elsevier (1)

Published on January 2017 | Categories: Documents | Downloads: 23 | Comments: 0 | Views: 321
of 14
Download PDF   Embed   Report

Comments

Content


Calibration of denitrifying activity of polyphosphate
accumulating organisms in an extended ASM2d model
F. Garcı´a-Usach
a
, J. Ribes
b,
*, J. Ferrer
a
, A. Seco
b
a
Instituto de Ingenierı´a del Agua y Medio Ambiente, Universidad Polite´cnica de Valencia, Camino de Vera s/n. 46022, Valencia, Spain
b
Departmento Ingenierı´a Quı´mica, Universitat de Vale`ncia, Doctor Moliner 50, 46100 Burjassot, Valencia, Spain
a r t i c l e i n f o
Article history:
Received 28 January 2010
Received in revised form
8 June 2010
Accepted 24 June 2010
Available online 3 July 2010
Keywords:
Anoxic yields
ASM2d
Biological phosphorus removal
Off-line calibration
Pilot plant study
a b s t r a c t
This paper presents the results of an experimental study for the modelling and calibration
of denitrifying activity of polyphosphate accumulating organisms (PAOs) in full-scale
WWTPs that incorporate simultaneous nitrogen and phosphorus removal. The conve-
nience of using different yields under aerobic and anoxic conditions for modelling bio-
logical phosphorus removal processes with the ASM2d has been demonstrated. Thus,
parameter h
PAO
in the model is given a physical meaning and represents the fraction of
PAOs that are able to follow the DPAO metabolism. Using stoichiometric relationships,
which are based on assumed biochemical pathways, the anoxic yields considered in the
extended ASM2d can be obtained as a function of their respective aerobic yields. Thus, this
modification does not mean an extra calibration effort to obtain the new parameters. In
this work, an off-line calibration methodology has been applied to validate the model,
where general relationships among stoichiometric parameters are proposed to avoid
increasing the number of parameters to calibrate. The results have been validated through
a UCT scheme pilot plant that is fed with municipal wastewater. The good concordance
obtained between experimental and simulated values validates the use of anoxic yields as
well as the calibration methodology. Deterministic modelling approaches, together with
off-line calibration methodologies, are proposed to assist in decision-making about further
process optimization in biological phosphate removal, since parameter values obtained by
off-line calibration give valuable information about the activated sludge process such as
the amount of DPAOs in the system.
ª 2010 Elsevier Ltd. All rights reserved.
1. Introduction
In recent years, the number of wastewater treatment plants
(WWTPs) designed for biological nutrient removal has been
increasing worldwide. Many research studies have focused on
giving more insight to biological nutrient removal processes
(BNR) over the last 20 years. Fromthe late 1980s, several studies
on activated sludge systems and laboratory cultures had
proven that PAOs can use nitrate as an electron acceptor in the
absence of molecular oxygen (Comeau et al., 1986, 1987; Vlekke
et al., 1988; Kerrn-Jespersen and Henze, 1993; Kuba et al., 1993;
Barker and Dold, 1996). However, it is still unclear whether the
same organisms are responsible for P removal under both
aerobic and anoxic conditions (Carvalho et al., 2007). Some
authors have suggested that PAOs could be divided into two
coexisting groups: denitrifying PAO (DPAO) and non-DPAO
* Corresponding author.
E-mail addresses: [email protected] (F. Garcı´a-Usach), [email protected] (J. Ribes), [email protected] (J. Ferrer), aurora.seco@uv.
es (A. Seco).
Avai l abl e at www. sci encedi r ect . com
j our nal homepage: www. el sevi er . com/ l ocat e/ wat r es
wa t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 5 2 8 4 e5 2 9 7
0043-1354/$ e see front matter ª 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.watres.2010.06.061
(Kerrn-Jespersen and Henze, 1993; Barker and Dold, 1996;
Meinhold et al., 1999; Freitas et al., 2005). DPAOs can utilize
either nitrate or oxygen as an electron acceptor, whereas non-
DPAOs can only use oxygen. Further, DPAOs have also shown
to be able to utilize nitrite as an electron acceptor despite it can
have inhibitory effect on the aerobic P-uptake (Sin et al., 2008).
In contrast, Zeng et al. (2003) suggested that PAOs and DPAOs
could be the same microorganisms and are able to change their
metabolism after a few hours of acclimatization. Recent
microbiological studies are focused on assessing the link
between the process operation conditions and their microbial
community structure under different conditions such as the
type of carbon source. Carvalho et al. (2007) found that different
PAO morphotypes have different behaviours, because they can
only use oxygen, or both nitrate and oxygen. Nevertheless, few
references can be found where full-scale or pilot plants are fed
with real wastewater (Hu et al., 2003).
Besides the microbiological studies, different modelling
approaches have beenproposed during the last fewyears. Most
of them are based on the assumption that, under anoxic
conditions, DPAOs are able to perform the same processes as
under aerobic conditions, although at lower rates (Kerrn-
Jespersen and Henze, 1993). This rate reduction is a conse-
quence of twopossiblehypotheses: (1) the minor energetic yield
obtained when nitrate, and not oxygen, is used as the electron
acceptor. When nitrate is used instead of oxygen as the final
electron acceptor of the oxidative phosphorylation, a lower
amount of ATP is generated and, consequently, a lower amount
of biomass is yielded (Kuba et al., 1996). (2) Only DPAOs are
capable of using nitrate as an electron acceptor.
ASM2d is one of the most widely used models for simu-
lating full-scale WWTPs that incorporate simultaneous
nitrogen and phosphorus removal. This model was developed
from ASM2 (Henze et al., 1995) by adding two new processes:
polyphosphate storage and growth of PAOs under anoxic
conditions (Henze et al., 1999). It uses the same kinetic equa-
tions as for the aerobic processes but corrected by a reduction
factor h
NO3
. This was explained by hypothesizing that either
only a fraction of the total PAOs are able to denitrify, or that
the anoxic activity of the entire PAO population is reduced by
a factor h
NO3
. The main drawback of ASM2d to represent the
denitrification of PAOs is that h
NO3
represents a simple
reduction in the kinetic expression for anoxic growth and
anoxic P-uptake, obviating any physical meaning of this
parameter. Many applications of ASM2d can be found in the
literature, where the same value of h
NO3
is used for anoxic P-
uptake and anoxic growth of PAOs (see e.g. Insel et al., 2006;
Makinia et al., 2006b; Yagci et al., 2006; Ruano et al., 2007).
Taking into account that deterministic models are more
useful for both process knowledge and decision-making in
diagnosis and optimization of WWTPs, several extensions
have been proposed to ASM2d, in an attempt to overcome this
drawback. On the one hand, some authors propose the use of
different h values for anoxic P-uptake and anoxic growth of
PAOs (Penya-Roja et al., 2002; Hu et al., 2003). Other authors
prefer to apply metabolic modelling approaches to give more
insight to the biological process that takes place in the acti-
vated sludge (Kuba et al., 1996; Murnleitner et al., 1997; Maurer
and Gujer, 1998). ASM2d and metabolic models have been
integrated together to simulate the behaviour of full-scale
EBPR plants, as is the case with the TUDP model (Meijer et al.,
2001) or with the EAWAG Bio-P module for activated sludge
model No. 3 (Rieger et al., 2001). When compared to conven-
tional ASM models, the combination of both metabolic and
Nomenclature
GAO glycogen accumulating organism
HAc acetic acid
NH
4
-N ammonia nitrogen
NO
3
-N nitrate nitrogen
NUR nitrate uptake rate (mg O
2
l
À1
min
À1
)
NUR
MAX,GROWTH
maximum specific nitrate uptake rate for
growth of PAOs (mg O
2
l
À1
min
À1
)
NUR
MAX,PeU
maximum specific nitrate uptake rate for
phosphorus uptake (mg O
2
l
À1
min
À1
)
OUR oxygen uptake rate (mg O
2
l
À1
min
À1
)
OUR
MAX,GROWTH
maximum specific oxygen uptake rate for
growth of PAOs (mg O
2
l
À1
min
À1
)
OUR
MAX,P-U
maximum specific oxygen uptake rate for
phosphorus uptake (mg O
2
l
À1
min
À1
)
PAO phosphorus accumulating organisms
PO
4
-P orthophosphate
r
pa
maximumspecific anoxic phosphorus uptake rate
(mg P l
À1
min
À1
)
r
po
maximum specific aerobic phosphorus uptake
rate (mg P l
À1
min
À1
)
SBOD soluble biological oxygen demand
SCOD soluble chemical oxygen demand
SRT solid retention time
TBOD total biological oxygen demand
TCOD total chemical oxygen demand
TN total nitrogen
TN
filtered
filtered total nitrogen
TP total phosphorus
TSS total suspended solids
TVS total volatile solids
VFA volatile fatty acids
VSS volatile suspended solids
WWTP wastewater treatment plant
X
DPAO
concentration of DPAO (mg CODl
À1
)
X
PAO
concentration of PAO (mg CODl
À1
)
X
PHA,DPAO
cell internal storage product for DPAOs. It
includes poly-hydroxyalkanoates (PHA)
(mg CODl
À1
).
X
PHA,PAO
cell internal storage product for PAOs. It includes
poly-hydroxyalkanoates (PHA) (mg CODl
À1
).
X
PP,DPAO
cell internal poly-phosphates for DPAOs (mg P l
À1
)
X
PP,PAO
cell internal poly-phosphates for PAOs (mg P l
À1
)
Y
PAO,NO
anoxic yield coefficient for DPAOs (gCODgCOD
À1
)
Y
PAO,O2
aerobic yield coefficient for PAOs (gCODgCOD
À1
)
Y
PHA,NO
PHA requirement for anoxic storage of phosphate
in the form of X
PP,DPAO
(gCODg
À1
P)
Y
PHA,O2
PHA requirement for aerobic storage of phosphate
in the form of X
PP,PAO
(gCODg
À1
P)
wa t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 5 2 8 4 e5 2 9 7 5285
ASM models enables an improved prediction of the process
performance. This combination also has the advantage of
requiring fewer parameters to be calibrated (Oehmen et al.,
2007). However, the complexity of these models may be
a main drawback for their application in practice.
The application of advanced models has been facilitated in
the recent years by the use of simulation platforms (Biowin
Ò
,
DESASS
Ò
, GPS-X
Ò
, WEST
Ò
, etc). With these programs, plant
operators can simulate their WWTPs and test different oper-
ation strategies to make decisions aimed at optimizing plant
performance (Ferrer et al., 2005). Also, new advances in
research can be transferred directly to technicians, designers,
and plant operators. More and more, plant operators are being
trained to use these platforms for decision-making about their
own WWTPs. However, the use of mathematical models
necessarily requires a prior calibration of the main kinetic and
stoichiometric parameters for each case study. Several cali-
bration protocols and guidelines have been proposed to cali-
brate the models (see e.g. Hulsbeek et al., 2002; Penya-Roja
et al., 2002; Vanrolleghem et al., 2003; Langergraber et al.,
2004; Makinia et al., 2006a). An important step of most of
these protocols consists of a series of batch tests to measure
the rate of either substrate or electron acceptor utilization.
The methodologies that include these batch tests are
currently known as off-line calibration methodologies since
they are carried out in the laboratory with the activated sludge
under controlled conditions. This allows the different
processes that are carried out by the different bacterial groups
considered by the model to be separated (Rieger et al., 2001;
Penya-Roja et al., 2002). Thus, the main kinetic and stoichio-
metric parameters can be obtained from the off-line calibra-
tion as well as a prior calibration of the steady state of the
plant. These parameters are then validated by steady state
simulations of the WWTP (Penya-Roja et al., 2002; Makinia
et al., 2006a). In case the model purpose requires it,
a dynamic model calibration can be performed to obtain an
accurate description of the plant performance (see, e.g. Sin
et al., 2005 for a detailed review on different calibration
protocols).
When model parameters are given a physical meaning, the
values obtained by off-line calibration provide valuable
information about the activated sludge process performance
(such as the amount of DPAOs in the system, the presence of
GAOs, etc). Thus, decision-making about further process
optimization that is carried out by simulation studies should
be based on deterministic parameter values obtained by off-
line calibration methodologies. These calibration methodolo-
gies are mostly based on respirometric techniques in order to
make them easy to implement in the laboratories that are
commonly available in a WWTP.
The principal aim of this work is to validate a modelling
approach that is based on using different yields for aerobic
and anoxic conditions in the ASM2d model. An off-line cali-
bration methodology developed for full-scale WWTPs that
incorporate simultaneous nitrogen and phosphorus removal
has been applied to validate the model. It is based on using
stoichiometric relationships between aerobic and anoxic
yields in order to simplify the off-line experimental calibra-
tion. The proposed model has been applied to a pilot plant
operated as a UCT scheme which was fed with municipal raw
wastewater coming from the screening of Pinedo WWTP
(Valencia, Spain). It has been validated for two steady states
under different sludge retention times, and the parameter
values obtained from the off-line calibration have been used
to simulate the pilot plant performance.
2. Materials and methods
2.1. Pilot plant operation and characterization
In order to obtain an activated sludge with a stable PAO pop-
ulation, a pilot plant was operated under a UCT configuration.
The pilot plant, which is represented in Fig. 1, was fed with
40 l h
À1
of municipal raw wastewater from the screening of
the Pinedo WWTP (Valencia, Spain). The plant was operated at
a constant controlled temperature of 20 Æ1

C and two
different sludge ages (9.8 and 5.7 d) until steady state condi-
tions were reached. Analytical monitoring was used to
conclude that steady state was reached. In the first experi-
mental phase (SRT¼9.8 d), the pilot plant was operated
during 119 d, whereas the second phase (SRT¼5.7 d) was
maintained for 42 d. The high length of the first phase was due
to several operational problems that caused disturbances in
the process. In order to prevent the GAO metabolism from
appearing in the system, several parameters were controlled:
the influent COD/P ratio was controlled below 30 (Liu et al.,
1997; Mino et al., 1998); the two operation SRTs were
selected to be under 10 d (Rodrigo et al., 1999; Whang and Park,
2006); and oxygen concentration was controlled at 2 mg l
À1
in
the aerobic reactor (Griffiths et al., 2002). Moreover, pH was
above 7.5 throughout the entire experimental period, which
Fig. 1 e Pilot plant configuration and operation conditions.
wa t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 5 2 8 4 e5 2 9 7 5286
favours the metabolism of PAO over GAO at 20

C (Lopez-
Vazquez, et al., 2009).
In order to evaluate the steady state conditions, daily
wastewater composite samples were collected by automatic
pumps placed at sampling points I and V (see Fig. 1). These
composite samples were kept at 4

Cina refrigerator during the
samplingperiod. Theactivatedsludge(samplingpointsII, III and
IV) was analyzed for the main parameters three times a week.
Once the steady state was reached, an exhaustive process
characterization was carried out. Composite and grab samples
were collected over two weeks in order to characterize the
pilot plant performance. The following parameters were
analyzed for the influent and effluent streams: total sus-
pended solids (TSS), volatile suspended solids (VSS), total and
soluble chemical oxygen demand (TCOD and SCOD), total and
soluble biological oxygen demand (TBOD and SBOD), volatile
fatty acids (VFA), ammonia nitrogen (NH
4
-N), nitrate nitrogen
(NO
3
-N), total nitrogen (TN), filtered total nitrogen (TN
filtered
),
orthophosphate (PO
4
-P), total phosphorus (TP), and alkalinity
(Alk). For the characterization of the activated sludge TSS,
VSS, TCOD, VFA, NH
4
-N, NO
3
-N, TN
filtered
, PO
4
-P, and TP were
determined. Nitrite was negligible in the activated sludge
during the experimental period. The average influent waste-
water characteristics for the two steady states are shown in
Table 1. The effluent and activated sludge characteristics for
each steady state are compared with the pilot plant simula-
tion results in Section 3.2 (see Tables 7 and 8). Acetic acid was
added to the influent wastewater in order to compensate for
the low COD in the influent.
2.2. Analytical methods
Most of the analytical techniques used in this work were in
accordance with Standard Methods (APHA, 2005) (TSS, VSS,
COD, NH
4
-N, TP, PO
4
-P). Specific MERCK kits were used to
determine COD (10e125 mg l
À1
), NO
3
-N, and TN. The
carbonate alkalinity and VFA were determined by titration
using the method proposed by WRC (1992). BOD was deter-
mined by pressure measurements using a WTW OXITOP
CONTROL system. In order to verify the absence of GAOs in
the system, Fluorescence in situ hybridisation (FISH) was
performed as described in Amann (1995) with the specific
probes for GAOs proposed in Crocetti et al. (2002).
2.3. Modelling the denitrifying activity of PAOs
In this work, an extension of ASM2d model has been proposed
based on using different yields for aerobic and anoxic condi-
tions in order to account for the less energetic efficiency of
microorganisms under anoxic conditions. Thus, the same
value of h
PAO
can be used for anoxic growth of PAO and anoxic
P-uptake. This is based on the assumption that the reduction
factor (h
PAO
) represents the proportion of X
PAO
that follows the
DPAO metabolism. This assumption differs from ASM2d,
where h
NO3
represents a simple reduction in the kinetic
expression for anoxic growth and anoxic P-uptake, obviating
any physical meaning of this parameter. Hence, the yields for
processes carried out by PAOs under anoxic and aerobic
conditions are represented by Y
PAO,NO
and Y
PHA,NO
, and Y
PAO,O2
and Y
PHA,O2
, respectively. In this approach metabolic consid-
erations explained below, which are based on biochemical
pathways proposed elsewhere (Smolders et al., 1994 and Kuba
et al., 1996), are used to obtain the anoxic yields as a function
of their respective aerobic yields, so that their experimental
calibration is not needed.
2.3.1. Metabolic approach for Y
PAO,NO
and Y
PHA,NO
Anoxic yields (Y
PAO,NO
and Y
PHA,NO
) are obtained from their
respective aerobic yields, based on the metabolic pathways
proposed by Smolders et al. (1994) and Kuba et al. (1996) for
PAO metabolism under aerobic and anoxic conditions,
respectively. According to these authors, both aerobic and
anoxic metabolism can be described by six metabolic
processes: PHB degradation within the TCA cycle; growth and
maintenance of biomass; polyphosphate formation; glycogen
formation; ATP production from NADH
2
(also known as elec-
tron transport phosphorylation); and phosphate uptake. Only
the last two processes depend on the type of electron
acceptor. The ATP production from NADH
2
under aerobic and
anoxic conditions is described by the following reactions
(Smolders et al., 1994; Kuba et al., 1996):
Aerobic : NADH
2
þ0:5O
2
/d
o
ATP þH
2
O (1)
Anoxic : NADH
2
þ
2
5
HNO
3
/
1
5
N
2
þd
n
ATP þ
6
5
H
2
O (2)
where d
o
is the amount of ATP produced per NADH
2
oxidized
in the electron-transport chain, and where d
n
is the amount of
ATP that can be produced per NADH
2
oxidized with nitrate as
the electronacceptor. Kuba et al. (1996) found a meanvalue for
d
n
of 1.0 for DPAOs under anoxic conditions. This value is
lower than the d
o
value proposed by Smolders et al. (1994) for
PAOs under aerobic conditions (1.85 mole ATP mole
À1
NADH
2
), which indicates the lower energetic efficiency of
DPAOs.
Taking into account the stoichiometry of Eqs. (1) and (2),
and assuming the abovementioned oxidative phosphoryla-
tion efficiencies, it can be deduced that the amount of electron
acceptor that is required to produce 1 mol of ATP under
aerobic conditions is 8.65 g O
2
mol
À1
ATP, and under anoxic
conditions is 5.6 g Nmol
À1
ATP.
Table 1 e Average influent wastewater characteristics
obtained for the two steady states considered in this
study.
Parameter Unit SRT¼9.8 d SRT¼5.7 d
TSS mg l
À1
57 Æ6 68 Æ9
VSS mg l
À1
41 Æ3 54 Æ6
TCOD mg CODl
À1
265 Æ11 310 Æ13
SCOD mg CODl
À1
157 Æ8 203 Æ7
TBOD mg CODl
À1
216 Æ4 246 Æ5
SBOD mg CODl
À1
150 Æ3 175 Æ4
VFA mg CODl
À1
83 Æ2 93 Æ2
NO
3
-N mg Nl
À1
<0.2 <0.2
NH
4
-N mg Nl
À1
26.2 Æ1.8 23.5 Æ0.6
TN mg Nl
À1
33.9 Æ7.4 35.2 Æ6.2
TN
filtered
mg Nl
À1
29.5 Æ6.3 28.0 Æ5.0
TP mg P l
À1
8.7 Æ1.2 9.7 Æ0.4
PO
4
-P mg P l
À1
6.8 Æ0.7 8.0 Æ0.6
Alk mg CaCO
3
l
À1
353 Æ7 385 Æ8
wa t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 5 2 8 4 e5 2 9 7 5287
Taking into account the definition of Y
PAO,O2
and Y
PAO,NO
,
the following relations can be obtained:
1 ÀY
PAO;O
2
Y
PAO;O
2

oxygen consumption
biomass production
¼
g O
2
mol ATP
$
mol ATP
g biomass
(3)
_
1 ÀY
PAO;NO
Y
PAO;NO
_
,
1
2:86 g O
2
=g N

nitrogen consumption
biomass production
¼
g N
mol ATP
$
mol ATP
g biomass
(4)
By dividing these two expressions, the following relation-
ship can be obtained between Y
PAO,O2
and Y
PAO,NO
:
_
1 ÀY
PAO;O
2
Y
PAO;O
2
_
_
1 ÀY
PAO;NO
Y
PAO;NO
_
¼
8:65 g O
2
=mol ATP
5:6 g N=mol ATP$2:86 g O
2
=g N
(5)
Similarly for Y
PHA,O2
and Y
PHA,NO
, the following expressions
can be obtained:
Y
PHA;O
2

oxygen consumption
phosphorus storage
¼
g O
2
mol ATP
$
mol ATP
g P storage
(6)
Y
PHA;NO
$
1
2:86g O
2
=g N

nitrogen consumption
phosphorus storage
¼
g N
mol ATP
$
mol ATP
g P storage
(7)
And then, the relationship between Y
PHA,O2
and Y
PHA,NO
is
obtained as:
Y
PHA;O
2
Y
PHA;NO
¼
8:65 g O
2
=mol ATP
5:6 g N=mol ATP$2:86 g O
2
=g N
(8)
Consequently, from Eqs. (5) and (8), the anoxic yields Y
PAO,NO
and Y
PHA,NO
can be obtained from the aerobic parameters as:
Y
PAO;NO
¼
_
1 þ
_
1-Y
PAO;O
2
_
Y
PAO;O
2
$
5:6
8:65
$2:86
_
À1
(9)
Y
PHA;NO
¼
5:6
8:65
$2:86$Y
PHA;O
2
(10)
2.3.2. Metabolic approach for Y
H,NO
Using the same stoichiometry of oxygen and nitrate require-
ments for ATP production as that proposed for PAOs (see Eqs.
(1) and (2)) and applying the same procedure as above for
anoxic growth of heteretrophs, the following relationship can
be obtained between Y
H,NO
and Y
H,O2
:
Y
H;NO
¼
_
1 þ
_
1 ÀY
H;O
2
_
Y
H;O
2
$
5:6
8:65
$2:86
_
À1
(11)
Hence, for a typical value of 0.67 g CODg
À1
COD, which is
conventionally adopted in ASM1 and similar models for acti-
vated sludge systems treating municipal type wastewaters, this
relationship gives an anoxic yield of 0.52 g CODg
À1
COD and an
anoxic to aerobic yield ratio of 0.78. This ratio is practically the
same as the one obtained by many other authors (see Muller
et al., 2003 for an extensive review of these ratios).
The experimental calibration methodology applied to
Y
H,O2
, Y
PAO,O2
, and Y
PHA,O2
as well as the rest of the high
influential parameters in the model is explained below.
2.4. Experimental calibration methodology
The calibration methodology is based on experimental batch
tests that were carried out with biomass and wastewater from
the WWTP. It consists of a selective calibration of highly
influential parameters by means of an individual analysis of
the different processes that are carried out by heterotrophic,
nitrifying, and polyphosphate accumulating organisms
(Penya-Roja et al., 2002). The experimental off-line calibration
methodology consisted of eight different experiments, which
are listed in Table 2. Sensitivity analysis was previously
carried out to identify the highly influential parameters in
each experiment and also to establish the proper experi-
mental conditions that maximize the sensitivity of the model
outputs to these parameters and minimize the sensitivity to
the rest of parameters that are not calibrated by these
experiments. Once the most important parameters were
estimated on the basis of batch experiments, the parameter
values were validated by simulating the plant in order to fit
the experimental data. Dynamic sensitivity analyses were
carried out for the batch tests after the calibration in order to
verify that the calibratedparameters were the most influential
ones for the initial conditions used in the experiments. The
exact sensitivity functions were obtained as proposed in Ribes
et al. (2004) and the methodology proposed in Ruano et al.
(2007) was applied to obtain the most influential parameters.
All the experiments were carried out with biomass samples
from the aerobic reactor under endogenous conditions. In
order to achieve endogenous conditions, biomass was first
exposed to aerobic conditions in the laboratory until complete
depletion of PHA reserves is reached. In order to identify the
endogenous conditions, OUR was continuously registered by
controlling the oxygen concentration between 2 and 4 mg l
À1
with oneoff aeration cycles. When necessary, phosphate was
added throughout the process (15e20 mg P l
À1
) to keep it from
becoming a limiting factor (see Penya-Roja et al., 2002).
Thefirst experiment (heterotrophicyield) wascarriedout by
measuring the total amount of oxygen consumed to complete
oxidation of a known amount of substrate. An amount of
filtered wastewater (soluble COD) was added to a batch reactor
with endogenous biomass, where the exogenous OUR was
continuously registered until COD was degraded. COD
degraded
was determined by means of filtered COD analysis before and
Table 2 e Experiments carried out and parameters
obtained in the off-line experimental calibration.
Microorganisms Experiment Parameters
obtained
Heterotrophs 1. Heterotrophic
yield
Y
H,O2
2. Monod curve m
H
, K
F
3. Aerobic and
anoxic batch
h
H
Autotrophs 4. Monod curve m
AUT
, K
NH4
PAOs 5. Anaerobic batch Y
PO4
, q
PHA
, K
A
6. Aerobic batch Y
PAO,O2
, Y
PHA,O2
, m
PAO
,
q
PP
, K
PHA
, K
IPP
, K
MAX
7. Aerobic and
anoxic batch
h
PAO
(check for Eqs. (9)
and (10)
All 8. Endogenous
respiration
b
H
, b
AUT
, b
PAO
, b
PP
, b
PHA
wa t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 5 2 8 4 e5 2 9 7 5288
after the experiment. Y
H,O2
was then obtained as (COD
de-
graded À !OURexo
$dt)/COD
degraded
. Thecorresponding anoxic yield
(Y
H,NO
) was then calculated from the aerobic one by using the
stoichiometric relationship shown in Eq. (11).
The reduction factor for heterotrophic denitrification
(experiment 3) was calculated from the maximum nitrate
uptake rate (NUR
MAX
) and the maximum oxygen uptake rate
(OUR
MAX
), which were obtained in two parallel batch reactors
after the addition of the same amount of filtered wastewater
to each reactor. The first batch was maintained under anoxic
conditions by adding NO
3
-N at the beginning, while the
second batch, which was maintained under aerobic condi-
tions, was the same batch as the one from experiment 1.
Monod curves for heterotrophs and autotrophs (experi-
ments 2 and 4) were determined by following the method
described by Cech et al. (1984). Different amounts of substrate
were added to endogenous sludge in order to achieve their
respective maximum exogenous respiration rates at different
substrate concentrations (up to 40 mg CODl
À1
for hetero-
trophs and 2.5 mg NH
4
-Nl
À1
for autotrophs). The added
substrates consisted of either filtered wastewater for the
heterotrophic biomass (experiment 2) or a known concentra-
tion of NH
4
-N solution for the autotrophic bacteria (experi-
ment 4). Optimal experimental design was carried out as
described in Ribes et al. (2004) in order to maximize the
sensitivity of parameters obtained from these experiments.
Nitrification process was inhibited in experiments 1e3
with the addition of thiourea to a concentration of 20 mg l
À1
in
order to compute OUR exclusively coming from COD oxida-
tion. Further details on experiments 1e4 can be found in
Penya-Roja et al. (2002). The autotrophic yield coefficient (Y
A
)
was not determined in this methodology because the value of
0.24 g CODg
À1
N is generally accepted for autotrophs (Henze
et al., 1999; Vanrolleghem et al., 1999).
For experiment 5, acetic acid (HAc) was added to the acti-
vated sludge to achieve an initial concentration around
100 mg CODl
À1
and the system was allowed to evolve in
anaerobic conditions. Time evolutions of HAc and phosphorus
concentrations were measured in order to obtain the param-
eters Y
PO4
, q
PHA
, and K
A
. Then Y
PO4
was directly obtained from
the linear fitting of S
PO4
vs. S
A
, whereas q
PHA
and K
A
were
obtained by non-linear parameter estimation based on least
squares method to fit S
A
and S
PO4
simulations to experimental
data (see Figs. 2 and 3).
For experiment 6, once acetate was exhausted from
experiment 5, the sludge sample was split into two fractions
(about 50%each) and then one fractionwas exposed to aerobic
conditions and the other fraction to anoxic conditions. When
necessary, phosphate was added to keep it from becoming
a limiting factor in P-uptake process. In the aerobic fraction
the OUR and phosphorus concentration were monitored until
the reaction was over (see Fig. 4). The oxygen consumed in
0
10
20
30
40
50
60
0
20
40
60
80
100
120
140
a
b
0 20 40 60 80 100
S
4
O
P
l
P
g
m
(
1
-
)
S
A
l
D
O
C
g
m
(
1
-
)
t (min)
SRT = 9.8 d
SA
SA sim
SPO4
SPO4 sim
y = -0.42x + 55.05
R² = 0.99
0
10
20
30
40
50
60
0 20 40 60 80 100 120
S
4
O
P
l
P
g
m
(
1
-
)
SA (mg COD l
-1
)
SRT = 9.8 d
Fig. 2 e Acetic acid and phosphorus experimental data measured in the anaerobic batch experiment for calibration 1
(SRT[9.8 d). (a) Curve fittings for q
PHA
and K
A
estimation; and (b) linear fitting for obtaining Y
PO4
.
0
10
20
30
40
50
a
b
0
20
40
60
80
100
0 20 40 60 80 100
S
4
O
P
l
P
g
m
(
1
-
)
S
A
l
D
O
C
g
m
(
1
-
)
t (min)
SRT = 5.7 d
SA exp
SA sim
SPO4 exp
SPO4 sim
y = -0.43x + 42.36
R² = 0.99
0
10
20
30
40
50
0 20 40 60 80 100
S
4
O
P
l
P
g
m
(
1
-
)
S
A
(mg COD l
-1
)
SRT = 5.7 d
Fig. 3 e Acetic acid and phosphorus experimental data measured in the anaerobic batch experiment for calibration 2
(SRT[5.7 d). (a) Curve fittings for q
PHA
and K
A
estimation; and (b) linear fitting for obtaining Y
PO4
.
wa t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 5 2 8 4 e5 2 9 7 5289
each process (aerobic growth and P-uptake) was determined
by measuring OUR together with the amount of phosphorus
that was removed from the medium. Similarly, anoxic batch
tests were carried out in parallel with these aerobic ones
(experiment 7) in order to obtain the anoxic reduction factor
(h
PAO
) (see Fig. 5). For the anoxic batch tests, nitrate was added
to obtain an initial concentration above 15 mg NO
3
-Nl
À1
in
order to avoid limited nitrate concentration.
Finally, an endogenous respiration test (experiment 8) was
carried out in order to obtain the global endogenous OUR in
the system, which is related to the decay coefficients for each
biomass fraction (Penya-Roja et al., 2002). According to Penya-
Roja et al. (2002), the different fractions of heterotrophs,
autotrophs, and PAOs in the active biomass (X
H
, X
A
and X
PAO
),
as well as their rate constants for lysis (b
H
, b
AUT
and b
PAO
) are
determined by using all the data obtained from the previous
experiments and taking into account that the sum of X
H
, X
A
and X
PAO
must fulfil the mass balance applied to the pilot
plant. The same value as the one obtained for b
PAO
is given to
b
PP
and b
PHA
. An optimization algorithm programmed in
Matlab
Ò
was applied for this data fitting, which is based on the
least squares method. In order to avoid local minima prob-
lems, a search method based on genetic algorithms was
implemented in the optimization algorithm.
Based on these laboratory experiments, the calibration
methodology for the denitrifying activity of PAOs was divided
into three steps. First, the aerobic yields (Y
PAO
and Y
PHA
), as
well as the other PAO kinetic parameters, were obtained by
fitting experimental values (S
PO4
and OUR) from the aerobic
batch test (experiment 5), which was drawn out until phos-
phate concentration remained constant (see Fig. 4). Second,
anoxic yields (Y
PAO, NO
and Y
PHA, NO
) were calculated from the
aerobic ones using the stoichiometric relationships shown in
Eqs. (9) and (10). Third, in order to obtain the anoxic reduction
factor, the following measurements were performed: the
maximal oxygen uptake rate (OUR
MAX
) and the maximal
specific aerobic phosphorus uptake rate (r
po
) were determined
by an aerobic batch test; the maximal nitrate uptake rate
(NUR
MAX
) and the maximal specific anoxic phosphorus uptake
rate (r
pa
) were determined by an anoxic batch test. A detailed
description of the last step of this methodology is explained
below.
2.4.1. Determination of the anoxic reduction factor (h
PAO
)
The calibration of the anoxic reduction factor and the yields
for PAOs is based on the assumption that h
PAO
exclusively
represents the fraction of denitrifying PAOs (DPAOs); there-
fore, different yields need to be assessed (aerobic and anoxic)
for the different fractions of PAOs. Hence, h
PAO
is defined as
X
DPAO
/X
PAO
; and thus X
DPAO
¼h
PAO
$X
PAO
. With this assump-
tion, the parameter h
PAO
can be calibrated by conducting two
phosphorus uptake batch tests: one test with oxygen (the
same as for calibrating Y
PAO
) and the other test with nitrate as
electron acceptors. First, the endogenous biomass is exposed
0
10
20
30
40
50
60
0
200
400
600
800
1000
1200
a
b
0.0 1.0 2.0 3.0 4.0
S
4
O
P
l
P
g
m
(
1
-
)
O
g
m
(
R
U
O
2
l
1
-
d
1
-
)
t (h)
SRT = 9.8 d
OUR exp.
OUR sim
SPO4 exp
SPO4 sim
0
10
20
30
40
50
60
70
80
0
200
400
600
800
1000
0.0 1.0 2.0 3.0 4.0
S
4
O
P
l
P
g
m
(
1
-
)
O
g
m
(
R
U
O
2
l
1
-
d
1
-
)
t (h)
SRT = 5.7 d
OUR exp.
OUR sim
SPO4 exp
SPO4 sim
Fig. 4 e S
PO4
and OUR experimental data and fitting in the aerobic batch experiment for (a) Calibration 1 and (b) Calibration 2.
0
10
20
30
40
50
60
a
b
S
4
O
P
l
P
g
m
(
1
-
)
t (min)
SRT = 9.8 d
SPO4 anox.
SPO4 aer. r
pa
r
po
0
15
30
45
60
75
90
0 50 100 150 200 250 0 50 100 150 200 250
S
4
O
P
l
P
g
m
(
1
-
)
t (min)
SRT = 5.7 d
SPO4 anox.
SPO4 aer.
r
pa
r
po
Fig. 5 e S
PO4
experimental data and maximal phosphorous uptake rates under anoxic and aerobic conditions obtained from
batch experiments for (a) Calibration 1 and (b) Calibration 2.
wa t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 5 2 8 4 e5 2 9 7 5290
to anaerobic conditions after the addition of HAc in a phos-
phorus release batch reactor. When this substrate is
completely consumed, the sludge is divided into two parts.
One of them is exposed to aerobic conditions, whereas the
other one is maintained in anoxic conditions. In the aerobic
batch reactor, OUR and phosphate concentration are period-
ically determined, whereas in the anoxic test, nitrate and
phosphate concentrations are measured. Similar procedures
can be found elsewhere to determine the proportion of deni-
trifying dephosphatation activity based on P-uptake rates and
NUR measurements (see, e.g. Wachtmeister et al., 1997 and
Sin et al., 2008).
The following differential equations canbe applied for both
phosphorus uptake rate (mg P l
À1
min
À1
) and oxygen or nitrate
uptake rate (mg O
2
l
À1
min
À1
) under aerobic or anoxic condi-
tions, respectively.
Phosphorus and oxygen uptake rates under aerobic conditions
(with X
PAO
)
À
_
dS
PO4
dt
_
AER
¼
_
q
PP
$
X
PHA;PAO
X
PAO
K
PHA
þ
X
PHA;PAO
X
PAO
$
K
MAX
À
X
PP;PAO
X
K
IPP
þ
_
K
MAX
À
X
PP;PAO
X
PAO
_$X
PAO
_
(13)
Phosphorus and nitrate uptake rates under anoxic conditions
(with X
DPAO
)
At the initial conditions in both batch experiments, when
phosphorus uptake rates, NUR, and OUR are maximum, X
PHA
is also maximum and X
PP
is minimum. Therefore, neither
X
PHA
nor X
PP
are limiting the process rates, and, thus, the
switching functions in Eqs. (13)e(16) are close to the unit.
Taking into account that X
DPAO
¼h
PAO
$X
PAO
, it follows that the
ratio between the maximal phosphorous uptake rates under
anoxic (r
pa
) and aerobic (r
po
) conditions is equal to h
PAO
.
À
_
dS
PO
4
dt
_
ANOX;MAX
À
_
dS
PO
4
dt
_
AER;MAX
¼
r
pa
r
po
¼
q
PP
$h
PAO
$X
PAO
q
PP
$X
PAO
¼ h
PAO
(17)
Similarly, by using the ratio between the maximal oxygen
uptake rate that is due to the growth of PAOs under aerobic
conditions (first term in Eq. (14)) and the maximal nitrate
uptake rate that is due to the growth of DPAOs under anoxic
conditions (first term in Eq. (16)) and taking into account the
relationship between Y
PAO,O2
and Y
PAO,NO
given by Eq. (5), the
following equation is obtained for h
PAO
:
À
_
dS
NO
3
dt
Â2:86
_
MAX;GROWTH
À
_
dS
O
2
dt
_
MAX;GROWTH
¼
NUR
MAX;GROWTH
Â2:86
OUR
MAX;GROWTH
¼
ð1-Y
PAO;NO
Þ
Y
PAO;NO
$m
PAO
$h
PAO
$X
PAO
_
1-Y
PAO;O
2
_
Y
PAO;O
2
$m
PAO
$X
PAO
¼
5:6
8:65
$2:86$h
PAO
(18)
The parameter h
PAO
can be obtained from either Eq. (17) or
(18). If Eq. (17) is used, only the S
PO4
experimental data
from aerobic and anoxic batch tests are needed. However, as
NUR
MAX,GROWTH
and OUR
MAX,GROWTH
cannot be measured
directly from laboratory experiments, using Eq. (18) to obtain
h
PAO
requires not onlyNURandOURexperimental data, but also
S
PO4
fromaerobicandanoxic batchtests inorder todiscriminate
the electron acceptor consumption in both growth and phos-
phorous uptake processes. Therefore, NUR
MAX,GROWTH
can be
obtainedfromNUR
MAX
andr
pa
(Eq. (19)); andOUR
MAX,GROWTH
can
be obtained from OUR
MAX
and r
po
(Eq. (20)).
NUR
MAX;GROWTH
¼ NUR
MAX
ÀNUR
MAX;PÀU
¼ NUR
MAX
ÀY
PHA;NO
$r
pa
(19)
Table 3 e Parameter values for heterotrophs and
autotrophs obtained by experimental calibration in the
two steady states.
Parameter Cal. 1 Cal. 2 Parameter Cal. 1 Cal. 2
Y
H,O2
(g CODg
À1
COD)
0.58 0.54 h
H
0.14 0.27
Y
H,NO
(g CODg
À1
COD)
0.43 0.39 m
AUT
(d
À1
) 1.14
a
m
H
(d
À1
) 1.38 2.10 K
NH4
(g Nm
À3
) 0.23
a
K
F
(g CODm
À3
) 0.84 3.95 b
AUT
(d
À1
) 0.20
a
b
H
(d
À1
) 0.39 0.58
a Autotrophic microorganisms were not present.
À
_
dS
O
2
dt
_
AER
¼
_
_
1 ÀY
PAO;O
2
_
Y
PAO;O
2
$m
PAO
$
X
PHA;PAO
X
PAO
K
PHA
þ
X
PHA;PAO
X
PAO
$X
PAO
þY
PHA;O
2
$q
PP
$
X
PHA;PAO
X
PAO
K
PHA
þ
X
PHA;PAO
X
PAO
$
K
MAX
À
X
PP;PAO
X
K
IPP
þ
_
K
MAX
À
X
PP;PAO
X
PAO
_$X
PAO
_
(14)
À
_
dS
NO
3
dt
Â2:86
_
ANOX
¼
_
ð1-Y
PAO;NO
Þ
Y
PAO;NO
$m
PAO
$
X
PHA;DPAO
X
DPAO
K
PHA
þ
X
PHA;DPAO
X
DPAO
$X
DPAO
þY
PHA;NO
$q
PP
$
X
PHA;DPAO
X
DPAO
K
PHA
þ
X
PHA;DPAO
X
DPAO
,
K
MAX
À
X
PP;DPAO
X
K
IPP
þ
_
K
MAX
À
X
PP;DPAO
X
DPAO
_$X
DPAO
_
(16)
À
_
dS
PO4
dt
_
ANOX
¼
_
q
PP
$
X
PHA;DPAO
X
DPAO
K
PHA
þ
X
PHA;DPAO
X
DPAO
,
K
MAX
À
X
PP;DPAO
X
DPAO
K
IPP
þ
_
K
MAX
À
X
PP;DPAO
X
DPAO
_$X
DPAO
_
(15)
wa t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 5 2 8 4 e5 2 9 7 5291
OUR
MAX;GROWTH
¼ OUR
MAX
ÀOUR
MAX;PÀU
¼ OUR
MAX
ÀY
PHA;O
2
$r
po
(20)
Hence, by Eq. (18), both stoichiometric relationships
between aerobic and anoxic yields (Y
PAO,NO
vs. Y
PAO,O2
and
Y
PHA,NO
vs. Y
PHA,O2
) are used in order to obtain h
PAO
.
In this work, both Eqs. (17) and (18) were used to test the
correctness of the calibration methodology. Similar results
from these two equations are expected if the stoichiometric
relationships that are proposed in Eqs. (9) and (10) are correct
and if h
PAO
exclusively represents the fraction of denitrifying
PAOs.
The same procedure can be applied for anoxic growth of
heterotrophs. As different yields are also used for anoxic and
aerobic growth, h
H
will represent the fraction of X
H
that can
grow under anoxic conditions. By using OUR
MAX
and NUR
MAX
data for heterotrophs obtained from experiment 3, h
H
can be
obtained as:
À
_
dS
NO
3
dt
Â2:86
_
0
À
_
dS
O
2
dt
_
0
¼
NUR
MAX
Â2:86
OUR
MAX
¼
ð1-Y
H;NOÞ
Y
H;NO
$m
H
$h
H
$X
H
ð1-Y
H;O
2
Þ
Y
H;O
2
$m
H
$X
H
¼
5:6
8:65
$2:86$h
H
(21)
3. Results and discussion
3.1. Calibration results
The calibration methodology was applied to the UCT pilot
plant at the two steady states obtained with different sludge
ages. The results obtained for calibration carried out with
sludge from 9.8 d steady state are referred to as “Calibration
1”; and the results obtained for 5.7 d steady state are referred
to as “Calibration 2”. Once each steady state was reached, the
experiments shown in Table 2 were conducted in the labora-
tory using biomass and wastewater from the pilot plant.
However, since the main objective is focused on the calibra-
tion of denitrifying activity of PAOs, only the results for
experiments 5, 6, and 7 are presented in detail below. For the
rest of experiments, only the parameter values obtained for
both calibrations are shown in Table 3. As can be deduced
from Table 3, autotrophs were not present when SRT was
reduced to 5.7 d in the pilot plant. This confirms that the high
N-NH
4
values observed in the effluent of the pilot plant under
this steady state were caused by the deterioration of the
nitrification process.
3.1.1. Anaerobic batch experiment for PAOs (experiment 5)
Figs. 2 and 3 show the results obtained in experiment 5 as well
as the model fittings used to obtain the parameters Y
PO4
, q
PHA
,
and K
A
for the two steady states. The parameter values
obtained in the two calibrations are presented in Table 4.
These parameter values allow good fitting of the experi-
mental values for S
PO4
and S
A
in both calibrations. There are
no significant differences between the values obtained for
both steady states, which indicates that SRT had little effect
on the parameter values. This small effect can be explained
because the biomass that was adapted to an SRT of 9.8 d
prevailed when SRT was changed to 5.7 d; and also because no
GAOs were present in the pilot plant inany of the steady states
(Rodrigo et al., 1999; Ferrer et al., 2004), as it was confirmed by
FISH technique.
3.1.2. Aerobic batch experiment for PAOs (experiment 6)
Fig. 4a and b showthe results obtained in experiment 6 as well
as the model fittings to obtain the parameter values for the
two steady states. The parameter values obtained in the two
calibrations are presented in Table 4.
In these calibrations, the parameter Y
PHA,O2
was fixed to
0.32 g CODg
À1
P, which is the value proposed by Smolders
et al. (1994) in their metabolic model. Lopez-Vazquez et al.
(2009) also proposed this value for PAOs when acetic acid is
the main substrate. From Y
PAO,O2
and Y
PHA,O2
values, and
taking into account the stoichiometric relationships from Eqs.
(9) and (10), the respective anoxic yields were obtained (see
Table 4). As Fig. 4 shows, the value of 0.32 for Y
PHA,O2
led to
a good concordance between experimental and simulated
values in both calibrations as well as in the pilot plant simu-
lations (see Tables 7 and 8).
As Table 4 shows, the value obtained for parameter Y
PO4
is
above 0.4 g P g
À1
COD, andthe value for K
MAX
is 0.34 g P g
À1
COD.
These values are expected when no GAO metabolism takes
place in the system (Rodrigo et al., 1999; Schuler and Jenkins,
2003; Ferrer et al., 2004). This underscores the utility of the
information obtained from experimental calibration method-
ologies for process performance evaluation, since parameter
values can indicate, for instance, the possible presence or
absence of GAOs in the system.
Table 5 e Values obtained for h
PAO
in both calibrations by
using Eq. (17) (i.e., from phosphorus uptake processes).
Parameter
(Units)
r
pa
(mg P l
À1
min
À1
)
r
po
(mg P l
À1
min
À1
)
h
PAO
Calibration 1 0.1798 0.7322 0.25
Calibration 2 0.3279 1.4051 0.23
Table 4 e Parameter values for PAOs obtained by experimental calibration in the two steady states.
Parameter Cal. 1 Cal. 2 Parameter Cal. 1 Cal. 2
Y
PO4
(g P g
À1
COD) 0.42 0.43 q
PP
(g P g
À1
CODd
À1
) 3.27 3.00
q
PHA
(g CODg
À1
CODd
À1
) 4.00 3.57 K
PHA
(g CODg
À1
COD) 0.07 0.06
K
A
(g CODm
À3
) 1.00 1.00 K
IPP
(g P g
À1
COD) 0.007 0.004
Y
PAO,O2
(g CODg
À1
COD) 0.65 0.74 K
MAX
(g P g
À1
COD) 0.34 0.34
Y
PHA,O2
(g CODg
À1
P) 0.32 0.32 Y
PAO,NO
(g CODg
À1
COD) 0.50 0.61
m
PAO
(d
À1
) 2.31 2.33 Y
PHA,NO
(g CODg
À1
P) 0.59 0.59
b
PAO
, b
PHA
, b
PP
(d
À1
) 0.20 0.20
wa t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 5 2 8 4 e5 2 9 7 5292
3.1.3. Determination of the anoxic reduction factor
(experiment 7)
As stated above, in order to obtain the h
PAO
parameter values
using bothEqs. (17) and(18), two batchtests were carriedout in
parallel (anoxic and aerobic). S
PO4
concentrations, as well as
electron acceptor concentrations (nitrate and oxygen), were
measured over time in order to obtain the maximal uptake
rates at the beginning of the experiment. Fig. 5a andb showthe
S
PO4
experimental data obtained under anoxic and aerobic
conditions for Calibration 1 and Calibration 2. The maximal
specificphosphorus uptakerates under anoxic(r
pa
) andaerobic
(r
po
) conditions wereobtainedfromtheseexperimental databy
linear regression of the first experimental points (where linear
behaviour canbe assumed). As observed in Sin et al. (2008) this
linear regression supposes the major source of uncertainty in
the estimation of maximal uptake rates. Hence, it is important
to consider only the time interval where experimental points
follow a linear profile. The values obtained for h
PAO
in both
calibrations using Eq. (17) are shown in Table 5.
In order to obtain the h
PAO
parameter values using Eq. (18),
electron acceptor concentrations (nitrate and oxygen) were
also needed. The S
NO3
experimental data obtained during the
anoxic batch tests for Calibration 1 and Calibration 2 are
shown in Fig. 6. The first two columns in Table 6 show the
maximal nitrate uptake rates (NUR
MAX
) obtained from these
data and the maximal oxygen uptake rates (OUR
MAX
) obtained
from the first respirometry in the aerobic batch test.
NUR
MAX,PeU
and NUR
MAX,GROWTH
as well as OUR
MAX,P-U
and
OUR
MAX,GROWTH
were obtained by using Eqs. (19) and (20),
respectively. The last column in Table 6 shows the h
PAO
parameter values obtained for Calibration 1 and Calibration 2
using Eq. (18).
As Tables 5 and 6 show, the same values were obtained for
h
PAO
using Eqs. (17) and (18). This confirms that the calibration
methodology is correct. On the one hand, these results vali-
date the assumption that h
PAO
exclusively represents the
fraction of denitrifying PAOs in the system. Hence, these
results validate the proposed modification in ASM2d, which
gives a physical meaning to this parameter. On the other
hand, the stoichiometric relationships proposed in Eqs. (9) and
(10) were also validated for the two steady states under
different sludge retention times. Therefore, the proposed
ASM2d modification does not mean extra calibration effort to
obtain the new parameters (Y
PAO,NO
and Y
PHA,NO
) since they
can be obtained as a function of their respective aerobic yields
and the proposed stoichiometric relationships. The main
advantage of this calibration approach is that general rela-
tionships can be used to simplify the number of parameters to
calibrate.
Since the nitrification process was totally deteriorated at
5.7 d, no nitrate was present in the pilot plant for this steady
state. However, the experimental calibrations show that the
fraction of denitrifying PAOs in the system did not change
significantly when SRT was decreased from 9.8 to 5.7 d.
Although the pilot plant was maintained for 42 d under SRT of
5.7 d, i.e. without nitrate, after biomass was exposed to nitrate
during the off-line calibration (experiment 7) the activated
sludge presented a value for h
PAO
similar to the first steady
state. Furthermore, no acclimatization period to nitrate was
observed during these experiments, contrarily to the
Table 6 e Maximum electron acceptor uptake rates and values obtained for h
PAO
in both calibrations by using Eq. (18) (i.e.,
from growth processes).
Process
Parameter Units
Total P-uptake Growth h
PAO
NUR
MAX
OUR
MAX
NUR
MAX,PeU
OUR
MAX,PeU
NUR
MAX,GROWTH
OUR
MAX,GROWTH
mg O
2
l
À1
min
À1
mg O
2
l
À1
min
À1
mg O
2
l
À1
min
À1
Calibration 1 0.2595 0.5840 0.1065 0.2343 0.1530 0.3497 0.25
Calibration 2 0.2656 0.6090 0.1943 0.4496 0.0713 0.1594 0.24
0
10
20
30
40
50
60
a
b
S
3
O
N
O
N
-
O
g
m
(
3
l
1
-
)
t (min)
SRT = 9.8 d
NUR
0
10
20
30
40
50
60
0 20 40 60 80 100 120 140 0 20 40 60 80 100 120 140
S
3
O
N
O
N
-
O
g
m
(
3
l
1
-
)
t (min)
SRT = 5.7 d
NUR
Fig. 6 e S
NO3
experimental data and maximal nitrate uptake rates (NUR
MAX
) obtained from batch experiments for (a)
Calibration 1 and (b) Calibration 2.
wa t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 5 2 8 4 e5 2 9 7 5293
observations of Zeng et al. (2003), which also indicates that the
same population remained in the system. This can be
explained by two possible hypothesis. (1) There existed two
different types of PAOs (DPAOs and non-DPAOs) and once
DPAOs were establishedinthe system(at previous steadystate
of 9.8 d) theycouldremainfor 42 dwithonlyoxygenas electron
acceptor, which indicates that they were able to use both
nitrate and oxygen. Similar results have been found by
Carvalho et al. (2007) and Flowers et al. (2009). And (2) the pilot
plant was maintained at 5.7 d for a too short time to observe
a change in population composition. Since it is usually
accepted that more than three times the SRT is enough to
assume steady state conditions for biomass the first hypoth-
esis seems more likely. However, it cannot be discarded
whether more time could lead to a more important change in
population composition or not. Whatever the reason, it is
important toemphasize that the same proportionof PAOs kept
the capability to use nitrate after 42 d without nitrate in the
pilot plant.
3.2. Pilot plant simulation
Finally, all the calibrated parameters were validated by
simulating the two steady states achieved in the pilot plant.
The simulations were carried out using the simulation soft-
ware DESASS (Ferrer et al., 2008), which incorporated the
proposed modified AMS2d model. DESASS can calculate
directly the steady state concentrations by converting the
differential equations of the model to algebraic ones (i.e.
setting the accumulation term to zero). The comparison
between experimental and simulated values for the two
steady states is shown in Tables 7 and 8. The experimental
value shown in these tables represents the average value
obtained from the different samples that were collected over
two weeks. Thus, the uncertainty associated to these experi-
mental values includes the standard deviation of the different
samples analyzed during this period as well as the coefficient
of variation associated to the analytical methods.
As Tables 7 and 8 show, the simulations carried out with
the calibrated parameters reproduce the pilot plant perfor-
mance under the two steady states with good concordance in
most of the measured parameters. The proposed ASM2d
modification allows the pilot plant to be simulated with the
same parameter values that can be obtained by off-line labo-
ratory experiments. The good concordance between experi-
mental and simulated values validates both the off-line
calibration methodology and the modified ASM2d model.
In summary, the use of different yields for aerobic and
anoxic conditions is based on the assumption that h
PAO
exclusively represents the proportion of X
PAO
that can follow
the DPAO metabolism. As these parameters are given a phys-
ical meaning, they can be used to simulate the plant in
a deterministic way, which allows WWTPs performance to be
diagnosed. The parameter values obtained by off-line cali-
bration provide valuable information about the activated
sludge process such as the amount of DPAOs in the system,
Table 8 ePilot plant experimental data obtained fromthe steady state of SRT[5.7 days, and simulated values obtained by
using parameter values from Calibration 2.
Parameter Unit Anaerobic reactor Anoxic reactor Aerobic reactor Effluent
Exp. Sim. Exp. Sim. Exp. Sim. Exp. Sim.
TCOD mgCODl
À1
1881 Æ150 1763 2347 Æ183 2344 2251 Æ316 2255 36 Æ3 41
SCOD mgCODl
À1
e e e e e e 31 Æ2 34
TSS mg l
À1
1431 Æ75 1353 1946 Æ140 1888 2047 Æ125 2003 7 Æ2 6
VSS mg l
À1
1129 Æ116 1194 1506 Æ178 1671 1489 Æ158 1787 6 Æ2 5
TP mgP l
À1
109 Æ9 103 149 Æ11 141 147 Æ9 141 1.2 Æ0.6 1.1
PO
4
-P mgP l
À1
34.0 Æ1.6 34.4 35.4 Æ1.1 35.3 0.2 Æ0.0 0.7 0.4 Æ0.1 0.7
NH
4
-N mgNl
À1
22.8 Æ1.5 23.2 22.5 Æ1.4 22.9 18.5 Æ0.9 18.4 17.7 Æ2.4 18.4
NO
3
-N mgNl
À1
<0.2 0.0 <0.2 0.0 0.3 Æ0.3 0.5 0.2 Æ0.2 0.5
TN mgNl
À1
e e e e e e 24.3 Æ6.8 23.7
TN
filtered
mgNl
À1
28.7 Æ6.0 27.7 27.7 Æ8.8 27.4 23.7 Æ7.4 23.2 23.7 Æ7.0 23.2
Table 7 ePilot plant experimental data obtained fromthe steady state of SRT[9.8 days, and simulated values obtained by
using parameter values from Calibration 1.
Parameter Unit Anaerobic reactor Anoxic reactor Aerobic reactor Effluent
Exp. Sim. Exp. Sim. Exp. Sim. Exp. Sim.
TCOD mg CODl
À1
2170 Æ166 2121 2978 Æ297 2946 2762 Æ295 2860 23 Æ2 24
SCOD mg CODl
À1
e e e e e e 22 Æ2 17
TSS mg l
À1
1701 Æ157 1779 2546 Æ218 2568 2704 Æ283 2658 6 Æ2 7
VSS mg l
À1
1263 Æ185 1474 1851 Æ285 2135 1882 Æ312 2225 5 Æ2 5
TP mg P l
À1
129 Æ12 147 172 Æ10 209 174 Æ15 209 2.1 Æ0.7 1.5
PO
4
-P mg P l
À1
28.6 Æ1.2 29.9 27.1 Æ1.6 28.8 0.6 Æ0.1 1.0 1.2 Æ0.3 0.7
NH
4
-N mg Nl
À1
18.9 Æ0.9 21.1 18.1 Æ0.6 18.7 0.6 Æ0.7 0.5 0.5 Æ0.2 0.5
NO
3
-N mg Nl
À1
0.7 Æ0.5 0.0 <0.2 0.2 16.9 Æ2.1 17.6 15.8 Æ2.1 15.5
TN mg Nl
À1
e e e e e e 18.3 Æ3.2 20.4
TN
filtered
mg Nl
À1
26.1 Æ5.3 24.2 24.3 Æ5.5 21.8 19.0 Æ3.9 20.2 17.9 Æ3.5 20.2
wa t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 5 2 8 4 e5 2 9 7 5294
the presence of GAOs, etc. In contrast, when h
NO3
is used in
ASM2d as a simple reduction in the kinetic expression for
anoxic growth and P-uptake, its value has to be adjusted by
simulating the plant performance and the values obtained
cannot give reliable information about the biological process.
Thus, decision-making about further process optimization
should be based on parameter values obtained by off-line
calibration methodologies, since these values can be inter-
preted by focusing on their physical meaning.
4. Conclusions
In this paper, the need for using different yields under aerobic
and anoxic conditions in nutrient removal process models has
been demonstrated experimentally. The ASM2d model has
been extended by adding anoxic yields as new parameters so
that parameter h
PAO
exclusively represents the fraction of
denitrifying PAOs (DPAOs). An off-line calibration method-
ology has been applied to validate the model, where general
relationships among stoichiometric parameters are proposed
to simplify the number of parameters to be calibrated.
The main conclusions of this work are as follows.
- Parameter h
PAO
in the modified ASM2d model can be
experimentally obtained from two different measures:
maximal electron acceptor uptake rates, and maximal
phosphorous uptake rates (obtained from anoxic and
aerobic batch tests). It has been shown both mathematically
and experimentally that the same value of h
PAO
will be
obtained from both of the measures if stoichiometric rela-
tionships between anoxic and aerobic yields are correct.
- The proposed modification in ASM2d does not mean an
extra calibration effort to obtain the new parameters
(Y
PAO,NO
and Y
PHA,NO
) since these parameters can be
obtained as a function of their respective aerobic yields and
the proposed stoichiometric relationships.
- Results from the off-line calibration methodology validate
the relationships between aerobic and anoxic yields as well
as the ASM2d modification. In this work, the stoichiometric
relationships have been validated for two steady states
under different sludge retention times.
- By simulating the pilot plant with the same parameter
values obtained by off-line laboratory experiments, we were
able to reproduce the pilot plant performance under both
steady states with good concordance between experimental
and simulated values.
- The parameter values obtained by the off-line calibration
methodology give valuable information about the activated
sludge process such as the amount of DPAOs in the system,
the presence of GAOs, etc. Hence, deterministic modelling
approaches together withoff-line calibration methodologies
can be used successfully in the study of biological phos-
phate removal to assist in decision-making about further
process optimization.
- After decreasing SRT from 9.8 to 5.7 d, the nitrification
process was completely deteriorated and nitrate was not
present in the pilot plant. However, the fraction of DPAOs
was similar for bothcalibrations. This indicates that the type
of DPAOs present in the pilot plant corresponds to PAOs that
are able to utilize either nitrate or oxygen as an electron
acceptor and they could retain the capability to use nitrate
after more thanone monthwithout nitrate inthe pilot plant.
Acknowledgements
The authors would like to acknowledge Entitat Pu´ blica de
Sanejament d’Aigu¨ es Residuals (EPSAR) and the SEARSA and
AQUAGEST companies for their financial support.
r e f e r e n c e s
Amann, R., 1995. Fluorescently labelled, rRNA-targeted
oligonucleotide probes in the study of microbial ecology.
Molecular Ecology 4 (5), 543e554.
APHA, 2005. Standard Methods for the Examination of Water and
Wastewater, 21st ed. American Public Health Association,
American Water Works Association and Water Environment
Federation, Washington DC, USA.
Barker, P.S., Dold, P.L., 1996. Denitrification behaviour in
biological phosphorus removal activated sludge systems.
Water Research 30 (4), 769e780.
Carvalho, G., Lemos, P.C., Oehmen, A., Reis, M.A.M., 2007.
Denitrifying phosphorus removal: linking the process
performance with the microbial community structure. Water
Research 41 (19), 4383e4396.
Cech, J.S., Chudoba, J., Grau, P., 1984. Determination of kinetics
constants of activated sludge microorganisms. Water Science
and Technology 17 (2/3), 259e272.
Comeau, Y., Hall, K.J., Handcock, R.E.W., Oldham, W.K., 1986.
Biochemical model for enhanced biological phosphorus
removal. Water Research 20 (12), 1511e1521.
Comeau, Y., Oldham, W.K., Hall, K.J., 1987. Dynamics of Carbon
Reserves in Biological Dephosphatation of Wastewater.
Biological Phosphate Removal from Wastewaters, Roma.
Pergamon Press, Oxford.
Crocetti, G.R., Banfield, J.F., Keller, J., Bond, P.J., Blackall, L.L., 2002.
Glycogen accumulating organisms in laboratory-scale and
full-scale activated sludge processes. Microbiology 148,
3353e3364.
Ferrer, J., Morenilla, J.J., Bouzas, A., Garcia-Usach, F., 2004.
Calibration and simulation of two large wastewater treatment
plants operated for nutrient removal. Water Science and
Technology 50 (6), 87e94.
Ferrer, J., Seco, A., Garcı´a-Usach, F., Bouzas, A., Barat, R., 2005.
Simulation of full scale plants: benefits and drawbacks. In:
Van Loosdrecht, Mark, Clement, Jonathan (Eds.), Water and
Environmental Management Series: 2nd IWA Leading-Edge on
Water and Wastewater Treatment Technologies. IWA
Publishing, London, pp. 155e163.
Ferrer, J., Seco, A., Serralta, J., Ribes, J., Manga, J., Asensi, E.,
Morenilla, J.J., Llavador, F., 2008. DESASS: a software tool for
designing, simulating and optimising WWTPs. Environmental
Modelling & Software 23 (1), 19e26.
Flowers, J.J., He, S., Yilmaz, S., Noguera, D.R., McMahon, K.D.,
2009. Denitrification capabilities of two biological phosphorus
removal sludges dominated by different ‘Candidatus
Accumulibacter’ clades. Environmental Microbiology Reports
1 (6), 583e588.
Freitas, F., Temudo, M., Reis, M.A.M., 2005. Microbial
population response to changes of the operating
conditions in a dynamic nutrient-removal sequencing
batch reactor. Bioprocess and Biosystems Engineering 28
(3), 199e209.
wa t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 5 2 8 4 e5 2 9 7 5295
Griffiths, P.C., Stratton, H.M., Seviour, R.J., 2002. Environmental
factors contributing to the “G bacteria” population in full-
scale EBPR plants. Water Science and Technology 46 (4e5),
185e192.
Henze, M., Gujer, W., Mino, T., Matsuo, T., Wentzel, M.C.,
Marais, G.V.R., 1995. Activated Sludge Model No. 2. IAWQ
Scientific and Technical Report No.2. IAWQ, London.
Henze, M., Gujer, W., Mino, T., Matsuo, T., Wentzel, M.C.,
Marais, G.v.R., Van Loosdrecht, M.C.M., 1999. Activated sludge
model No. 2d. Water Science and Technology 39 (1), 165e182.
Hu, Z., Wentzel, M.C., Ekama, G.A., 2003. Modelling biological
nutrient removal activated sludge systems e a review. Water
Research 37 (14), 3430e3444.
Hulsbeek, J.J.W., Kruit, J., Roeleveld, P.J., van Loosdrecht, M.C.M.,
2002. A practical protocol for dynamic modelling of activated
sludge systems. Water Science and Technology 45 (6),
127e136.
Insel, G., Sin, G., Lee, D.S., Nopens, I., Vanrolleghem, P.A., 2006. A
calibration methodology and model-based systems analysis
for SBRs removing nutrients under limited aeration
conditions. Journal of Chemical Technology and
Biotechnology 81 (4), 679e687.
Kerrn-Jespersen, J.P., Henze, M., 1993. Biological phosphorus
uptake under anoxic and aerobic conditions. Water Research
27 (4), 617e624.
Kuba, T., Smolders, G., van Loosdrecht, M.C.M., Heijnen, J.J., 1993.
Biological phosphorus removal from wastewater by
anaerobic-anoxic sequencing batch reactor. Water Science
and Technology 27 (5), 241e252.
Kuba, T., Murnleitner, E., van Loosdrecht, M.C.M., Heijnen, J.J.,
1996. A metabolic model for the biological phosphorus
removal by denitrifying organisms. Biotechnology and
Bioengineering 52 (6), 685e695.
Langergraber, G., Rieger, L., Winkler, S., Alex, J., Wiese, J.,
Owerdieck, C., Ahnert, M., Simon, J., Maurer, M., 2004. A
guideline for simulation studies of wastewater treatment
plants. Water Science and Technology 50 (7), 131e138.
Liu, W.T., Nakamura, K., Matsuo, T., Mino, T., 1997. Internal
energy-based competition between polyphosphate- and
glycogen-accumulating bacteria in biological phosphorus
removal reactors e effect of P/C feeding ratio. Water Research
31 (6), 1430e1438.
Lopez-Vazquez, C.M., Oehmen, A., Hooijmans, C.M.,
Brdjanovic, D., Gijzen, H.J., Yuan, Z., van Loosdrecht, M.C.M.,
2009. Modeling the PAOeGAO competition: effects of carbon
source, pH and temperature. Water Research 43 (2), 450e462.
Makinia, J., Rosenwinkel, K.H., Spering, V., 2006a. Comparison of
two model concepts for simulation of nitrogen removal at
a full-scale biological nutrient removal pilot plant. Journal of
Environmental Engineering 132 (4), 476e487.
Makinia, J., Rosenwinkel, K.H., Swinarski, M., Dobiegala, E., 2006b.
Experimental and model-based evaluation of the role of
denitrifying polyphosphate accumulating organisms at two
large scale WWTPs in northern Poland. Water Science and
Technology 54 (8), 73e81.
Maurer, M., Gujer, W., 1998. Dynamic modelling of enhanced
biological phosphorus and nitrogen removal in activated
sludge systems. Water Science and Technology 38 (1),
203e210.
Meijer, S.C.F., van Loosdrecht, M.C.M., Heijnen, J.J., 2001.
Metabolic modelling of full-scale biological nitrogen and
phosphorus removing WWTP’s. Water Research 35 (11),
2711e2723.
Meinhold, J., Filipe, C.D.M., Daigger, G.T., Isaacs, S., 1999.
Characterization of the denitrifying fraction of phosphate
accumulating organisms in biological phosphate removal.
Water Science and Technology 39 (1), 31e42.
Mino, T., van Loosdrecht, M.C.M., Heijnen, J.J., 1998. Microbiology
and biochemistry of the enhanced biological phosphate
removal process. Water Research 32 (11), 3193e3207.
Muller, A., Wentzel, M.C., Loewenthal, R.E., Ekama, G.A., 2003.
Heterotroph anoxic yield in anoxic aerobic activated sludge
systems treating municipal wastewater. Water Research 37
(10), 2435e2441.
Murnleitner, E., Kuba, T., van Loosdrecht, M.C.M., Heijnen, J.J.,
1997. An integrated model for the aerobic and denitrifying
biological phosphorus process. Biotechnology and
Bioengineering 54 (5), 434e450.
Oehmen, A., Lemos, P.C., Carvalho, G., Yuan, Z., Keller, J.,
Blackall, L.L., Reis, M.A.M., 2007. Advances in enhanced
biological phosphorus removal: from micro to macro scale.
Water Research 41 (11), 2271e2300.
Penya-Roja, J.M., Seco, A., Ferrer, J., Serralta, J., 2002. Calibration
and validation of Activated Sludge Model No. 2d for Spanish
municipal wastewater. Environmental Technology 23 (8),
849e862.
Ribes, J., Keesman, K., Spanjers, H., 2004. Modelling
anaerobic biomass growth kinetics with a substrate
threshold concentration. Water Research 38 (20),
4502e4510.
Rieger, L., Koch, G., Ku¨ hni, M., Gujer, W., Siegrist, H., 2001. The
EAWAG Bio-P module for Activated Sludge Model no. 3. Water
Research 35 (16), 3887e3903.
Rodrigo, M.A., Seco, A., Ferrer, J., Penya-Roja, J.M., 1999. The effect
of sludge age on the deterioration of the enhanced biological
phosphorus removal process. Environmental Technology 20
(10), 1055e1063.
Ruano, M.V., Ribes, J., De Pauw, D.J.W., Sin, G., 2007. Parameter
subset selection for the dynamic calibration of activated
sludge models (ASMs): experience versus systems analysis.
Water Science and Technology 56 (8), 107e115.
Schuler, A.J., Jenkins, D., 2003. Enhanced biological phosphorus
removal from wastewater by biomass with different
phosphorus contents, part I: experimental results and
comparison with metabolic models. Water Environment
Research 75 (6), 485e498.
Sin, G., van Hulle, S.W.H., de Pauw, D.J.W., van Griensven, A.,
Vanrolleghem, P.A., 2005. A critical comparison of systematic
calibration protocols for activated sludge models: A SWOT
analysis. Water Research 39 (12), 2459e2474.
Sin, G., Niville, K., Bachis, G., Jiang, T., Nopens, I., van Hulle, S.,
Vanrolleghem, P.A., 2008. Nitrite effect on the phosphorus
uptake activity of phosphate accumulating organisms (PAOs)
in pilot-scale SBR and MBR reactors. Water SA 34 (2),
249e260.
Smolders, G.J.F., van der Meij, J., van Loosdercht, M.C.M.,
Heijnen, J.J., 1994. Stoichiometric model of the aerobic
metabolism of the biological phosphorus removal
process. Biotechnology and Bioengineering 44 (7),
837e848.
Vanrolleghem, P.A., Insel, G., Petersen, B., Sin, G., De Pauw, D.,
Nopens, I., Weijers, S., Gernaey, K., 2003. A comprehensive
model calibration procedure for activated sludge models. In:
Proceedings: WEFTEC 2003, 76th Annual Technical Exhibition
and Conference, October 11e15, 2003, Los Angeles, CA, USA
(on CD-ROM).
Vanrolleghem, P.A., Spanjers, H., Petersen, B., Ginestet, P.,
Takacs, I., 1999. Estimating (combinations of) Activated Sludge
Model No.1 parameters and components by respirometry.
Water Science and Technology 39 (1), 195e214.
Vlekke, G.J.F.M., Comeau, Y., Oldham, W.K., 1988. Biological
phosphate removal from wastewater with oxygen or nitrate in
sequencing batch reactors. Environmental Technology Letters
9 (8), 791e796.
wa t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 5 2 8 4 e5 2 9 7 5296
Wachtmeister, A., Kuba, T., van Loosdrecht, M.C.M., Heijnen, J.J.,
1997. A sludge characterization assay for aerobic and
denitrifying phosphorus removing sludge. Water Research 31
(3), 471e478.
Whang, L.M., Park, J.K., 2006. Competition between
polyphosphate and glycogen-accumulating organisms in
enhanced biological-phosphorus-removal systems: effect of
temperature and sludge age. Water Environment Research 78
(1), 4e11.
WRC, 1992. Simple Titration Procedures to Determine H
2
CO
3
*
Alkalinity and Short-chain Fatty Acids in Aqueous Solutions
Containing Known Concentrations of Ammonium, Phosphate
and Sulphide Weak Acid/bases. Report No. TT 57/92. Water
Research Commission, University of Cape Town, Pretoria,
Republic of South Africa.
Yagci, N., Insel, G., Tasli, R., Artan, N., Randall, C.W., Orhon, D.,
2006. A new interpretation of ASM2d for modelling of SBR
performance for enhanced biological phosphorus removal
under different P/HAc ratios. Biotechnology and
Bioengineering 93 (2), 258e270.
Zeng, R.J., Saunders, A.M., Yuan, Z., Blackall, L.L., Keller, J., 2003.
Identification and comparison of aerobic and denitrifying
polyphosphate-accumulating organisms. Biotechnology and
Bioengineering 83 (2), 140e148.
wa t e r r e s e a r c h 4 4 ( 2 0 1 0 ) 5 2 8 4 e5 2 9 7 5297

Sponsor Documents

Or use your account on DocShare.tips

Hide

Forgot your password?

Or register your new account on DocShare.tips

Hide

Lost your password? Please enter your email address. You will receive a link to create a new password.

Back to log-in

Close