Separation and Purification Technology 126 (2014) 30–38

Contents lists available at ScienceDirect

Separation and Purification Technology

journal homepage: www.elsevier.com/locate/seppur

The operating cost of an anaerobic membrane bioreactor (AnMBR)

treating sulphate-rich urban wastewater
R. Pretel a,⇑, A. Robles b, M.V. Ruano c, A. Seco b, J. Ferrer a
a
Institut Universitari d’Investigació d’Enginyeria de l’Aigua i Medi Ambient, IIAMA, Universitat Politècnica de València, Camí de Vera s/n, 46022 Valencia, Spain
b
Departament d’Enginyeria Química, Escola Tècnica Superior d’Enginyeria, Universitat de València, Avinguda de la Universitat s/n, 46100, Burjassot, Valencia, Spain
c
Aqualia, Gestión Integral del Agua, S.A., Avenida del Camino de Santiago, 40 28050 Madrid, Spain

a r t i c l e i n f o a b s t r a c t

Article history: The objective of this study was to evaluate the operating cost of an anaerobic membrane bioreactor
Received 20 November 2013 (AnMBR) treating sulphate-rich urban wastewater (UWW) at ambient temperature (ranging from 17
Received in revised form 14 February 2014 to 33 °C). To this aim, energy consumption, methane production, and sludge handling and recycling to
Accepted 18 February 2014
land were evaluated. The results revealed that optimising specific gas demand with respect to permeate
Available online 22 February 2014
volume (SGDP) and sludge retention time (for given ambient temperature conditions) is essential to max-
imise energy savings (minimum energy demand: 0.07 kW h m3). Moreover, low/moderate sludge pro-
Keywords:
ductions were obtained (minimum value: 0.16 kg TSS kg1 CODRemoved), which further enhanced the
Energy consumption
Industrial-scale hollow-fibre membranes
overall operating cost of the plant (minimum value: €0.011 per m3 of treated water). The sulphate con-
Operating cost tent in the influent UWW significantly affected the final production of methane and thereby the overall
Anaerobic membrane bioreactor (AnMBR) operating cost. Indeed, the evaluated AnMBR system presented energy surplus potential when treating
Sulphate-rich urban wastewater low-sulphate UWW.
Ó 2014 Elsevier B.V. All rights reserved.

1. Introduction competition between Methanogenic Archaea (MA) and Sulphate

Reducing Bacteria (SRB) for the available substrate [4] when there
Nowadays, a key issue in global sustainable development is the is significant sulphate content in the influent, reducing therefore
dependency on fossil fuels for electricity production, which repre- the available COD for methanisation [5]. For urban wastewater,
sents up to the 80% of the global energy consumption [1]. In this which can easily present low COD/SO4–S ratio, this competition
respect, electricity consumption is a key element in the overall can critically affect the amount and quality of the biogas produced.
environmental performance of a wastewater treatment plant Specifically, 2 kg of COD are consumed by SRB in order to reduce
(WWTP) [2]. Hence, it is particularly important to implement 1 kg of influent SO4–S (see, for instance, [5]). According to the the-
new energy-saving technologies that reduce the overall energy oretical methane yield under standard temperature and pressure
balance of the WWTP, such as anaerobic membrane bioreactors conditions (350 LCH4 kg1COD), SRB reduces the production of ap-
(AnMBRs). This technology focuses on the sustainability benefits prox. 700 L of methane per kg of influent SO4–S (considering reduc-
of anaerobic processes compared to aerobic processes, such as: tion of all sulphate to sulphide). Therefore, higher biogas
minimum sludge production due to low biomass yield of anaerobic productions would be achieved when there is little sulphate con-
organisms; low energy demand since no aeration is required; and tent in the influent (typical sulphate concentration in UWW fluctu-
methane production that can be used to fulfil process energy ates around 7–17 mg SO4–S L1 [6]). On the other hand, due to the
requirements [3]. low-growth rate of anaerobic microorganism, high sludge reten-
Several issues have been recognised elsewhere as potential tion times (SRTs) are required when operating at low temperatures
drawbacks which may affect the sustainability of AnMBR technol- in order to achieve suitable organic matter removal rates, espe-
ogy treating urban wastewater (UWW). One key issue is the cially for low-strength wastewaters like urban ones (typical COD
levels below 1 g L1 [6]). However, as regards filtration process,
operating AnMBRs at high SRT may imply operating at high mixed
liquor total solid (MLTS) levels. This is considered to be one of the
⇑ Corresponding author. Tel.: +34 96 387 61 76; fax: +34 96 387 90 09.
main constraints on membrane operating because it can result in a
E-mail addresses: rutprejo@upv.es (R. Pretel), angel.robles@uv.es (A. Robles),
mavictoria.ruano.garcia@fcc.es (M.V. Ruano), aurora.seco@uv.es (A. Seco), jferrer@ high membrane fouling propensity and therefore high energy
hma.upv.es (J. Ferrer). demand for membrane scouring by gas sparging [7].

http://dx.doi.org/10.1016/j.seppur.2014.02.013
1383-5866/Ó 2014 Elsevier B.V. All rights reserved.
 R. Pretel et al. / Separation and Purification Technology 126 (2014) 30–38 31

The objective of this study was to evaluate the operating cost of (SRFMT and SRFAnR respectively): 2.7 and 1 m3 h1 respectively;
an AnMBR system treating sulphate-rich urban wastewater specific gas demand per square metre of membrane area (SGDm):
(UWW) at ambient temperature (ranging from 17 to 33 °C). To this controlled at 0.17 and 0.23 m3 h1 m2; and biogas recycling flow
aim, power requirements, energy recovery from methane (biogas to the anaerobic reactor (BRFAnR): 1.5 m3 h1.
methane and/or methane dissolved in the effluent), and sludge
handling and recycling to land were evaluated at different operat- 2.2.2. Biological process
ing conditions. In order to obtain reliable results that can be Variations in SRT and seasonal temperature were studied to ac-
extrapolated to full-scale plants, this study was carried out in an count for the dynamics in methane and sludge productions over
AnMBR using industrial-scale hollow-fibre membrane units. This time. During the 920-day experimental period the plant was oper-
system was operated using effluent from the pre-treatment of ated at ambient temperature ranging from 17 to 33 °C and SRT var-
the Carraixet WWTP (Valencia, Spain). ied from 30 to 70 days. Three different experimental scenarios
related to biological process (BP33 °C, SRT 70 days, BP22 °C, SRT 38 days
2. Materials and methods and BP17 °C, SRT 30 days) were considered to evaluate the energy con-
sumption of the AnMBR plant (see Table 3): (1) a summer period of
2.1. AnMBR plant description 2 months of operation resulting in high methane and low sludge
productions (BP33 °C, SRT 70 days) due to operating at high tempera-
A semi-industrial AnMBR plant was operated using the effluent ture (33 °C in average) and high SRT (70 days); (2) one year of
of a full-scale WWTP pre-treatment. The average AnMBR influent operation resulting in moderate methane and sludge productions
characteristics are shown in Table 1. This influent UWW was char- (BP22° C, SRT 38 days) due to operating at variable temperature
acterised by a low COD (around 650 mg L1) and high sulphate (22 °C in average) and moderate SRT (38 days); and (3) a winter
concentration (around 105 mg SO4–S L1). period of 2 months of operation resulting in low methane and
The AnMBR plant consists of an anaerobic reactor with a total moderate sludge productions (BP17 °C, SRT 30 days) due to operating
volume of 1.3 m3 connected to two membrane tanks (MT1 and at relatively low temperature (17.1 °C in average) and moderate
MT2) each one with a total volume of 0.8 m3. Each membrane tank SRT (30 days). These three scenarios represent boundary (BP33 °C,
includes one ultrafiltration hollow-fibre membrane commercial SRT 70 days: best conditions; and BP17 °C, SRT 30 days: worst conditions)

system (PURONÒ, Koch Membrane Systems, 0.05 lm pore size, and average (BP22 °C, SRT 38 days) of the operating conditions evalu-
30 m2 total filtering area). The filtration process was studied from ated in the plant.
experimental data obtained from MT1 (operated recycling contin- In addition, several simulation scenarios were calculated in or-
uously the obtained permeate to the system), whilst the biological der to assess the AnMBR performance within the whole range of
process was studied using experimental data obtained from MT2 temperature (17–33 °C) and SRT (30–70 days) evaluated in this
(operated for the biological process without recycling the obtained study. Simulation results were obtained using the WWTP simulat-
permeate). Hence, different 20 °C-standardised transmembrane ing software DESASS [9]. This simulation software features the
fluxes (J20) were tested in MT1, without affecting the hydraulic mathematical model BNRM2 [10], which was previously validated
retention time (HRT) of the plant. using experimental data obtained in the AnMBR plant. Fig. 1 shows
In addition to conventional membrane operating stages the resulting effluent COD without including dissolved methane
(filtration, relaxation and back-flushing), two additional stages concentration (see Fig. 1a); total methane production (see
were considered in the membrane operating mode: degasification Fig. 1b); and sludge production (Fig. 1c) for the different tempera-
and ventilation. Further details on this AnMBR can be found in ture and SRT conditions simulated.
Giménez et al. [5] and Robles et al. [8].
2.2.2.1. Influent sulphate concentration. The effect of the influent sul-
2.2. AnMBR operating conditions phate on the AnMBR operating cost was also evaluated. As men-
tioned before, the UWW fed to the AnMBR plant was characterised
The AnMBR plant was operated for around 920 days within a by relatively low COD and high sulphate concentrations
wide range of operating conditions for both filtration and biologi- (see Table 1). Therefore, an important fraction of the influent COD
cal process. was consumed by SRB. To be precise, the sulphate content in the
influent was approx. 105 mg S–SO4 L1, from which approx. 98%
2.2.1. Filtration process was reduced to hydrogen sulphide (around 103 mg S–SO4 L1).
Five operating scenarios related to filtration process (FP1-FP5) Therefore, about 206 mg L1 of influent COD were consumed by SRB.
were considered to evaluate the energy consumption of the AnMBR The results obtained in this study were compared to the theoret-
plant (see Table 2). As Table 2 shows, the main operating conditions ical results obtained in an AnMBR system treating low-sulphate
in these five scenarios were as follows: transmembrane pressure UWW (10 mg S–SO4 L1). To this aim, the methane production
(TMP) during filtration: from 0.09 to 0.35 bar; J20 from 9 to 20 when treating low-sulphate UWW was calculated on the basis of
LMH; MLTS entering the membrane tank: from 12.5 to 32.5 g L1; the theoretical methane yield under standard temperature and
sludge recycling flow in anaerobic reactor and membrane tank pressure conditions: 350 LCH4 kg1 COD. Table 4 shows the theoret-
ical methane production (including both biogas methane and meth-
ane dissolved in the effluent) obtained for cases BP33 °C, SRT 70 days,
Table 1
Average characteristics of AnMBR influent. BP22 °C, SRT 38 days and BP17 °C, SRT 30 days when treating low-sulphate
UWW (10 mg S–SO4 L1). The distribution between gas and liquid
Parameter Mean ± SD
phase of the produced methane was established on the basis of
Treatment flow rate (m3 day1) 3.2 ± 0.7 the experimental distribution obtained in the AnMBR plant.
TSS (mg L1) 313 ± 45
VSS (mg L1) 257 ± 46
COD (mg L1) 650 ± 147 2.3. Analytical monitoring
SO4–S (mg L1) 105 ± 13
NH4–N (mg L1) 35 ± 3 The following parameters were analysed in mixed liquor and
PO4–P (mg L1) 4±1 influent stream according to Standard Methods [11]: total solids
(TS); total suspended solids (TSS); volatile suspended solids
32 R. Pretel et al. / Separation and Purification Technology 126 (2014) 30–38

Table 2
Main operating conditions in scenarios FP1–FP5.

Scenario Period TMP J20 MLTS SRFMT SRFAnR SGDm BRFAnR

(days) (bar) (LMH) (g L1) (m3 h1) (m3 h1) (m3 h1 m2) (m3 h1)
FP1 137–170 0.35 10.0 32.5 2.7 1 0.23 1.5
FP2 361–404 0.13 13.3 12.5 2.7 1 0.23 1.5
FP3 556–600 0.26 9.0 22.5 2.7 1 0.23 1.5
FP4 807–850 0.09 15.0 14 2.7 1 0.17 1.5
FP5 853–896 0.20 20.0 13 2.7 1 0.23 1.5

TMP: transmembrane pressure; J20: 20 °C-standardised transmembrane flux; MLTS: mixed liquor total solids; SRFMT and SRFMT: sludge recycling flow to membrane tank and
anaerobic reactor, respectively; SGDm: specific gas demand per square metre of membrane area; and BRFAnR: biogas recycling flow to anaerobic reactor.

Table 3
Operating temperature (T) and sludge retention time (SRT), total methane production (VCH4), biogas methane (VCH4,biogas), and methane dissolved in the effluent (VCH4,effluent) per
m3 of treated water, and sludge production, for cases BP33 °C, SRT 70 days, BP22 °C, SRT 38 days and BP17 °C, SRT 30 days.

T (°C) SRT (days) VCH4 (biogas + effluent) (L m3) VCH4,biogas (L m3) VCH4,effluent (L m3) Sludge production (kg TSSkg1 COD removed)
BP33 °C, SRT 70 days 33 70 41.1 26.5 14.6 0.16
BP22 °C, SRT 38 days 22 38 16.8 8.4 8.4 0.43
BP17 °C, SRT 30 days 17 30 8.5 1.4 7.1 0.55

(VSS); sulphate (SO4–S); nutrients (ammonium (NH4–N) and calculated by applying the corresponding theoretical equations
orthophosphate (PO4–P)); and chemical oxygen demand (COD). (Eqs. (1)–(3), respectively).
The methane fraction of the biogas was measured using a gas chro-   " a1 #
matograph equipped with a Flame Ionization Detector (GC-FID, J ðM  R  T gas Þ p2 a
PB ¼ 1 ð1Þ
Thermo Scientific) in accordance with Giménez et al. [5]. The S ða  1Þ  gblower p1
dissolved methane fraction of the effluent was determined in
accordance with Giménez et al. [12]. AMPTSÒ (Automatic Methane where PB is the blower power requirement (adiabatic compression),
Potential Test System, Bioprocess Control) was employed for eval- M (mol s1) is the molar flow rate of biogas, R (J mol1 K1) is the
uating the biochemical methane potential (BMP) of the wasted gas constant for biogas, P1 (atm) is the absolute inlet pressure, P2
sludge. Due to the low microbial activity of this sludge, BMP tests (atm) is the absolute outlet pressure, Tgas (K) is the biogas temper-
were inoculated using biomass coming from the anaerobic digester ature, a is the adiabatic index and gblower is the blower efficiency.
of the Carraixet WWTP. VSS and TSS levels in the wasted sludge P1 and M were taken from the data obtained in the AnMBR plant;
were measured at the beginning and at the end of the BMP test, P2 and Tgas were calculated by the simulation software; and a value of
allowing the percentage of biodegradable volatile suspended solids 0.8 was considered for gblower as a theoretical typical value.
 
(%BVSS) to be calculated. In this study, the sludge stabilisation cri- J
terion was set to 35% of BVSS. Pg ¼ qimp  qliquor  g
S
nh    i o
ðLþLeq Þf V 2 ðLþLeq Þf V 2
D2g
asp: þ D2g
imp: þ ½Z 1  Z 2 
2.4. Energy balance description  ð2Þ
gpump
The energy balance of the AnMBR system consisted of: power where Pg is the power requirement by the general pump, consider-
requirements (W), and energy recovery from both biogas methane ing both pump aspiration and pump impulsion section, calculated
(Ebiogas) and methane dissolved in the effluent (Edissolved methane). The from the impulsion volumetric flow rate (qimp. in m3 s1), liquor
heat energy term (Q) was assumed negligible since the process was density (qliquor in kg m3), acceleration of gravity (g in m s1), pipe
evaluated at ambient temperature conditions. length (L in m), pipe equivalent length of the punctual pressure
Therefore, the AnMBR energy consumption was evaluated in drops (Leq in m), liquor velocity (V in m s1), friction factor (f,
this study assuming the following terms: (1) energy consumption dimensionless), diameter (d in m), difference in height (Z1–Z2, in
when non-capture of methane is considered; (2) net energy con- m) and pump efficiency (gpump).
sumption including energy recovery from biogas methane; and qimp and qliquor were taken from the data obtained in the AnMBR
(3) net energy consumption including energy recovery from both plant; L, Leq, D and Z1–Z2 were taken from the dimensions of the
biogas methane and methane dissolved in the effluent. AnMBR plant; V and f were calculated by the modelling software;
The equipment considered in the W term consisted of the fol- and a value of 0.8 was considered for gpump as a theoretical typical
lowing: one anaerobic reactor feeding pump; one membrane tank value.
sludge feeding pump; one anaerobic reactor sludge mixing pump;  
one permeate pump; one anaerobic reactor biogas recycling J qstage  TMP stage
Pstageðfiltration; degasification or backflushingÞ ¼ ð3Þ
blower; one membrane tank biogas recycling blower; one rotofil- S gpump
ter; and one dewatering system.
where Pstage is the permeate pump power requirement during filtra-
The energy requirements for each of the scenarios evaluated in
tion, degasification or back-flushing calculated from transmem-
this study were calculated using the simulation software DESASS,
brane pressure (TMPstage in Pa), pump volumetric flow rate
which includes a general tool that enables calculating the energy
(qstage in m3 s1) and pump efficiency (gpump).
consumption of the different units comprising a WWTP.
TMPstage and qstage were taken from the data obtained in the
AnMBR plant.
2.4.1. Power requirements (W) To calculate the net power required by the permeate pump
As proposed by Judd and Judd [13], the energy consumption (Ppermeate), the sum of the power consumed in the following four
related to pumps and blowers (adiabatic compression), was membrane operating stages was considered: filtration (Pfiltration),
 R. Pretel et al. / Separation and Purification Technology 126 (2014) 30–38 33

back-flushing (Pback-flushing), degasification (Pdegasification) and ventila-

tion (Pventilation). Eq. (4) was used to calculate the power in filtration,
60 back-flushing and degasification. Eq. (3) was used to calculate the
power in ventilation since the fluid does not pass through the
g· L )

50
-1

membrane.
40 The energy consumption related to the rotofilter was obtained
(m

from a catalogue for full-scale implementation [14].

COD effluent

30
Concerning sludge handling, centrifuges with an average power
20
32 consumption of 45 kW h t1 TSS [15] were selected in our study as
10 29 sludge dewatering system.
26

)
ºC
0 23

T(
65 60
55 50 20 2.4.2. Energy recovery from methane
45 40
35 30 17 Since microturbines can run on biogas, they were selected as
SRT (
days) combined heat and power (CHP) technology [16]. Microturbine-
(a) based CHP technology has an overall efficiency of around 65.5%,
assuming power energy efficiency of about 27% (see Eq. (4)).

V biogas  ð%CH4  CV CH4 Þ  %power effciency CHP

W biogas ðkWÞ ¼ ð4Þ
1000  24  3600
50 where Wbiogas is the power generated by the Microturbine-based
CHP system using biogas, Vbiogas (L d1) is the biogas volume, %CH4
40 is the methane percentage and CVCH4 (kJ m3) is the methane calo-
VCH4 ( L· m )
-3

30
rific power.
It must be said that methane dissolved in the effluent was con-
20 sidered to be captured for obtaining power energy by using the
32 Microturbine-based CHP system. Theoretical capture efficiency
10 29 for the dissolved methane of 100% was considered in order to as-
26
)

sess the maximum energy potential.

ºC

0 23
T(

65 60
55 50 20
45 40
35 30 17 2.5. Operating cost assessment
SRT (
days)
The operating cost analysis was limited in this study to net en-
(b) ergy demand, and sludge handling and recycling to land.
The net energy demand in scenarios FP1-FP5 was evaluated for
cases BP33 °C, SRT 70 days, BP22 °C, SRT 38 days and BP17 °C, SRT 30 days
assuming, as previously mentioned, the following terms: (1) non-
capture of methane; (2) energy recovery from biogas methane;
oved

1.0
and (3) energy recovery from both biogas methane and methane
kg TSS·kg COD rem

dissolved in the effluent. The energy term considered in this study

was €0.138 per kW h (according to the current electricity rates and
prices in Spain [17]).
0.5
Concerning sludge handling and recycling to land, centrifuges
-1

require the use of polyelectrolyte for proper sludge conditioning.

32
29 The dose of polyelectrolyte considered in our study was 6 kg t1
26 TSS [18], and the assumed polyelectrolyte cost was €2.52 per kg
)
(ºC

0.0 23
65 60 Polyelectrolyte [19]. The produced sludge was considered to be
T

55 50 20
45 40 used as a fertiliser in agricultural land. The assumed cost for sludge
35 30 17
SRT ( recycling to land was €4.81 per t TSS [19].
Days)

(c) 3. Results and discussion

Fig. 1. AnMBR performance at different temperature and SRT conditions: (a) 3.1. Overall process performance
effluent COD (without including dissolved methane concentration); (b) total
methane production (VCH4) (biogas methane and methane dissolved in the
effluent); and (c) sludge production measured in kg TSSkg1 COD removed.
Fig. 2 shows the 20 °C-standardised membrane permeability
(K20) and the MLTS level in the anaerobic sludge fed to the mem-
brane tanks during 920 days of operation. Both K20 and MLTS are

Table 4
Theoretical methane production (VCH4), biogas methane (VCH4,biogas), and methane dissolved in the effluent (VCH4,effluent) per m3 of treated water for cases BP33 °C, SRT 70 days, BP22 °C,
SRT 38 and BP17 °C, SRT 30 days when treating low-sulphate UWW.

VCH4, (biogas + effluent) (L m3) VCH4,biogas (L m3) VCH4,effluent (L m3)

BP33 °C, SRT 70 days 105.8 68.1 37.7
BP22 °C, SRT 38 days 81.5 40.8 40.7
BP17 °C, SRT 30 days 73.2 11.7 61.5
34 R. Pretel et al. / Separation and Purification Technology 126 (2014) 30–38

the most unfavourable conditions evaluated in this study (i.e.

BP17 °C, SRT 30 days). It is important to highlight that one key sustain-
able benefit of AnMBR technology is that the produced sludge is
stabilised and no further digestion is required for its disposal on
farmland. In addition, sludge production in anaerobic processes is
expected to be lower than in aerobic processes.

3.2. Energy consumption and operating cost of the AnMBR system

3.2.1. Power requirements

Table 5 shows the power requirements of the AnMBR plant for
each of the five scenarios shown in Table 2 (FP1–FP5). This table also
illustrates the weighted average distribution for the energy
consumption of each particular equipment, i.e. pumps, blowers
and rotofilter. The dotted line between scenario FP3 and FP4 differ-
Fig. 2. Evolution of K20 and MLTS throughout 920 days of operation.
entiates the scenarios evaluated before and after chemically clean-
ing the membranes. Comparing the different scenarios assessed, it is
worth to say that scenarios studied prior to chemically cleaning the
referred to its daily average value. This experimental period is di-
membranes present higher energy consumptions (0.44, 0.32 and
vided into two stages, represented in Table 2 by a horizontal
0.49 kW h m3 for FP1, FP2 and FP3, respectively) than those stud-
dashed line. Energy consumption was firstly evaluated in a period
ied afterwards (0.20 and 0.19 kW h m3 for FP4 and FP5, respec-
of about 790 days, which was mostly operated at sub-critical filtra-
tively). This is mainly due to the higher J20 applied in the second
tion conditions (scenarios FP1–FP3). Overall, during this stage K20
operating stage whilst operating at similar SGDm. Specifically, the
decreased due to increasing membrane fouling over time (see days
specific gas demands per permeate volume (SGDP) resulted in the
300–790 in Fig. 2). Around day 790 the membranes were
range from 21 to 32 in scenarios FP1–FP3, decreasing to approx.
chemically cleaned. After this chemical cleaning, the energy con-
14 in scenarios FP4 and FP5.
sumption was evaluated in a period of about 140 days, which
Fig. 3 shows the weighted average distribution for the power
was operated at critical filtration conditions (scenarios FP4 and
requirements in the first (scenarios FP1–FP3 in Table 5) and second
FP5). During this second stage higher J20 were applied (see days
operating period (scenarios FP4 and FP5 in Table 5). This figure
790–920 in Fig. 2), making the AnMBR performance comparable
shows that the most important item contributing the power input
to full scale aerobic MBRs [13].
was the membrane tank biogas recycling blower, representing
Regarding the biological process, methane production increased
about two-thirds (60–75%) of the total AnMBR power require-
significantly when operating at both high temperature and high SRT
ments. The next in importance was the membrane tank sludge
(BP33 °C, SRT 70 days). To be precise, the average experimental methane
feeding pump, which represented about 15–20% of the total
production was 41.1, 16.8 and 8.5 LCH4 m3 for case BP33 °C, SRT 70 days,
AnMBR power requirements. Therefore, the main terms contribut-
BP22 °C, SRT 38 days and. BP17 °C, SRT 30 days (see Table 3), respectively. It
ing the total AnMBR power requirements were related to filtration
can be considered that an increase in the ambient temperature and/
(representing about 85–90%). This highlights the need of optimis-
or SRT leads to offset the low growth rate of MA [20]. In this respect,
ing filtration in any operating range to improve the feasibility of
simulation results in Fig. 1 show adequate effluent COD concentra-
AnMBR technology to treat UWW.
tions and increasing methane productions and decreasing sludge
To keep long operating periods without applying membrane
productions as temperature and/or SRT increases, and reducing
chemical cleaning (i.e. minimising irreversible fouling problems:
sludge production as temperature and/or SRT increases, within the
first 790 days in Fig. 2), low J20 and/or high SGDm are required. On
range of operating conditions evaluated in this study.
the other hand, increasing the chemical cleaning frequency allows
Concerning sludge production, low/moderate amounts of
operating at high J20 and/or low SGDm (i.e. low SGDP), which reduces
sludge were generated. As Table 3 shows, the sludge production re-
considerably the net energy demand (days 790–920 in Fig. 2). To be
sulted in 0.16, 0.43 and 0.55 kg TSS kg1 CODREMOVED in average for
precise, scenario FP5 was operated with the lowest SGDP (14.4),
cases BP33 °C, SRT 70 days, BP22 °C, SRT 38 days and. BP17 °C, SRT 30 days,
resulting in the lowest power input (0.19 kW h m3). Hence, it is
respectively. The minimum sludge production corresponded to
of vital importance to reduce the energy consumption by minimis-
case BP33 °C, SRT 70 days, due to operating at high temperature
ing SGDP, which indirectly increases the membrane chemical clean-
(33 °C) and high SRT (70 days). On the other hand, the experimen-
ing frequency. Nevertheless, increasing the frequency of membrane
tally determined %BVSS resulted in values below 35% within the
chemical cleaning means high chemical reagent consumption and
whole range of evaluated operating conditions, which indicated
may affect the membrane lifetime, resulting therefore in an increase
adequate sludge stabilities of the wasted sludge. For instance,
in membrane replacement and maintenance costs. Therefore,
%BVSS resulted in the highest value (31%) when operating under
further research is required to evaluate the most suitable AnMBR

Table 5
Power requirements in scenarios FP1–FP5.

Scenario Total energy Permeate Membrane tank biogas Membrane tank sludge Stirring power Anaerobic reactor Rotofilter
consumption pump (%) recycling blower (%) feeding pump (%) reactor (%) feeding pump (%) (%)
(k Wh m3)
FP1 0.44 2.34 73.15 14.54 8.20 0.52 1.25
FP2 0.32 1.26 73.18 14.69 8.43 0.72 1.73
FP3 0.49 1.61 73.94 14.58 8.27 0.47 1.13
FP4 0.20 1.38 61.73 21.02 11.89 1.17 2.81
FP5 0.19 3.06 67.46 16.19 9.18 1.21 2.90
 R. Pretel et al. / Separation and Purification Technology 126 (2014) 30–38 35

Fig. 3. Weighted average distribution for the AnMBR power requirements in scenarios: (a) FP1–FP3; and (b) FP4 and FP5.

operating strategy from an economical and environmental point of Therefore, operating at high ambient temperature and/or high
view including not only energy consumption but also investment SRT allows achieving significant energy savings whenever the
and maintenance costs. methane generated is captured and used as energy resource.

3.2.2. Net energy consumption 3.2.3. Operating cost

Fig. 4 shows the net energy consumption of the AnMBR for each Fig. 5 shows the operating cost of the AnMBR system including
of the five scenarios shown in Table 2 (FP1–FP5). This net energy energy recovery from methane (biogas methane and methane dis-
consumption includes both power requirements and energy recov- solved in the effluent) and sludge handling and recycling to land.
ery from methane. As mentioned earlier, each scenario (FP1–FP5) As Fig. 5 illustrates, the most favourable situation as regards
was evaluated for three different methane productions (BP33 °C, operating cost corresponded to case BP33 °C, SRT 70 days. By way of
SRT 70 days, BP22 °C, SRT 38 days and BP17 °C, SRT 30 days) and two different example, the operating cost in scenario FP5 when capturing both
levels of energy recovery (biogas methane, and biogas methane the biogas methane and the methane dissolved in the effluent
and methane dissolved in the effluent). was €0.011, €0.027 and €0.032 per m3 of treated water for cases
Fig. 4 shows considerable reductions in the AnMBR energy BP33 °C, SRT 70 days, BP22 °C, SRT 38 days and BP17 °C, SRT 30 days,
demand (in comparison with results shown in Table 5) whenever respectively. In this respect, savings of up to 64% from winter to
the generated methane is used as energy resource. For example, summer seasons could be achieved. This highlights the feasibility
the energy consumption in scenario FP5 was 0.19 kW h m3 when of AnMBR technology to treat UWW in warm climate regions, as
methane was not captured (see Table 5); whilst the net energy de- well as the necessity of optimising SRT for a given ambient temper-
mand in scenario FP5 decreased to 0.17 kW h m3 for case ature to maximise methane production and minimise sludge
BP17 °C, SRT 30 days when capturing both the biogas methane and production.
the methane dissolved in the effluent. In addition, operating at high On the other hand, it is worth pointing out the reduction in the
ambient temperature and/or high SRT further enhances the energy operating cost if energy is recovered from methane. To be precise,
balance of the system. For instance, the energy consumption in scenario FP5 for case BP33 °C, SRT 70 days resulted in an operating cost
scenario FP5 could be reduced up to 0.07 and 0.14 kW h m3 when of €0.028, €0.017 and €0.011 per m3 of treated water when consid-
recovering energy from both biogas methane and methane dissolved ering non-energy recovery from methane, energy recovery from
in the effluent for cases BP33 °C, SRT 70 days and BP22 °C, SRT 38 days, biogas methane, and energy recovery from biogas methane and
respectively (see Fig. 4b). methane dissolved in the effluent, respectively (see Fig. 5).
36 R. Pretel et al. / Separation and Purification Technology 126 (2014) 30–38

0.5 0.08

Operating cost ( /m 3)
Net energy demand (kWh·m -3 )

0.4 0.06

0.04
0.3

0.02
0.2
0
0.1 FP1 FP2 FP3 FP4 FP5

(a)
0
FP1 FP2 FP3 FP4 FP5 0.08

Operating cost ( /m 3)
(a)
0.06
0.5
0.04
Net energy demand (kWh·m-3)

0.4
0.02

0.3 0
FP1 FP2 FP3 FP4 FP5
0.2 (b)
0.1 0.08
Operating cost ( /m 3)
0.06
0
FP1 FP2 FP3 FP4 FP5
0.04
(b)
0.02
Fig. 4. Net energy consumption in scenarios FP1-FP5 for cases BP33 °C, SRT 70 days (j),
BP22 °C, SRT 38 ( ) and BP17 °C, SRT 30 days ( ) including energy recovery from: (a) 0
biogas methane; and (b) biogas methane and methane dissolved in the effluent. FP1 FP2 FP3 FP4 FP5

(c)
Therefore, the energy recovery from methane enables reducing
considerably the operating cost of AnMBRs treating sulphate-rich Fig. 5. Operating cost (net energy consumption and sludge handling and recycling
to land) in scenarios FP1–FP5 for cases BP33 °C, SRT 70 days (j), BP22 °C, SRT 38 days ( )
UWW at ambient temperature. This highlights the need of devel-
and BP17 °C, SRT 30 days ( ): (a) non-capture of methane; (b) energy recovery from
oping feasible technologies for capturing the methane dissolved biogas methane; and (c) energy recovery biogas methane and methane dissolved in
in the effluent stream not only to reduce its environmental impact the effluent.
(e.g. due to methane release to the atmosphere from the effluent),
but also to enhance the economic feasibility of AnMBR technology.
As previously commented, several simulation scenarios were
calculated in order to assess the AnMBR performance within the
whole range of temperature and SRT evaluated in this study.
l cost (€ /m )
3

Fig. 6 shows the simulation results regarding the theoretical influ- 0.08
ence of temperature and SRT on the AnMBR operating cost (when
treating sulphate-rich UWW), including energy recovery from 0.06
methane (biogas methane and methane dissolved in the effluent)
AnMBR operationa

and sludge handling and recycling to land. Specifically, this study 0.04
shows the results obtained for three SGDP levels (22.3, 33.4 and
14.4) corresponding to scenarios FP2, FP3 and FP4, respectively.
0.02 32
As shown in Fig. 6, from a biological process perspective, the oper- 29
ating cost is reduced when temperature and/or SRT increase; 26
0.00
)

23
(ºC

whilst, from a filtration process perspective, the operating cost is 65 60

55 50 20
T

reduced when SGDp decreases. 45 40

SRT ( 35 30 17
days
)
3.3. Effect of influent sulphate content on AnMBR operating cost
Fig. 6. AnMBR operational cost (power requirements, energy recovery from total
methane production, and sludge handling and recycling to land) at different
As mentioned before, Table 4 shows the total volume of meth- temperature and SRT conditions for three SGDP levels: ( ) SGDp 33.4; ( ) SGDp
ane produced (including both biogas methane and methane dis- 22.3; and (j) SGDp 14.3.
solved in the effluent) for the cases referred as BP33 °C, SRT 70 days,
BP22 °C, SRT 38 days and BP17 °C, SRT 30 days when treating low-sulphate
UWW (10 mg S–SO4 L1). Similar to treating high-sulphate UWW, methane increases significantly compared to treating high-sul-
methane production increases significantly when operating at high phate UWW (see Table 4 and Table 3).
ambient temperature and/or high SRT (BP33 °C, SRT 70 days). When Fig. 7 illustrates the operating cost of the AnMBR system when
treating low-sulphate UWW, since a little amount of COD is con- treating low-sulphate UWW. As Fig. 7 shows, a significant decrease
sumed by SRB, the amount of influent COD transformed into in the AnMBR operating cost could be achieved when treating
 R. Pretel et al. / Separation and Purification Technology 126 (2014) 30–38 37

low-sulphate UWW in comparison with treating high-sulphate demand of around 0.34 and 0.55 kW hvm3, which is low com-
UWW. For instance, for scenario FP5 and case BP33 °C, SRT 70 days, pared to the consumption of other full-scale municipal aerobic
the operating cost could be reduced from €0.017 per m3 (see MBRs (e.g. Running Springs MBR WWTP, USA, consuming around
Fig. 5c) to €0.001 per m3 (see Fig. 7c) when recovering energy from 1.3–3 kW h m3). On the other hand, the conventional activated
biogas methane. This highlights the possibility of improving the sludge system in Schilde (Belgium) consumed 0.19 kW h m3
feasibility of AnMBR technology when treating low/non sulphate- [21]. In our study, the theoretical minimum energy requirements
loaded wastewaters. treating sulphate-rich UWW resulted in 0.07 kW h m3. Therefore,
Mention must also be made of the potential of AnMBR to be net from an energy perspective, AnMBR operating at ambient temper-
energy producer (surplus electricity that can be exploited in other ature is a promising sustainable system compared to other existing
parts of the WWTP) when treating low-sulphate UWW. Specifi- urban wastewater treatment technologies. Nevertheless, it is
cally, Fig. 7c shows that when methane is captured from both bio- important to consider that the energy demand from the AnMBR
gas and effluent, scenario FP5 presents very low operating cost system evaluated in our study does not take into account the en-
(€0.006 per m3) for case BP17 °C, SRT, 30 days; whilst this cost de- ergy needed for nutrient removal, which it is considered in the
creases up to €0.002 per m3 for case BP22 °C, SRT 38 days. Moreover, wastewater treatment plants that has been mentioned as
null operating cost (or even income if the surplus energy is references.
exploited and/or sold to the market) could be achieved for case According to Xing et al. [22], sludge production in activated
BP33 °C, SRT 70 days: theoretical maximum benefit of up to €0.014 sludge processes is generally in the range of 0.3–0.5 kg TSSkg1
per m3. CODREMOVED. As expected, low/moderate amounts of sludge were
Therefore, in mild/warm climates (i.e. tropical or Mediterra- obtained in our study (0.16, 0.43 and 0.55 kg TSSkg1 CODREMOVED
nean), AnMBR technology is likely to be a net energy producer for cases BP33 °C, SRT 70 days, BP22 °C, SRT 38 days and. BP17 °C, SRT 30 days,
when treating low/non sulphate-loaded wastewaters: a theoretical respectively). Moreover, the produced sludge was considered sta-
maximum energy production of up to 0.11 kW h m3 could be ob- bilised, which allows, as mentioned before, its direct disposal on
tained by capturing the methane from both biogas and effluent. farmland without requiring further digestion.

3.4. Comparison with other existing technologies 4. Conclusions

According to recent literature [13], the full-scale aerobic MBR The results obtained reinforce the importance of optimising
from Peoria (USA) has a membrane and total aeration energy SGDP and SRT (for given ambient temperature conditions) to mini-
0.08
mise the energy requirements of AnMBRs treating sulphate-rich
UWW (minimum value: 0.07 kW h m3). Operating at high ambient
Operating cost ( /m 3)

0.06 temperature and/or high SRT allows achieving significant energy

savings whenever the methane generated is used as energy re-
0.04 source. Moreover, low/moderate sludge productions were obtained
(minimum value: 0.16 kg TSSkg1 CODREMOVED), which further en-
0.02 hanced the AnMBR operating cost (minimum value: €0.01 per m3).
On the other hand, the sulphate content in the UWW significantly
0
FP1 FP2 FP3 FP4 FP5 affected the final production of methane and thereby affected the
overall energy consumption. Indeed, AnMBR technology is likely
(a) to be a net energy producer when treating low/non sulphate-loaded
0.08
wastewaters in warm/hot climates: theoretical maximum energy
productions of up to 0.11 kW h m3 could be achieved.
Operating cost ( /m 3)

0.06

Acknowledgements
0.04

0.02
This research work was possible thanks to projects CTM2011-
28595-C02-01/02 (funded by the Spanish Ministry of Economy
0 and Competitiveness jointly with the European Regional Develop-
FP1 FP2 FP3 FP4 FP5 ment Fund) and aqualia INNPRONTA IISIS IPT-20111023 (partially
(b) funded by the Centre for Industrial Technological Development
(CDTI) and supported by the Spanish Ministry of Economy and
0.08 Competitiveness).
Operating cost ( /m 3)

0.06

0.04 References

0.02 [1] X.M. Guo, E. Trably, E. Latrille, H. Carrère, J.P. Steyer, Hydrogen production from
agricultural waste by dark fermentation: a review, Int. J. Hydrog. Energy 35
0 (2010) 10660–10673.
[2] A. Gallego, A. Hospido, M.T. Moreira, G. Feijoo, Environmental performance of
-0.02 wastewater treatment plants for small populations, Resour. Conser. Recy. 52
FP1 FP2 FP3 FP4 FP5 (2008) 931–940.
(c) [3] L. Raskin, Anaerobic Membrane Bioreactors for Sustainable Wastewater
Treatment, WERF, Report U4R08, June 2012.
[4] L.W. Hulshoff Pol, Treatment of sulphate-rich wastewaters: microbial and
Fig. 7. Operating cost (net energy consumption and sludge handling and recycling process technological aspects TMR summer school programme, The Biological
to land) in scenarios FP1–FP5 for cases BP33 °C, SRT 70 days (j), BP22 °C, SRT 38 days ( ) Sulfur Cycle: Environmental Science and Technology, April, Wageningen, The
and BP17 °C, SRT 30 days ( ) when treating low-sulphate UWW: (a) non-capture of Netherlands, 1998.
methane; (b) energy recovery from biogas methane; and (c) energy recovery biogas [5] J.B. Giménez, A. Robles, L. Carretero, F. Durán, M.V. Ruano, M.N. Gattib, J. Ribes,
methane and methane dissolved in the effluent. J. Ferrer, A. Seco, Experimental study of the anaerobic urban wastewater
38 R. Pretel et al. / Separation and Purification Technology 126 (2014) 30–38

treatment in a submerged hollow-fiber membrane bioreactor at semi- [13] S.J. Judd, C. Judd, Principles and Applications of Membrane Bioreactors in
industrial scale, Bioresour. Technol. 102 (2011) 8799–8806. Water and Wastewater Treatment, second ed., Elsevier, London, UK, 2011
[6] Miliarium Aureum, S.L, Portal de referencia para la Ingeniería Civil, Construcción (Chapters 3 and 5).
y el Medio Ambiente, <http://www.miliarium.com/prontuario/Tablas/Aguas/ [14] AGUA TÉCNICA, Equipos mecánicos para tratamiento de aguas, Tamiz rotativo
CaracterizacionAguasResidualesUrbanas.asp>, (accessed 06.02.14). para sólidos finos, <http://www.aguatecnica.es/index.php?option=com_
[7] A. Robles, M.V. Ruano, J. Ribes, J. Ferrer, Sub-critical long-term operation of content&view=article&id=50&Itemid=59>, (accessed 07.09.12).
industrial scale hollow-fibre membranes in a submerged anaerobic MBR (HF- [15] X. Elias-Castell, Vías de tratamiento y valorización de fangos de depuradora,
AnMBR) system, Sep. Purif. Technol. 100 (2012) 88–96. Ediciones Díaz de santos, Madrid, España, 2012.
[8] A. Robles, M.V. Ruano, J. Ribes, J. Ferrer, Factors that affect the permeability of [16] EPA, Environmental Protection Agency, Combined Heat and Power
commercial hollow-fibre membranes in a submerged anaerobic MBR (HF- Partnership, Agency of the United States federal government, 2012.
AnMBR) system, Water Res. 47 (2013) 1277–1288. [17] Spanish Electricity Tariff, Iberdrola S.A., 2013.
[9] J. Ferrer, A. Seco, J. Serralta, J. Ribes, J. Manga, E. Asensi, J.J. Morenilla, F. [18] J.A. Sainz-Lastre, Tecnologías para la sosteniblidad, Procesos y operaciones
Llavador, DESASS: a software tool for designing, simulating and optimising unitarias en depuración de aguas residuales, Colección EOI (2005).
WWTPs, Environ. Modell. Softw. 23 (2008) 19–26. [19] Miliarium Aureum, S.L, Portal de referencia para la Ingeniería Civil,
[10] R. Barat, J. Serralta, M.V. Ruano, E. Jiménez, J. Ribes, A. Seco, J. Ferrer, Biological Construcción y el Medio Ambiente, <http://www.miliarium.com/Proyectos/
nutrient removal model no 2 (BNRM2): a general model for wastewater Depuradoras/introduccion/edar.asp>, (accessed 05.05.13).
treatment plants, Water Sci. Technol. 67 (2013) 1481–1489. [20] G. Lettinga, S. Rebac, G. Zeeman, Challenge of psychrophilic anaerobic
[11] Standard methods for the Examination of Water and Wastewater, 21th ed., wastewater treatment, Trends Trends Biotechnol. 19 (2001) 363–370.
American Public Health Association/American Water Works Association/ [21] A. Fenu, J. Roels, T. Wambecq, K. De Gussem, C. Thoeye, G. De Gueldre, B. Vand De
Water Environmental Federation, Washington DC, USA, 2005. Steene, Energy audit of a full scale MBR system, Desalination 262 (2010) 121–128.
[12] J.B. Giménez, N. Martí, J. Ferrer, A. Seco, Methane recovery efficiency in a [22] C.-H. Xing, W.-Z. Wu, Y. Quian, E. Tardieu, Excess sludge production in
submerged anaerobic membrane bioreactor (AnMBR) treating sulphate-rich membrane bioreactors: a theoretical investigation, J. Environ. Eng. ASCE 129
urban wastewater: evaluation of methane losses with the effluent, Bioresour. (2003) 291–297.
Technol. 118 (2012) 67–72.