Download as pdf or txt
Download as pdf or txt
You are on page 1of 11

Renewable and Sustainable Energy Reviews 110 (2019) 132–142

Contents lists available at ScienceDirect

Renewable and Sustainable Energy Reviews


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

Customizing anaerobic digestion-coupled processes for energy-positive and T


sustainable treatment of municipal wastewater
Bao-Cheng Huanga, Wen-Wei Lia,∗, Xu Wangb, Yan Luc, Han-Qing Yua,∗∗
a
CAS Key Laboratory of Urban Pollutant Conversion, Department of Applied Chemistry, University of Science & Technology of China, Hefei, 230026, China
b
State Key Joint Laboratory of Environment Simulation and Pollution Control, Research Center for Eco-Environmental Sciences, Chinese Academy of Sciences, Beijing,
100085, China
c
School of the Environment, Nanjing University, Nanjing, 210023, China

ARTICLE INFO ABSTRACT

Keywords: Anaerobic digestion (AD) technologies offer exciting opportunities for implementation of sustainable wastewater
Anaerobic digestion management, which are currently still an energy-intensive and costly process. However, selecting an appropriate
Energy recovery AD process is challenging because the treatment performances vary significantly with the technology type,
Municipal wastewater wastewater properties and environmental conditions and a trade-off between the treatment efficiency and other
Sustainability
cost and environmental impacts is usually applied. In this study, a multi-criteria evaluation method which gives
Customization
a balanced consideration between multiple sustainability dimensions and the practical implementation condi-
tions for a specific treatment process was developed for AD-coupled processes evaluation. Six indicators across
three metrics (i.e., environmental, economic, and technical performances) were covered in the method. Two
representative AD-coupled processes for municipal wastewater treatment and four countries (United States,
United Kingdom, China, and India) were selected for the evaluation. A national-specific evaluation shows that
the high-rate activated sludge + sludge anaerobic digestion always outperforms the other processes (presently
and in the near future) except for India, where up-flow anaerobic sludge blanket + activated sludge process
stands out as the currently optimal option based on the 2015 data. Moreover, it is suggested that implementing
AD-coupled nutrient recovery technologies will be economically more feasible for the developed countries than
the developing ones, because the latter will still prioritize the economically- and environmentally-beneficial
technologies in the near future. This work may help identifying key area of improvement needed for municipal
wastewater sectors in different countries and provide guidance for national and global policy-makers to promote
customized implementation of AD-coupled processes towards energy-positive and sustainable wastewater
treatment.

1. Introduction [7,8]. Especially, the anaerobic digestion (AD) processes for methane
production are sensitive to environmental upset [9]. Thus, pollutant
As a key link of the global cycling of water resources, wastewater conversion and methanogenic activity are readily restricted by low
treatment offers essential protection for the urban environmental sus- temperature [10] and complex MW composition [11], resulting in low
tainability. Unfortunately, the treatment processes themselves are energy production. These factos restrict the application of AD processes
usually costly and energy intensive. Especially, the currently prevailing in practical MW treatment today. As a trade-off between aerobic and
activated sludge (AS) process for municipal wastewater (MW) treat- anaerobic processes, their hybrid processes are widely accepted as a
ment presents a huge energy and economic burden for the society more feasible option for practical MW treatment [12]. Such coupled
[1–3]. Changing this situation entails a paradigm shift from aerobic to processes, where AD technologies are implemented either at side-
anaerobic treatment, due to the avoided aeration and possibilities for stream (i.e., sludge anaerobic digestion, SAD) or mainline (i.e., direct
energy and resource recovery in anaerobic processes [4–6]. However, treatment of dilute wastewater) of MW treatment line, enable good
anaerobic processes for MW treatment also have drawbacks such as effluent quality and energy performance and have been increasingly
unstable effluent quality, higher investment, and long start-up time implemented in wastewater treatment plants (WWTPs). Two most


Corresponding author.
∗∗
Corresponding author.
E-mail addresses: wwli@ustc.edu.cn (W.-W. Li), hqyu@ustc.edu.cn (H.-Q. Yu).

https://doi.org/10.1016/j.rser.2019.04.064
Received 6 November 2018; Received in revised form 7 April 2019; Accepted 22 April 2019
Available online 02 May 2019
1364-0321/ © 2019 Elsevier Ltd. All rights reserved.
B.-C. Huang, et al. Renewable and Sustainable Energy Reviews 110 (2019) 132–142

List of abbreviations GHG greenhouse gas


HRAS high-rate activated sludge
AD anaerobic digestion MCE multi-criteria evaluation
AOTR actual oxygen transfer rate MW municipal wastewater
AS activated sludge SAD sludge anaerobic digestion
BOD5 5-day biochemical oxygen demand SRT solids retention time
CAS conventional activated sludge UASB up-flow anaerobic sludge blanket
CHP combined heat and power US united states
ENPV economic net present value VSS volatile suspended solid
GDP gross domestic product WWTPs wastewater treatment plants

successful examples are the high-rate activated sludge (HRAS) + SAD evaluating the integrated sustainability of different AD-coupled pro-
process (demonstrated in several WWTPs in United States of America cesses was developed. With this method, the possible sustainability
(USA) [13] and Austria [14]) and the up-flow anaerobic sludge blanket benefits of the AD-coupled processes were evaluated and the future
(UASB) + AS process (applied in several tropic countries for mainline opportunities for implementation in practical MW facilities were dis-
AD treatment [15,16]). cussed.
Although both AD-coupled processes promise energy recovery from The MCE method took into account the environmental, economic,
MW, their actual energy performances and implementation feasibilities and technical metrics covering six key indicators. Each indicator was
are case specific. The bioenergy production of AD process could vary scored and weighted based on the local conditions for different coun-
drastically with the MW characteristics (e.g., organic carbon strength) tries to yield an integrated score. The highest-scored process was con-
and environmental conditions (e.g., temperature). The higher tem- sidered as the optimal one for a specific country. Four representative
perature in tropic regions favors the implementation of UASB + AS countries that together contribute to more than a half of the produced
process [17]. In addition, the sustainability of a MW treatment process MW globally were selected for the MCE study. Besides the large-sized
is not only about energy, but also covers the environmental, economic MW sector, these countries are in significantly different economic de-
and social metrics and other aspects of technology implementation velopment levels, and have different wastewater properties and en-
(e.g., technology resilience, scalability and ease of integration). Hence, vironmental conditions to facilitate the comparison. USA is the world's
developing an appropriate evaluation criterion based on multiple me- largest developed country with a high MW organic strength, while
trics of sustainability is essential for implementing case-specific, sus- China is the world's largest developing country with relatively dilute
tainable AD technologies. MW. United Kingdom (UK) and India also differ significantly not only in
The performances and application potential of AD-coupled pro- the economic development level but also in the temperature due to
cesses have been evaluated in several earlier studies [18,19]. However, their different geographic latitudes. It should be noted that these four
such evaluations were mostly based on individual technical, econom- countries were studied as an example only, and the established MCE
ical or environmental indicators and did not taking into account the method might be readily extended to other countries or regions. In all,
diversity in wastewater properties, environmental conditions and eco- this work lays a theoretical basis for case-specific, rational selection of
nomic development levels that might significantly affect the technology AD-coupled processes for MW treatment, and might provide valuable
performance. In addition, life-cycle assessment has been widely applied implications for future wastewater management at global and national
to evaluate the technical performance and environmental footprint of scales.
MW treatment processes, but mostly were case studies without con-
sidering the global diversity in WWTP conditions [20–22]. A global-
2. The AD-coupled MW treatment flow
scale, case-specific evaluation on the comprehensive sustainability of
wastewater treatment processes is still lacking. This might be an im-
2.1. Treatment flow and operating parameters of AD-coupled processes
portant reason for the blind development and usually suboptimal im-
plementation of AD technologies for MW treatment at the present stage.
In the CAS process, MW flows through the aeration tank and sec-
Optimization of the technical sustainability requires a balanced con-
ondary clarifier, and the settled sludge from the secondary clarifier is
sideration among environmental, economic, and technical impacts, and
recirculated to the aeration tank to maintain a certain biomass con-
such impacts might be further compounded by the involvement of
centration (Fig. 1a). Diffused air aeration is usually adopted to ensure
various stakeholders [23]. Multi-criteria decision analysis, which has
sufficient oxygen supply. The residual sludge is regularly discharged
been used to aid the optimization of energy production and storage
from the system to keep a relatively constant sludge biomass. Here, the
technologies [24,25], may offer as a suitable tool to solve such multi-
average solids retention time (SRT) is set as 12 days to guarantee suf-
objective optimization problems.
ficient organic matter (represented by 5-day biochemical oxygen de-
Hence, this work aims to develop a multi-criteria evaluation (MCE)
mand, BOD5) and good ammonium removal performance. By contrast,
method that covers the environmental, economic, and technical metrics
a very short SRT (1 day) is adopted for the HRAS process, where the
for assessing the integrated sustainability of AD-coupled processes and,
organic matters are mainly absorbed instead of degraded by the sludge
based on this method, to evaluate the suitability of different treatment
[26]. The organic carbon-rich residual sludge from both the CAS and
processes for individual countries. Specifically, three AD-coupled pro-
HRAS processes is thickened and transported to the AD tank for biogas
cesses, including HRAS+SAD, UASB+AS and the currently prevailing
production, followed by the dewatering treatment. A primary sedi-
process of conventional activated sludge (CAS) +SAD, were selected
mentation tank is placed before the aeration tank to remove the
for the technology comparison. Although other alternative AD pro-
settleable solids in raw MW for the CAS+SAD and HRAS+SAD pro-
cesses are available, the HRAS+SAD and UASB+AS were selected here
cesses, but such tank is canceled for the UASB+AS process (Fig. 1b).
mainly due to their relative technical maturity and hence higher social
In the UASB reactor, the MW flows upwards and BOD is removed by
acceptance. To identify the impacts of MW characteristics and en-
anaerobic microorganisms, which yields only a small amount of sludge.
vironmental conditions on the treatment performance, the energy
The biomass is separated by gravity in the settling zone at the upper
performances of the three AD-coupled processes at various MW organic
part of the reactor and returned to the reaction zone, while the clarified
strengths and temperatures were examined. Then, an MCE method for
effluent is discharged. The residual sludge is anaerobically stabilized

133
B.-C. Huang, et al. Renewable and Sustainable Energy Reviews 110 (2019) 132–142

mainly in the form of electricity and thermal energy (Ep, kWh/m3) [29].
The amount of produced bioenergy can be estimated as:
Ep = BODinf XR (1)
3
where BODinf (g/m ) is the influent BOD5 concentration; X is the BOD5
removal efficiency; R (mol-CH4 recovered/g BOD5removed) is the me-
thane production; ξ is the theoretical energy value of CH4, which is
0.222 kWh/mol CH4; η is the CHP efficiency, for which a value of 33%
electric efficiency and 38% heat efficiency is taken assuming a state-of-
the-art reciprocating engine CHP is used [29]. Dissolved methane in the
effluent was estimated according to the Henry's law.
Total energy consumption (qener, kWh/m3-MW) includes the energy
for aeration in AS (qae), pumping (qpe), residue sludge handling (qsh),
mixing (qmix), and recovery of dissolved methane by degassing mem-
brane (qdm). qae was calculated according to Metcalf & Eddy [30]:
0.283
AOTR RT1 P2 24
qae = 1
EA a Qw 3600 29.7n P1 Q (2)
Fig. 1. Schematic illustrations of (a) side-stream AD (CAS+SAD and HRAS 3
where EA is the aerator oxygen diffusion efficiency, ρa (kg/m ) is the air
+SAD), and (b) mainline AD (UASB+AS) processes. density, Qw (kg O2/kg air) is the oxygen weight in the air, R (J/mol·K) is
the engineering gas constant for air, n is 0.283 for air, Q (m3/d) is in-
before being discharged from the system. To ensure sufficient removal fluent flow rate, P1 and P2 are the suction and discharge pressure for the
of pollutants, UASB is typically followed by an AS process. No ni- aerator (pa). AOTR (kg O2/h) is the actual oxygen transfer rate under
trification takes place in the subsequent AS unit to ensure good effluent field conditions.
quality. The detailed operational parameters of the above processes are qpe (kWh/m3) was estimated with the following equation [31]:
listed in Table 1. 1 24
qpe = P
Q (3)
2.2. Methods for evaluating the impacts of MW strength and temperature on
3
the energy performances of AD-coupled processes where is the density of permeate (kg/m ), is the pumping energy
efficiency (0.85 in this work), is the conversion achieved by a single
Before evaluating the integrated sustainability of the AD-coupled passage of fluid along the length of the module (normally 100% for
processes, the energy performance, a key indicator of the technical submerged systems), P is the sum of all individual pressure losses
metric, in response to different MW organic strengths was first in- (Pa).
vestigated. The BOD concentrations of MW vary geographically. In qsh is the energy consumption for gravity thickening (EGT) (kWh/
general, higher MW BOD5 (up to 560 mg/L) is found in the developed m3) and centrifuge dewatering (ECD) (kWh/m3). It could be estimated
countries over the developing ones (mostly below 350 mg/L) [27]. as follows [18]:
Based on the reported MW BOD5 range, three BOD5 concentrations 422832 (Flowinf , MGD )0.9248 1
(230, 350 and 560 mg/L) were selected to represent the low-, medium-, EGT =
365.25 Q (4)
and high-strength MW, respectively. The detailed compositions of the
three types of MW are given in Table 2. 5024825 (Flowinf , MGD ) + 39693 1
Temperature is another crucial factor governing methane produc- ECD =
365.25 Q (5)
tion performance in AD processes (Table S1). Low temperature not only
impairs the microbial activity but also lowers the methane recovery where Flowinf (MGD) is the sludge flow, Q (m3/d) is the influent flow
ratio due to increased dissolved methane. Considering that the annual- rate.
average temperatures in most parts of the world are below 40 °C and the The energy consumption for mixing was assumed to be 0.007 kW/
methanogenic activity is severely suppressed at < 8 °C [28], four tem- m3 reactor volume [30]. To evaluate the energy potential for recovering
peratures (10, 15, 25, and 35 °C) were used here to evaluate the tem- dissolved methane in effluent, degassing membrane is selected as po-
perature impacts on AD-coupled processes. tential technology by assuming a methane recovery efficiency of
The model calculation was based on the construction and operation 70–90% [32–34]. The energy consumption of such methane degassing
of a hypothetical, small-sized WWTP. The WWTP treatment capacity is processes (qdm, kWh/m3) was estimated as [34]:
assumed to be 2 × 104 m3/d, the minimal size required for an eco- y / NS
NS qs P2 24
nomical operation of combined heat and power (CHP) system for biogas qdm = 354 1
P1 Q (6)
combustion [29]. The residual sludge is subjected to AD treatment at
35 °C with the following operating parameters: methane production of
where Ns is the number of vacuum stages (1); qs is the flow rate at the
0.3 m3/kg VSS; sludge yield of 0.6, 0.6 and 0.05 g VSS/g BOD removed
for CAS+SAD, HRAS+SAD, and UASB+AS processes, respectively
Table 1
[30]. Main operation parameters for the studied AD-coupled processes.
AD coupled process Aeration tank Primary
3. Algorithm and model construction
clarifier
DO (mg/ SRT (day) Nitrification
3.1. Algorithm for estimating energy performances and greenhouse gas L) function
(GHG) emission
CAS+SAD 2 12 ✓ ✓
HRAS+SAD 1.5 1 × ✓
In the AD-coupled processes, methane-rich biogas can be obtained, UASB+AS 2 7 × ×
and a further combustion of the biogas results in bioenergy generation

134
B.-C. Huang, et al. Renewable and Sustainable Energy Reviews 110 (2019) 132–142

Table 2 economic, and technical metrics for integrated sustainability quantifi-


Compositions of MW with different organic strengths [27]. cation into account is illustrated in Fig. 2. MCE provides a useful tool
Item MW strength (mg/L) for evaluating and optimizing complicated wastewater treatment pro-
cesses [51,52]. Here, six indicators across environmental, economic,
Low Medium High and technical metrics were taken into account in MCE (Table S2). These
indicators (including land footprint, residue sludge production, GHG
Suspended solids, total 250 400 600
Suspended solids, volatile 200 320 480 emission, construction and operation cost, bioenergy production, and
Biochemical oxygen demand, 5 day 230 350 560 ease of operation) constitute the key factors that govern the sustain-
Chemical oxygen demand 500 750 1200 ability of MW treatment processes [52,53]. Another criterion for se-
Nitrogen (total as N) 20 45 75 lecting these indicators is that their metric weights can be derived from
Organic nitrogen 10 15 25
processing global databases, implying that the integrated scores here
Ammonia 10 30 50
Phosphorus (total as P) 6 15 25 can be updated for the following years to track the changes of national
sustainability in implementing AD-coupled processes.
The MCE score for each treatment process was obtained with the
suction end of the vacuum pump (m3/s); η is vacuum pump efficiency following formula:
(0.65); P1 and P2 are the suction and discharge pressure for the vacuum n
pump (pa), respectively; y is the isentropic expansion of gas. MCE(a) = wi Pi
The total GHG emission (qghg-total, kg CO2-equivalent/m3 MW) in- i (9)
cludes the on-site and off-site emissions. The on-site GHG emissions are
where Pi is the normalized value of indices i; wi is the weight coefficient;
mainly from methane combustion via CHP (qghg-mc, kg CO2-equivalent/
subscript n is the total number of indices (n = 6). The raw value of
m3 MW), direct methane (qghg-methane, kg CO2-equivalent/m3 MW) and
indices i is normalized to obtain Pi:
N2O emissions (qghg-N2O, kg CO2-equivalent/m3 MW). The direct me-
thane emissions are from the dissolved methane in the effluent and the Pi =
qi
m
leaked part during transportation and CHP operation. The latter ac- i
qi (10)
counted for about 3% of total methane production [35,36]. The ni-
trification process in the CAS process involves N2O emission, for which where qi is the raw value of indices i, and subscript m is the total
a N2O emission factor of 0.6% (relative to the total N-load) was adopted number of evaluated treatment systems. Due to the metric difference of
according to the literature [37]. The global warming potential of me- different indices, the reciprocals of qeco, qpro, qlan, qslu, and qghg were
thane and N2O is set as 28 and 298 times of CO2, respectively [38]. The adopted to calculate Pi to make it consistent with qbio. The importance
off-site GHG emission (qghg-off, kg CO2-equivalent/m3 MW) was calcu- of each indicator is ascertained via assigning a weight coefficient (wi).
lated based on the electricity consumption: All the weight coefficients were summed to be 1 following the “linear
weighted sum” rule [51]. wi were determined based on the national-
qghg off = qener fi (7) level data:
3
where qener (kWh/m -MW) is the total energy consumption, fi (kg CO2/ winormanized
kWh) is the estimated GHG emission intensity from electricity genera- wi = n
winormanized (11)
tion (Table 3). i

winormanized (dimensionless) is ascertained as:


3.2. MCE evaluation on sustainability of AD-coupled processes in different
countries wiobjective
winormanized =
wiaverage (12)
3.2.1. Estimation of MW strength and temperature in the studied countries
where is the raw value for a specific country in the objective
wiobjective
At a global scale, the MW discharge behaviors (i.e., BOD, nitrogen, year, which was obtained from the international databases; wiaverage is
and phosphorus discharge per capita) are in general correlated to the the average value of operator i in the past 25 years (1990–2015). To
local income level with certain uncertainty [41,42]. Therefore, the estimate the weight coefficients in 2025, the autoregressive integrated
pollutant discharge amounts in the four representative countries were moving average was adopted to fit the historic trends and then obtain
estimated according to the economic development levels as classified the predicted results. The uncertainties were analyzed by using 100000
by the World Bank [43]. The sigmoid curve which indicates the MW Monte Carlo stimulations with Crystal Ball software (Edition 11.1.2.4,
production intensity (m3/capita/day) with income (GDP per capita) Oracle Co., USA).
was modeled (Fig. S1) by using historical data [44]. Then, the total MW In this way, the different AD-coupled processes were scored on a
production was estimated by multiplying the MW production intensity national-specific basis. A higher-scored technology implies it has a
with the forecasting population [45,46]. greater potential to be implemented in a specific country towards in-
The MW temperature distribution was approximated from the am- tegrated sustainability. Thus, the optimal (highest-scored) technology
bient temperature using the following nonlinear model [47,48]: for different countries can be identified, laying a basis for customization
µ
Tw (MW ) = µ + BTa
1+e ( Ta) (8) Table 3
Estimated GHG emission intensity from electricity generation according to
where Tw (MW) is the estimated MW temperature; Ta is the measured
International Energy Agency [39] and BP [40] reports.
ambient temperature; µ and are the minimum and maximum MW
temperature, respectively; is the steepest slope of the function; re- Country Electricity GHG emissions of electricity GHG emission
production and heat production intensity
presents the ambient temperature at the inflection point. The model
was established and validated with literature data [49]. The annual (TWh) (million t CO2) (g-CO2/kWh)
average ambient temperature data were acquired from the Climatic
Research Unit [50]. China 5649 4384 769
India 1252 1046 835
USA 4363 2125 489
3.2.2. Procedures for the MCE UK 338 145 429
A framework of the MCE method that takes all the environmental,

135
B.-C. Huang, et al. Renewable and Sustainable Energy Reviews 110 (2019) 132–142

Fig. 2. The framework of the multi-criteria evaluation on the sustainability of MW treatment process.

Fig. 3. Energy balance and net energy production performances of (a) CAS+SAD, (b) HRAS+SAD, and (c) UASB+AS processes under different MW temperature and
organic strength conditions.

136
B.-C. Huang, et al. Renewable and Sustainable Energy Reviews 110 (2019) 132–142

of AD-coupled processes. ammonium recovery while iron-based phosphorus adsorption is applied


for main-line P recovery.
3.2.3. Data sources
The projected population of each country was collected from the 4. Results and discussion
United Nations Population Division [54]. The GDP data were acquired
from the US Department of Agriculture Economic Research Service 4.1. Energy performances of AD-coupled processes at various MW organic
[55]. To calculate the weight coefficients, the data of each country in strengths and temperatures
the following indicators over the past 25 years were collected: total
renewable electricity production (billion kWh/capita) [56], gross do- For all the studied processes, both the energy production and con-
mestic product ($/capita) [57], occupied area per capita (km2/capita) sumption increase with the raising organic strength of MW, but the
[58], government effectiveness (dimensionless) [59], municipal wastes increments vary for different processes. In the CAS+SAD process, the
production (million t/year) [60,61], and GHG emission (million t CO2/ energy consumption (mainly for aeration and heating sludge digester)
capita) [40]. increases more than its production, resulting in net energy consumption
(Fig. 3a). In comparison, the HRAS+SAD and UASB+AS processes
3.3. Cost-benefit analysis of the AD-coupled nutrients recovery scenarios in need slightly higher energy consumption but enable significantly im-
different countries proved bioenergy production due to the strengthened methane pro-
duction under high BOD conditions (Fig. 3b and c). Raising the MW
Economic net present value (ENPV) is used to evaluate the potential temperature also significantly affects the energy performances of the
economic impact throughout the entire life-time of WWTP [62]. The three processes. Specifically, the CAS+SAD process shows more net
ENPV is estimated as: energy consumption due to the increased energy consumption and
n n decreased production; the energy balance of the HRAS+SAD process
ENPV = Bt (1 + i 0) t Ct (1 + i 0) t remains almost unchanged; the UASB+AS achieves more net energy
t=0 t=0 (13) production.
The above results suggest that the HRAS+SAD and UASB+AS
where Bt is the total benefit; Ct represents the economic cost; i0 is the
processes hold a great potential for net energy production. Especially,
social discount rate, which is set as 7% according to US Office of
the UASB+AS process shows the highest net energy production when
Management and Budget [57]; t is the computation period, assuming 3
the MW organic strength and temperature are both high. Thus, from the
years construction period for a new-built project. The WWTP service-
perspective of maximizing energy production, the HRAS+SAD and
life is assumed to be 40 years. Total benefit for installation of a re-
UASB+AS processes have different optimal application ranges in terms
source-oriented treatment project can be estimated as:
of temperature and MW organic strength (Fig. 4). A case-specific se-
Bt = Bpoll + Bele + Bnutrients (14) lection of AD-coupled processes is thus very necessary.
In light of the global diversity in MW organic strength and tem-
where Bpoll ($/yr) is the benefit for pollution removal (BOD5, SS, TN,
perature, customizing the AD-coupled processes for different countries
and TP elimination); Bele ($/yr) is the benefit for sale of generated
is highly desirable. As a first step, tailored processes were assumed for
electricity; and Bnutrients ($/yr) is the benefit stems from the potential
the four representative countries targeting on energy performance op-
value of recovered nutrients.
timization. The studied countries differ significantly in their MW or-
The potential environmental benefits from elimination of BOD5, SS,
ganic strengths and temperatures due to geographical and diet differ-
TN, and TP were assumed to be 0.0388, 0.0059, 17, and 31 $/kg, re-
ence (Table S3). The evaluation results show that HRAS+SAD and
spectively [63,64]. The electricity price was set as 0.215 $/kWh ac-
UASB+AS processes both enable net energy production in India, USA,
cording to the electricity price for industrial consumers in Eurostat
and UK. However, the UASB+AS is a net energy-consuming process for
database [65]. The recovered N and P were used as an alternative to
China due to its relatively low MW organic strength (Fig. 5a). Based on
conventional chemical fertilizer and their values were assumed to be
the net energy production result, the UASB+AS appears to be the most
212 and 119 $/kg [66].
energy-beneficial process for USA, whereas the HRAS+SAD should be
Total cost (Ct) of energy recovery project is calculated as:
the best option for the rest countries.
Ct = Ccon + Co & m + Cloan + Ctax (15) Besides the energy performance, the economic and environmental

where Ccon is the construction costs; Co&m is the operation and main-
tenance costs; Cloan is the loan interests; and Ctax is the tax fee.
Ccon and Co&m are estimated via CpadetWorks (Version 3.0,
Hydromantes Environmental Software Solutions Inc., Canada). It is
assumed that the different costs for constructing a project in developed
and developing countries are attributed mainly to different labor costs.
The average labor costs are relatively cheaper in developing countries.
The labor costs for developing countries (10 $/hour) and developed
countries (33 $/hour) were set according to the International Labor
Organization data [67]. The capital for project construction was ob-
tained from bank loan. The loan was assumed to be paid off within 10
years and the corresponding interest for a developed and developing
country was 5.479% and 4.251%, respectively [68]. The tax cost for
electricity sale was set as 16.3% according to the World Bank Group
report [69]. All the economic value or prices were converted through
US consumer price index adjustment.
In this work, the ENPVs of the conventional biological nutrients
removal process (Anaerobic/Anoxic/Oxic) and nutrients recovery cou-
pled HRAS+SAD process were calculated and compared. According to Fig. 4. Suitable regions for the AD-coupled processes to maximize energy re-
the US EPA report [70], the ion-exchange process is used for covery from MW.

137
B.-C. Huang, et al. Renewable and Sustainable Energy Reviews 110 (2019) 132–142

Fig. 5. Estimated performances of AD-coupled processes in four representative countries for different indicators: Net energy production (a), GHG emission (b),
construction and operation costs (c), residue sludge production (d).

Fig. 6. Temporal trends of weight coefficients for metric assessment between different countries (a-China, b-India, c-USA, and d-UK).

138
B.-C. Huang, et al. Renewable and Sustainable Energy Reviews 110 (2019) 132–142

Table 4
Integrated scores of AD-coupled processes for different countries.
CAS+SAD HRAS+SAD UASB+AS

China 0.32 0.39 0.29


India 0.30 0.34 0.35
USA 0.30 0.36 0.32
UK 0.32 0.38 0.30

Fig. 9. Economic net present value estimation for constructing a 20000 m3/d
scale plant with conventional and nutrients recovery processes.

appropriately weights the different criteria for maximizing the process


sustainability.

4.2. Multi-criteria evaluation of AD-coupled processes

MCE was performed to evaluate the integrated sustainability of


different AD-coupled processes in the four countries by taking into six
indicators across the environmental, economic, and technical metrics
account (Fig. 2). To set appropriate weight coefficients for the different
indicators, national-specific historical data (during 1990–2015) were
used. The results show that the coefficients of construction and opera-
Fig. 7. Potential merits comparing to CAS process by application of suitable AD tion costs kept growing for China (from 0.04 to 0.29) and India (from
process. The concentric lines are at intervals of 0.2 and a higher polygonal area 0.09 to 0.26) over the years, but the growth rates slowed down in the
indicates a better performance. 1-financial profit (0.1 $/m3), 2-bioenergy pro- recent five years. In contrast, USA and UK showed a decreasing trend in
duction (kWh/m3), 3-GHG reduction (kg CO2/m3), 4-residue sludge reduction this coefficient lately (Fig. 6). The relatively higher values for China and
(kg/m3), 5-total energy saving (kWh/m3). India imply that the WWTP construction and operation still present a
significant financial burden for the developing countries. The GHG
performances are also important considerations for evaluating MW emission weight in UK decreased from 0.23 (1990) to 0.08 (2015),
treatment technologies. Especially, for developing countries like China which should be mainly ascribed to increased share of renewable en-
and India, these aspects are of even higher national priority because of ergy (from 23% in 1990 to 47% in 2015) in this country [71]. Appar-
the severe environmental challenges and economical restriction. The ently, the priorities and weight coefficients of the different indicators
incorporation of economic and environmental metrics in the evaluation change dynamically and differ significantly among the four countries
would significantly affect the evaluation results and technology due to their different social-economic development levels and priorities.
selection.Taking the UASB+AS process as an example, this process is Based on the national-specific weight coefficients and the perfor-
favorable for energy production only under high organic strength and mance of each indicator in 2015, the overall MCE scores of the different
temperature conditions. It stands out as the best option in terms of AD-coupled processes were estimated to quantify the integrated sus-
residue sludge production and costs (Fig. 5c and d) but shows the tainability (Table 4). The MCE score of the same process varied sig-
poorest performance in GHG emission. Such a conflicting result for nificantly for different countries. Taking the UASB+AS process as an
different evaluation indicators implies a necessity for MCE that example, it fell between the HRAS+SAD and CAS+SAD in MCE score

Fig. 8. Predicted weight coefficients (a) and integrated scores of different technologies to be implemented in different countries (b) in the future (2025). The center
lines are median values while the boxes represent 10th to 90th percentiles.

139
B.-C. Huang, et al. Renewable and Sustainable Energy Reviews 110 (2019) 132–142

for USA. However, it stands out as the highest-score option when ap- (Fig. 9). Therefore, in the near future, implementation of AD-coupled,
plied in India (benefited from the small land footprint and low con- resource-oriented processes seems to be economically more preferable
struction and operation costs) and the most unfavorable one for China for developed countries. For the developing world, the shift to AD-
and UK. Thus, a national-specific customization of the AD-coupled coupled processes would depend strongly on the development of low-
processes is essential. cost technologies for energy and nutrients recovery. Hence, it might be
Apart from the overall scores, the contribution of individual in- more desirable for the developing countries to prioritize energy-saving
dicators to integrated sustainability (Fig. S2) might also provide useful nutrient removal technologies. Currently, many promising technologies
information for identifying the areas of improvement for different have emerged in this direction. For example, the decentralized treat-
countries. For instance, in the cases of the HRAS+SAD and CAS+SAD ment and source separation practices are expected to significantly lower
processes, it would be more desirable for UK to prioritize the im- the initial investment (20–50% of conventional plants) and operation
provement of indicators other than bioenergy production, but in China costs (5–25% of CAS based plants) of wastewater facilities [74–76]. The
a higher priority on the reduction of land footprint and residue sludge development of enhanced methanogenesis process and anaerobic
production is preferred. Such individual indicator evaluation would be membrane bioreactors also provide opportunities for direct mainline
helpful to identify key areas of improvement aiming at higher in- treatment of MW with improved energy efficiency [4,77–79]. In addi-
tegrated sustainability of the treatment process. tion, technologies are already available to further lower the soluble
Next, the highest-scored process for each country was selected to methane loss (Fig. S3) and to improve the energy efficiency of mainline
evaluate the best-achievable performances of MW treatment in the in- AD treatment. These new technological advances may greatly motivate
dividual countries. The CAS process was used as a control for quanti- the development and global implementation of AD-coupled processes
tative comparison. The results show that, with the implementation of for sustainable MW treatment.
the AD-coupled process with the best integrated score, India would
achieve the highest financial profit (0.06 $/m3) among the four coun- 5. Conclusions
tries (Fig. 7b). In comparison, USA would achieve the highest profits in
terms of total energy saving, bioenergy production, GHG reduction, and In this work, a multi-criteria evaluation method was developed to
residue sludge reduction. To improve the overall benefits for im- evaluate the country-specific integrated sustainability of several anae-
plementing suitable AD-coupled process, different strategies should be robic digestion-coupled processes for municipal wastewater treatment.
taken for these countries. For example, China and UK should prioritize By taking into account the evaluation criteria of environmental, eco-
MW treatment processes that enable low cost, GHG emission reductions nomic and technical performances, the multi-criteria evaluation scores
and bioenergy production, whereas seeking for further financial profits for each anaerobic digestion-coupled process in different countries
improvement in the MW sector might be undesirable for India. (China, India, USA and UK) were obtained and their future potentials
Notably, the weight coefficients of the evaluated indicators are ex- were predicted. The high-rate activated sludge plus sludge anaerobic
pected to change in the future with continued economic and social digestion process is identified as the most sustainable technology to be
development of each country. To provide guidance for future MW applied in China, USA, and UK at any time, while the up-flow anaerobic
management, the hypothetical performance of MW treatment processes sludge blanket plus activated sludge process is more preferable for India
in individual indicators and the overall MCE scores were estimated at the present stage. In the future, economics will still be key factor
based on the predicted data in 2025. The results show that, in China governing the selection of municipal wastewater treatment technology
and UK, the bioenergy production will have the highest weight coeffi- in China and India. To maximize the sustainability, it will be econom-
cients among other factors (Fig. 8a), implying a growing importance of ically more feasible for the developed countries to explore and imple-
bioenergy production in both countries. The relatively high weight ment the resource-oriented anaerobic technologies, while developing
coefficient of construction and operation costs in China and India in- countries would still lay a high priority on the development of low-cost
dicates that economic cost would still be a critical consideration for the and environmentally-beneficial technologies in the near future.
developing countries in selecting MW treatment technology. In terms of
integrated sustainability, the HRAS+SAD would outcompete the other Competing financial interests
two processes in the future for all the countries including India, due to
the increased weight of construction and operation costs and GHG The authors declare no competing financial interests.
emission (Fig. 8b). Relatively lower MCE scores of the UASB+AS pro-
cess were predicted for China (0.27) and UK (0.30), indicating this Acknowledgements
process was unsuitable for the two countries. The low score of the UASB
+AS was attributed mainly to its relatively high GHG emission and low This work was supported by the National Natural Science
bioenergy production, which is against the growing priority for China Foundation of China (grant numbers 51538011, 51522812 and
and UK to develop clean energy and reduce GHG emission for sus- 51821006).
tainable development [72,73]. Therefore, such a mainline AD process
should be applied with caution. Appendix A. Supplementary data

4.3. Future perspectives for practical application of AD-coupled processes Supplementary data to this article can be found online at https://
for energy and nutrient recovery doi.org/10.1016/j.rser.2019.04.064.

Although the HRAS+SAD process has the highest score in in- References
tegrated sustainability, its widespread application in the future still
faces many challenges. One key issue is that the nutrients in effluent of [1] US EPA. Water and energy: leveraging voluntary programs to save both water and
the HRAS+SAD process still need to be removed or recovered. Thus, it energy. 2008. Washington DC.
[2] Yan P, Qin R, Guo J, Yu Q, Li Q, Chen Y, et al. Net-zero-energy model for sustainable
has to be followed by carbon-independent nutrient removal/recovery wastewater treatment. Environ Sci Technol 2017;51:1017–23.
technologies, which are still immature today [71]. Another challenge is [3] He L, Du P, Chen Y, Lu H, Cheng X, Chang B, et al. Advances in microbial fuel cells
the economic affordability. The achievable ENPV by shifting from CAS for wastewater treatment. Renew Sustain Energy Rev 2017;71:388–403.
[4] McCarty PL, Bae J, Kim J. Municipal wastewater treatment as a net energy pro-
+SAD process to resource-oriented treatment paradigm varies sig- ducer-Can this be achieved? Environ Sci Technol 2011;45:7100–6.
nificantly for different countries. Specifically, USA and UK will have [5] Tauseef SM, Abbasi T, Abbasi SA. Energy recovery from wastewaters with high-rate
positive ENPV values, while China and India have negative values anaerobic digesters. Renew Sustain Energy Rev 2013;19:704–41.

140
B.-C. Huang, et al. Renewable and Sustainable Energy Reviews 110 (2019) 132–142

[6] Li WW, Yu HQ, Rittmann BE. Reuse water pollutants. Nature 2015;528:29–31. [39] International Energy Agency. CO2 emission from fuel combustion. IEA statistics;
[7] Jewell WJ. Anaerobic sewage treatment. Environ Sci Technol 1987;21(1):14–21. 2016.
[8] Seghezzo L, Zeeman G, van Lier JB, Hamelers HVM, Lettinga G. A review: the [40] Statistical BP. Review of world energy. https://www.bp.com/en/global/corporate/
anaerobic treatment of sewage in UASB and EGSB reactors. Bioresour Technol energy-economics/statistical-review-of-world-energy/electricity.html, Accessed
1998;65:175–90. date: November 2017.
[9] Hao X, Cai Z, Fu K, Zhao D. Distinguishing activity decay and cell death from [41] Henze M, Harremoes P, La Cour Jansen J, Arvin E. Wastewater treatment: biological
bacterial decay for two types of methanogens. Water Res 2012;46(4):1251–9. and chemical processes. New York: Springer Science & Business Media; 2001.
[10] Lettinga G, Rebac S, Zeeman G. Challenge of psychrophilic anaerobic wastewater [42] UN Environment Programme. The Water Supply & Sanitation Collaborative
treatment. Trends Biotechnol 2001;19(9):363–70. Council, the World Health Organization. Water pollution control: a guide to the use
[11] Chen Y, Cheng JJ, Creamer KS. Inhibition of anaerobic digestion process: a review. of water quality management principles. London: E & FN Spon; 1997.
Bioresour Technol 2008;99(10):4044–64. [43] World Bank. http://data.worldbank.org/about/country-and-lending-groups.
[12] Li WW, Yu HQ. Advances in energy-producing anaerobic biotechnologies for mu- [Accessed November 2017].
nicipal wastewater treatment. Engineering 2016;2:438–46. [44] UNFAO AQUASTAT database. http://www.fao.org/nr/water/aquastat/data.
[13] Sheboygan Regional Wastewater Treatment Facility. Treatment plant brochure. [Accessed November 2017].
2017. [Acessed Novembener, 2017]. http://www.sheboyganwwtp.com/files/2012/ [45] Florke M, Kynast E, Barlund I, Eisner S, Wimmer F, Alcamo J. Domestic and in-
pdf/Treatment%20Plant%20Brochure%202012.pdf. dustrial water uses of the past 60 years as a mirror of socio-economic development:
[14] Jonasson M. Energy Benchmark for wastewater treatment processes. Sweden: Lund a global simulation study. Glob Environ Chang 2013;23:144–56.
University; 2007. [46] Alcamo J, Doll P, Henrichs T, Kaspar F, Lehner B, Rosche T, et al. Development and
[15] Walia R, Kumar P, Mehrotra I. Performance of UASB based sewage treatment plant testing of the WaterGAP 2 global model of water use and availability. Hydrol Sci J
in India: polishing by diffusers an alternative. Water Sci Technol 2011;63:680–8. 2003;48:317–37.
[16] Monroy O, Famá G, Meraz M, Montoya L, Macarie H. Anaerobic digestion for [47] Kamalesh JM. An analysis of wastewater temperature variations in six remote
wastewater treatment in Mexico: state of the technology. Water Res monitored onsite systems. West Virginia University; 2008.
2000;34:1803–16. [48] Mohseni O, Stefa HG, Erickson TRA. Nonlinear regression model for weekly stream
[17] Zhang L. Anaerobic treatment of municipal wastewater in a UASB-digester system: temperatures. Water Resour Res 1998;34:2685–92.
temperature effect on system performance, hydrolysis and methanogenesis [dis- [49] Cipolla SS, Maglionico M. Heat recovery from urban wastewater: analysis of the
sertation]. Netherland: Wageningen University; 2016. variability of flow rate and temperature in the sewer of bologna. Italy. Energ.
[18] Smith AL, Stadler LB, Cao L, Love NG, Raskin L, Skerlos SJ. Navigating wastewater Buildings 2014;69:122–30.
energy recovery strategies: a life cycle comparison of anaerobic membrane bior- [50] Climatic Research Unit. http://www.cru.uea.ac.uk.[Accessed November 2017].
eactor and conventional treatment systems with anaerobic digestion. Environ Sci [51] Flores-Alsina X, Rodriguez-Roda I, Sin G, Gernaey KV. Multi-criteria evaluation of
Technol 2014;48:5972–81. wastewater treatment plant control strategies under uncertainty. Water Res
[19] Mo W, Zhang Q. Can municipal wastewater treatment systems be carbon neutral? J 2008;42:4485–97.
Environ Manag 2012;112:360–7. [52] Wang X, Liu J, Ren NQ, Yu HQ, Lee DJ, Guo X. Assessment of multiple sustainability
[20] Buonocore E, Franzese PP, Ulgiati S. Assessing the environmental performance and demands for wastewater treatment alternatives: a refined evaluation scheme and
sustainability of bioenergy production in Sweden: a life cycle assessment perspec- case study. Environ Sci Technol 2012;46:5542–9.
tive. Energy 2012;37:69–78. [53] Cornejo PK, Becker J, Pagilla K, Mo WW, Zhang Q, Mihelcic JR, et al. Sustainability
[21] Zhang C, Anadon LD. Life cycle water use of energy production and its environ- metrics for assessing water resource recovery facilities of the future. Water Environ
mental impacts in China. Environ Sci Technol 2013;47:14459–67. Res 2019;91:45–53.
[22] Foley JM, Rozenda RA, Hertle CK, Lant PA, Rabaey K. Life cycle assessment of high- [54] UN Population Division. World population prospects: the 2015 revision. 2015.
rate anaerobic treatment, microbial fuel cells, and microbial electrolysis cells. p. 40.
Environ Sci Technol 2010;44:3629–37. [55] US Department of Agriculture Economic Research Service. International macro-
[23] Huang IB, Keisler J, Igor L. Multi-criteria decision analysis in environmental sci- economic data set, www.ers.usda.gov/data-products/international-
ences: ten years of applications and trends. Sci Total Environ macroeconomic-data-set. 41[Accessed February 2017].
2011;409(19):3578–94. [56] US Energy Information Administration. International energy statistics: renewables
[24] Murrant D, Radcliffe J. Assessing energy storage technology options using a multi- (electricity generation). https://www.eia.gov/beta/international/data/[Accessed
criteria decision analysis-based framework. Appl Energy 2018;231:788–802. November 2017].
[25] Caldeira C, Freire F, Olivetti EA, Kirchain R, Dias LC. Analysis of cost-environmental [57] World Bank. World development indicators: GDP per capita. http://databank.
trade-offs in biodiesel production incorporating waste feedstocks: a multi-objective worldbank.org/data/reports.aspx?source=world-development-indicators,
programming approach. J Clean Prod 2019;216:64–73. Accessed date: November 2017.
[26] Jimenez J, Miller M, Bott C, Murthy S, De Clippeleir H, Wett B. High-rate activated [58] World Bank. World development indicators: population density. http://databank.
sludge system for carbon management-Evaluation of crucial process mechanisms worldbank.org/data/reports.aspx?source=world-development-indicators
and design parameters. Water Res 2015;87:476–82. [Accessed November 2017].
[27] Henze M, van Loosdrecht MC, Ekama GA, Brdjanovic D. Biological wastewater [59] World Bank. Worldwide governance indicators: government effectiveness. http://
treatment. London: IWA Publishing; 2008. databank.worldbank.org/data/reports.aspx?source=worldwide-governance-
[28] Bowen EJ, Dolfing J, Davenport RJ, Read FL, Curtis TP. Low-temperature limitation indicators [Accessed November 2017].
of bioreactor sludge in anaerobic treatment of municipal wastewater. Water Sci [60] Organisation for Economic Co-operation and Development. OECD statistics. stats.
Technol 2014;69:1004–13. oecd.org [Accessed November 2017].
[29] US EPA. Opportunities for combined heat and power at wastewater treatment fa- [61] Kumar S, Smith SR, Fowler G, Velis C, Kumar SJ, Arya S, et al. Challenges and
cilities: market analysis and lessons from the field. 2011. Washington DC. opportunities associated with waste management in India. Open Science
[30] Metcalf and Eddy. Wastewater engineering: treatment and reuse. New York: 2017;4:160764.
McRAW-Hill publishing; 2003. [62] US Office of Management and Budget. Circular A-94 revised: guidelines and dis-
[31] Judd S, Judd C. The MBR book: chapter 3-design. Oxford: Elsevier Science; 2006. count rates for benefit-cost analysis of federal programs. 1994https://www.
[32] Bandaraa WM, Satoha H, Sasakawab M, Nakaharab Y, Takahashia M, Okabe S. whitehouse.gov/omb/circulars_a094, Accessed date: June 2017.
Removal of residual dissolved methane gas in an upflow anaerobic sludge blanket [63] Hernández-Sancho F, Molinos-Senante M, Sala-Garrido R. Economic valuation of
reactor treating low-strength wastewater at low temperature with degassing environmental benefits from wastewater treatment processes: an empirical ap-
membrane. Water Res 2011;45:3533–40. proach for Spain. Sci Total Environ 2010;408:953–7.
[33] Cookney J, Cartmell E, Jefferson B, McAdam EJ. Recovery of methane from anae- [64] Molinos-Senante M, Hernández-Sancho F, Sala-Garrido R. Economic feasibility
robic process effluent using poly-di-methyl-siloxane membrane contactors. Water study for wastewater treatment: a cost–benefit analysis. Sci Total Environ
Sci. Tehcnol. 2012;65:604–10. 2010;408:4396–402.
[34] Luo G, Wang W, Angelidaki IA. New degassing membrane coupled upflow anae- [65] Eurostat database. http://ec.europa.eu/eurostat/data/database [Accessed June
robic sludge blanket (UASB) reactor to achieve in-situ biogas upgrading and re- 2017].
covery of dissolved CH4 from the anaerobic effluent. Appl Energy [66] World Bank Group. Paying taxes. http://www.doingbusiness.org/reports/thematic-
2014;132:536–42. reports/paying-taxes/[Accessed June 2017].
[35] Lelieveld J, Lechtenböhmer S, Assonov SS, Brenninkmeijer CAM, Dienst C, [67] International Labor Organization. Labor costs, http://www.ilo.org/global/
Fischedick M, et al. Greenhouse gases: low methane leakage from gas pipelines. statistics-and-databases/lang-en/index.htm [Accessed June 2017].
Nature 2005;434:841–2. [68] World Bank Group Finances. Loan average interest by country. https://finances.
[36] Shahabadi MB, Yerushalmi L, Haghighat L. Impact of process design on greenhouse worldbank.org/Loans-and-Credits/WB-Loan-Average-Interest-Rate-by-Country/
gas (GHG) generation by wastewater treatment plants. Water Res cfc3-hhwe/data [Accessed June 2017].
2009;43:2679–87. [69] World Bank. Global economic monitor commodities. http://databank.worldbank.
[37] Kampschreur MJ, Temmink H, Kleerebezem R, Jetten MSM, Van Loosdrecht MCM. org/data/reports.aspx?source=global-economic-monitor-commodities. [Accessed
Nitrous oxide emission during wastewater treatment. Water Res June 2017].
2009;43:4093–103. [70] US EPA. Emerging technologies for wastewater treatment and in-plant wet weather
[38] Myhre G, Shindell D, Bréon FM, Collins W, Fuglestvedt J, Huang J, et al. management. 2013.
Anthropogenic and natural radiative forcing. Climate change 2013: the physical [71] IEA. Atalas of energy. http://energyatlas.iea.org/[Accessed November 2017].
science basis. Contribution of working group i to the fifth assessment report of the [72] National Development and Reform Commission. The thirteen five-year plan of re-
intergovernmental panel on climate change. Cambridge, United Kingdom and New newable energy development. www.ndrc.gov.cn/zcfb/zcfbtz/201612/
York, NY, USA: Cambridge University Press; 2013. W020161216659579206185.pdf [in Chiense, accessed Novembe 2017].

141
B.-C. Huang, et al. Renewable and Sustainable Energy Reviews 110 (2019) 132–142

[73] UNFCCC. China's intended nationally determined contribution: enhanced actions treatment systems: techno-economical approach. Water Sci Technol
on climate change. 2015. 2015;71:468–78.
[74] Connor R, Renata A, Ortigara C, Koncagül E, Uhlenbrook S, Lamizana-Diallo BM, [77] Feng Y, Zhang Y, Quan X, Chen S. Enhanced anaerobic digestion of waste activated
et al. The United Nations world water development report. Wastewater: the un- sludge digestion by the addition of zero valent iron. Water Res 2014;52:242–50.
tapped resource. 2017. [78] Mao C, Feng Y, Wang X, Ren G. Review on research achievements of biogas from
[75] Larsen TA, Hoffmann S, Lüthi C, Truffer B, Maurer M. Emerging solutions to the anaerobic digestion. Renew Sustain Energy Rev 2015;45:540–55.
water challenges of an urbanizing world. Science 2016;352:928–33. [79] Li WW, Yu HQ, He Z. Towards sustainable wastewater treatment by using microbial
[76] Singh NK, Kazmi AA, Starkl MA. Review on full-scale decentralized wastewater fuel cells-centered technologies. Energy Environ Sci 2014;7:911–24.

142

You might also like