many process combinations, both for BOD-removal and nutrient
removal. It has been used for small as well as large plants. Even though this paper
focused on municipal wastewater applications, the reactor has also been used extensively
for industrial wastewater treatment, particularly in the food industry and the pulp and
paper industry. Presently, there are more than 300 wastewater treatment plants in 22
countries based on the MBBR processes.
The primary advantage of the process, as compared to activated sludge reactors, is its
compactness and no need for sludge recirculation. The advantage over other biofilm pro-
cesses is its flexibility. One may use almost any reactor shape and one may choose different
operating loadings in a given reactor volume, simply by choice of carrier filling fraction.
In Scandinavia, the MBBR processes are normally combined with chemical P-removal
and this process combination may result in extremely compact plants high rate biomass/
floc separation processes (i.e. flotation) being used. The MBBR processes have favour-
ably been used for nitrification as well as for denitrification. High nitrification and denitri-
fication rates have been demonstrated even at low temperatures. When using the MBBR
process for nitrogen removal, a combined pre- and post-denitrification process is normally
recommended because this process combination is superior when it comes to process
control and performance.
ReferencesÆsøy, A., Ødegaard, H., Hægh, M., Risla, F. and Bentzen, G. (1998). Upgrading wastewater treatment plants
by the use of biofilm carriers, oxygen addition and pre-treatment in the sewer network. Wat. Sci. Tech.,
37(9), 159–166.
EthanolMethanolMonopropyleneglycol
4
3
2
1
04 6 8 10 12 14 16
Temperature, ˚C
Den
iitri
fica
tion
rate
g N
O-N
equi
v./m
2 d
Figure 13 DN-rates versus temperature obtained with various external carbon sources
H.Ø
deg
aard
32
Downloaded from https://iwaponline.com/wst/article-pdf/53/9/17/432865/17.pdfby gueston 08 November 2018
Helness, H. and Ødegaard, H. (2001). Biological phosphorous and nitrogen removal in a sequencing batch
moving bed biofilm reactor. Wat. Sci. Tech., 41(1), 233–240.
Hem, L., Rusten, B. and Ødegaard, H. (1994). Nitrification in a moving bed biofilm reactor. Water Res.,
28(6), 1425–1433.
Levine, A.D., Tchobanoglous, G. and Asano, T. (1985). Characterization of the size distribution of
contaminants in wastewater; Treatment and reuse implications. J. WPCF, 57(2), 805.
Mosey, F.E. (1996). Biofilms on Kaldnes carrier particles. Proc. from seminar organised by Anglian Water.
Cambridgeshire, UK 27–29 November 1996.
Ødegaard, H. (1992). Norwegian experiences with chemical treatment of raw wastewater. Wat. Sci. Tech.,
25(12), 255–264.
Ødegaard, H. (1998). Optimized particle separation in the primary step of wastewater treatment. Wat. Sci.
Tech., 37(10), 43–53.
Ødegaard, H., Rusten, B. and Westrum, T. (1994). A new moving bed bofilm reactor-Applications and
results. Wat. Sci. Tech., 29(10–11), 157–165.
Ødegaard, H., Rusten, B. and Siljudalen, J. (1999). The development of the moving bed biofilm
process–from idea to commercial product. Eur. Wat. Manage., 2(2).
Ødegaard, H., Gisvold, B. and Strickland, J. (2000). The influence of carrier size and shape in the moving
bed biofilm process. Wat. Sci. Tech., 41(4–5), 383–392.
Ødegaard, H., Melin, E., Helness, H. and Kenakkala, T. (2004). High-rate Wastewater treatment based on
moving bed biofilm reactors, polymer coagulation and flotation. In Chemical Water and Wastewater
Treatment VIII, Hahn, H.H., Hoffmann, E. and Ødegaard, H. (eds.), IWA Publishing.
Rusten, B. and Nedland, K.T. (2003). Upgrading of the TAU WWTP with a biological stage (In Norwegian).
Aquateam report no. 03–006.
Rusten, B. and Nedland, K.T. (2004). Pilot testing of biological-chemical treatment at the TAU WWTP
(In Norwegian). Aquateam report no. 04–008.
Rusten, B., Siljudalen, J.G. and Nordeidet, B. (1994). Upgrading to nitrogen removal with the KMT moving
bed biofilm process. Wat. Sci. Tech., 29(12), 185–195.
Rusten, B., Hem, L. and Ødegaard, H. (1995a). Nitrification of municipal wastewater in novel moving bed
biofilm reactors. Water Environ. Res., 67(1), 75–86.
Rusten, B., Hem, L. and Ødegaard, H. (1995b). Nitrogen removal from dilute wastewater in cold climate
using moving bed biofilm reactors. Water Environ. Res., 67(1), 65–74.
Rusten, B., Wien, A. and Skjefstad, J. (1996). Spent aircraft deicing fluid as external carbon source for
denitrification of municipal wastewater: From waste problem to beneficial use. 51. Purdue Ind. Waste
Conf. Proc., Ann Arbor Press, Inc. Chelsea, Michigan.
Rusten, B., McCoy, M., Proctor, R. and Siljudalen, J.G. (1998). The innovative moving bed biofilm reactor/
solids contact reaeration process for secondary treatment of municipal wastewater. Water Environ. Res.,
70(4), 1083–1089.
Stolz, J. (2004). Personal communication.
Tschui, M. (2004). Personal communication.
van Nieuwenhuijzen, A.F. and Mels, A.R. (2002). Characterization of particulate matter in municipal
wastewater. In Chemical Water and Wastewater Treatment VII, in Hahn, H.H., Hoffmann, E. and
Ødegaard, H. (eds.), IWA Publishing, London, pp. 203–212.
H.Ø
deg
aard
33
Downloaded from https://iwaponline.com/wst/article-pdf/53/9/17/432865/17.pdfby gueston 08 November 2018