Journal of Cleaner Production 170 (2018) 1185e1194

Contents lists available at ScienceDirect

Journal of Cleaner Production

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

Technical, economic and environmental assessment of coagulation/

filtration tertiary treatment processes in full-scale wastewater
treatment plants
Dan Wang a, Fang Guo b, Yihui Wu b, Zhiping Li b, Guangxue Wu a, *
a
Key Laboratory of Microorganism Application and Risk Control (MARC) of Shenzhen, Graduate School at Shenzhen, Tsinghua University, Shenzhen 518055,
China
b
Kunming Dianchi Water Treatment Co. Ltd., Kunming 650228, China

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

Article history: The technical performance, economic cost and environmental impact of six full-scale tertiary coagula-
Received 2 June 2017 tion/filtration processes located in Kunming, China were evaluated. All tertiary treatment processes
Received in revised form removed total phosphorus (TP) and total suspended solids (TSS) efficiently, with the removal percentages
23 September 2017
of 55.0%e80.0% and 50.0%e74.0%, respectively. Polyaluminium chloride (PAC) consumption for TP and
Accepted 26 September 2017
Available online 26 September 2017
TSS removal in the six tertiary treatment processes were quite different, with chemical dosages of 7.9
e38.5 g PAC/g TP removed and 0.3e1.7 g PAC/g TSS removed, respectively. The multiple linear regression
analysis showed that the PAC dosage closed to the optimal value benefited TP and TSS removal, and this
Keywords:
Coagulation/filtration tertiary treatment
also reduced the economic cost. For environmental impacts, the main source of greenhouse gas was
process electricity consumption and the coagulation/filtration process had positive effect on reducing eutro-
Technical performance phication. The comprehensive assessment including technical, economic and environmental aspects was
Economic cost characterized by the composite cost index. The composite cost index showed that the tertiary treatment
Greenhouse gas emission process of micro-flocculation with D type/cloth media filtration achieved the best comprehensive per-
Eutrophication formance, while D type filter had great potential for energy saving and chemical reduction.
© 2017 Elsevier Ltd. All rights reserved.

1. Introduction
Wastewater reuse is an effective way to alleviate the shortage of
water resources. As an important approach to protect water envi-
ronment, WWTPs (wastewater treatment plants) play an important
role in wastewater purification and reclamation. The final effluent
of WWTPs is often reused as scenic environment water, industrial
water, agricultural irrigation water, etc. The secondary effluent of
WWTP usually cannot meet the discharge standard and re-
quirements for reclaimed water reuse. Therefore, it is necessary to
adopt tertiary treatment for secondary effluent. The tertiary treat-
ment processes mainly include coagulation, sedimentation, bio-
logical filter, physical/membrane filtration, ozonation, disinfection,
~o et al., 2013; Friedler et al.,
etc. (Plakas et al., 2016; Norton-Branda
2008; Ebeling et al., 2003). Coagulation/filtration processes are
widely applied in tertiary treatment because of its simplicity, high
* Corresponding author.
E-mail address: wu.guangxue@sz.tsinghua.edu.cn (G. Wu).
removal efficiency and low cost. To the best of our knowledge, there
are lots of studies on jar tests and pilot scale tests related to
coagulation and flocculation (Illueca-Mun ~ oz et al., 2008; Guida
et al., 2007). However, the full-scale coagulation/filtration pro-
cesses have been relatively less investigated. The performance
evaluation of full-scale coagulation/filtration process would be
beneficial to optimize the operation of the tertiary treatment pro-
cess and to provide references for the selection of sound tertiary
treatment processes.
At present, the technical performance evaluation of WWTPs or
wastewater treatment processes includes qualitative and quanti-
tative evaluation. Qualitative indicators usually refer to the reli-
ability, simplicity and stability of the wastewater treatment
systems. The removal efficiency, removal loading, effluent con-
centration distribution, etc., are denoted as quantitative technical
performance indicators (Quadros et al., 2010; Bott et al., 2012).
Nevertheless, the main drawbacks of the above mentioned tech-
nical performance indicators are that the improvement measures
and suggestions to optimize the operating performance cannot be
provided. To identify the key factors affecting pollutant removal

https://doi.org/10.1016/j.jclepro.2017.09.231
0959-6526/© 2017 Elsevier Ltd. All rights reserved.
1186 D. Wang et al. / Journal of Cleaner Production 170 (2018) 1185e1194
Abbreviations
CO2 Carbon dioxide
COD Chemical oxygen demand
DPF Discharge pollution fee
GHG Greenhouse gas
HRT Hydraulic retention time (min)
LCA Life cycle assessment
NH4-N Ammonia-nitrogen
NO3-N Nitrate-nitrogen
PAC Polyaluminium chloride
PACTP Ratio of PAC consumption to the removed total
phosphorus (g PAC/g TP removed)
PACTSS Ratio of PAC consumption to the removed total
suspended solids (g PAC/g TSS removed); PAM
Polyacrylamide
PO4-P Phosphate
will benefit to achieve maximum removal of pollutants, so as to
improve operating performance of wastewater treatment systems.
Multiple linear regression has been widely applied to explore the
influence factors of dependent variable, which is the mathematical
statistical method for quantitative study of the functional rela-
tionship between a dependent variable and multiple independent
variables (Hijosa-Valsero et al., 2011). However, for the full-scale
coagulation/filtration processes, the application of multiple
regression method to explore factors affecting the removal of pol-
lutants is relatively less.
The performance of wastewater treatment process has a great
relation with the economic cost. For example, Remy et al. (2014)
evaluated the environmental impact of the five tertiary treatment
processes by the life cycle assessment (LCA) method, and found
that although the effluent of the membrane bioreactor was the best,
the costs of electricity and chemical consumption were quite high.
Therefore, it is necessary to evaluate the tertiary treatment pro-
cesses combined with technical performance and economic cost.
The main purpose of the previous WWTPs was to achieve the
effluent quality standards so as to protect the receiving water
bodies (Mannina et al., 2016). Although the tertiary treatment
process greatly improves the effluent quality, it would have a direct
or indirect impact on the environment due to the consumption of
energy, chemicals and other materials (Rahman et al., 2016).
Nowadays, the environmental impact caused by the wastewater
treatment process has been received more attention. As an effective
tool for environmental impact evaluation, LCA has been widely
applied in wastewater treatment processes. Various environmental
impact indicators are involved in LCA, such as greenhouse gas
(GHG), eutrophication, acidification, photochemical oxidation,
toxicity-related aspects, ozone layer depletion and abiotic resource
depletion, etc. (Corominas et al., 2013). However, GHG and effluent
eutrophication have been received more attention compared with
other impact categories, because of their most significant contri-
bution to the environmental pollution (Li et al., 2013; Corominas
et al., 2013; Lorenzo-Toja et al., 2016). GHG or eutrophication has
been considered as the only environmental indicator in many
studies related to environmental impact assessment of WWTPs,
while other environmental factors have been neglected. For
instance, Rodriguez-Garcia et al. (2011) mainly focused on the
effluent eutrophication of WWTPs, while global warming was only
considered in the study of Dong et al. (2017). For the environmental
impact evaluation of tertiary treatment processes in this study,
RTP Removal efficiency of TP (%)
RTSS Removal efficiency of TSS (%)
SPSS Statistical Package for the Social Science
T Temperature (
C)
TN Total nitrogen
TP Total phosphorus
TPin Influent total phosphorus of tertiary treatment process
(mg/L)
TPS Technology performance statistics
TSS Total suspended solids
TSSin Influent total suspended solids of tertiary treatment
process (mg/L)
VD Filtration rate of D type filter (m/h)
VIF Variance inflation factor
VV Filtration rate of V type filter (m/h)
WWTP Wastewater treatment plant
GHG and eutrophication are mainly concerned.
The limitation would exist if technical performance, economic
cost and environmental impact are evaluated separately, which
cannot give an overall and comprehensive evaluation results and is
also unfavorable to make a comparative analysis for multiple
evaluation objects (Molinos-Senante et al., 2014; Plakas et al.,
2016). Therefore, a single index system needs to be integrated
into a comprehensive index contained multi-dimensional infor-
mation. Many methods can be used to integrate several indicators
into a composite indicator (Nardo et al., 2008). The comprehensive
evaluation methods used commonly include analytic hierarchy
process (Molinos-Senante et al., 2014), gray relational analysis
(Zeng et al., 2007), data envelopment analysis (DEA) (Molinos-
Senante et al., 2016) and the combination of the above analytical
methods (Pophali et al., 2011). Majority methods except the DEA
method need to be assigned weights, leading to the strong
subjectivity. But the defect of the DEA method is that it is not
suitable for the evaluation of fewer decision units or objects. In
addition, some other comprehensive evaluation methods are also
proposed. For example, Molinos-Senante et al. (2014) proposed the
weighted average value of economy, environment and society to
evaluate the sustainability of the seven secondary treatment pro-
cesses. Based on the combination of environmental and economic
indicators, Lorenzo-Toja et al. (2016) segregated 22 WWTPs into A,
B and C three different categories, in which the A and C rating
represented the best and the worst eco-efficiency standard,
respectively.
According to the monetization approach, composite cost was
taken as the comprehensive index to evaluate tertiary treatment
processes in the present study. The monetization approach is based
on the idea that the seriousness of different categories or indicators
can be measured by money (Wu et al., 2005). Namely, the influence
of different pollutants emission on environment can be quantified
by money. “Green tax” levied on pollutant emissions denotes the
social willingness to pay for the environmental pollution (Wu et al.,
2005). “Green tax” is widely used in environmental impact
assessment. To promote the environmental friendly earthwork
construction, Li et al. (2010) identified the most important envi-
ronmental impact indicators during the construction process by the
“green tax” method. Li et al. (2013) also evaluated the environ-
mental costs of twelve environmental impact indicators of a WWTP
by the “green tax” method, indicating that eutrophication and
global warming are two most expensive environmental impacts.
 D. Wang et al. / Journal of Cleaner Production 170 (2018) 1185e1194
The purpose of this study was to evaluate the technical perfor-
mance, economic cost and two most important environmental
impacts of GHG and eutrophication in six full-scale coagulation/
filtration processes. The multiple regression analysis was adopted
to clarify factors affecting the removal efficiency of key pollutants,
so as to provide theoretical references for operating and optimizing
the coagulation/filtration process.
2. Materials and methods
2.1. Coagulation/filtration processes
The tertiary treatment processes of six WWTPs are located in
Kunming, China. The tertiary effluent is discharged into the river or
lake, applying for scenic environment application. As the effluent of
six secondary treatment processes could not meet the effluent
discharge standards, especially for TSS and TP, so the tertiary
treatment process of coagulation/filtration was adopted to polish
the secondary effluent.
The flowchart and system boundary of tertiary treatment sys-
tem is shown in Fig. 1. The detailed information about six tertiary
treatment processes are presented in Table 1. The solid line in Fig. 1
shows the complete tertiary treatment processes including coag-
ulation, sedimentation, filtration, ultraviolet (UV) disinfection and
chemical sludge disposal. The system boundary applied in this
study is shown as the dashed line. The removal of pathogenic
bacteria was not examined in the present study and UV disinfection
had no effect on the removal of pollutants such as chemical oxygen
demand (COD), total nitrogen (TN), TP and TSS, thus, the UV
disinfection tank was excluded from the coagulation/filtration
system boundary. There was no chemical sludge data to support the
study in this paper, so chemical sludge disposal was not considered.
The Actiflo of WWTPC was quite different from other coagula-
tion facilities, which was integrated with coagulation (PAC addi-
tion), ballasted flocculation and settling. The ballasted flocculation
possesses a high settlement rate with the addition of microsand
and polyacrylamide (PAM), which is one of the main feature of
Actiflo. Micro-flocculation is the process of increasing the size of
particles as a result of particle collisions, and the particle aggre-
gation is brought about by the random thermal motion of fluid
molecules. There were only two types of filters for six tertiary
Fig. 1. The flowchart and system boundary of six tertiary treatment processes.
1187
treatment processes, namely V type filter and D type filter. The V
type filter is a descending current open sand filter. It retains sus-
pended solids in raw water owing to a thick layer of sand. D type
filter was developed on the basis of V type filter. The main differ-
ence between the V type filter and D type filter was the filter media,
with the filter media of sand and fiber, respectively. The D type filter
has the advantages of less land occupation, less investment cost
and more convenient maintenance than that of V type filter,
because of the high filtration rate, long service life and easy
replacement of fiber media. D type and V type filters were equipped
with three types of backwashing, i.e., air backwashing, water
backwashing and air-water combined backwashing.
2.2. Functional unit
The functional unit is a measure of the function of the studied
system and it is a standardized parameter used to describe the final
evaluation result. The most widely used functional unit for the
evaluation of WWTP is the volume of treated wastewater, such as
1 m3 treated wastewater (Corominas et al., 2013). However, the
functional unit based on the volume of treated wastewater is not
always representative, because it cannot reflect the influent quality
or removal efficiency of a WWTP (Corominas et al., 2013).
In order to reflect the removal efficiency of six tertiary treatment
processes, per ton equivalent of removed COD (1 t CODeq removed)
was used as the functional unit in this study. According to Copp
et al. (2002) and Benedetti et al. (2008), the removed TSS, TP and
TN were converted into the COD equivalent by the discharge
pollution fee (DPF). The DPF was derived from the pollutant
discharge costs of legislation in Flanders, Belgium. The discharge of
pollutants could be charged with a corresponding fine. Similarly,
the reduction of pollutant emissions (removal of pollutants) will
reduce the corresponding fines. According to the proportion of
fines for different pollutants, the COD equivalents of TSS, TN and TP
were 2 kg CODeq/kg TSS, 20 kg CODeq/kg TN and 100 kg CODeq/kg
TP, respectively (Benedetti et al., 2008).
2.3. Technical performance evaluation
2.3.1. Removal efficiency and technology performance statistic
Based on the pollutant indicators measured weekly of six
1188 D. Wang et al. / Journal of Cleaner Production 170 (2018) 1185e1194

Table 1
The characteristics of six full-scale coagulation/filtration tertiary treatment processes.

unit WWTPA WWTPB WWTPC WWTPD WWTPE WWTPF

Designed capacity 104 m3/ 12 10 21 18.5 13 30

d
Coagulation tank e e Micro- Actiflob Coagulation Folded plate flocculation Micro-flocculation tank
flocculationa tank tank
Sedimentation tank e e e Inclined plate setting e Inclined tube setting e
Filter e D type filter V type filter D type filter V type filter D type filter D type/Cloth media
filterc
Chemical (material) e PAC PAC PAC, PAM, microsand PAC PAC PAC
HRT of coagulation tank min e e 10.4 2.4 16.0 5.8
Filtration rate m/h 14.88 4.80 17.36 8.45 14.11 12.86
Air backwash intensity L/(m2$s) 23 18 23 55 23 23
Water backwash intensity L/(m2$s) 6 5 6 17 6 6
Air backwash time min 3 4 2 5 5 5
Air and water backwash min 5 6 3 6 15 15
time
Water backwash time min 3 5 2 5 5 5
Discharging e Dianchi Dianchi lake Dianchi lake, Daguan Dianchi lake Dianchi lake Dianchi lake
lake river
a
There was no coagulation tank before the V type filter of WWTPB, but micro-flocculation function still existed. A long distant between the secondary clarifier and V type
filter of WWTPB provided the sufficient reaction time for the chemical and wastewater.
b
The coagulation tank of Actiflo refers to the coagulation and ballasted flocculation.
c
The designed capacity of WWTPF was 30*104 m3/d, 20*104 m3/d filtered by D type filter, another 10*104 m3/d filtered by cloth media filter. It was impossible to distinguish
the effluent between D type filter and cloth media filter of WWTPF, because the effluent of these two filters were mixed together as the final effluent. In this study, the function
of cloth media filter was regarded as the function of D type filter.

coagulation/filtration processes in 2014, the influent, effluent and
removal efficiency of TP, TSS, COD, TN and ammonia-nitrogen (NH4-
N) were used to evaluate the technical performance of the system.
Technology performance statistics (TPS) (Bott et al., 2012) was also
used to evaluate the performance of tertiary treatment processes
by analyzing the effluent pollutant concentrations.
TPS was a probability statistical method for the calculation of
effluent pollutant concentration by the PERCENTILE function in
Excel. The three levels of 3.84th percentile (14/365 ¼ 3.84%), 50th
percentile (TPS-50%) and 95th percentile (TPS-95%) were evaluated
to represent the ideal, the median, and the reliably achievable
performance, respectively (Bott et al., 2012). TPS-3.84%, TPS-50%
and TPS-95% refer to the monitored data of 3.84%, 50% and 95% after
ranking from the minimum data to the maximum data. The
detailed explanation of TPS was from Bott et al. (2012). In a word,
the effluent concentration of different tertiary treatment processes
was compared by technical performance statistics.
2.3.2. Factors affecting the removal of TP and TSS
The coagulation/filtration tertiary treatment process mainly
removes pollutants such as TSS and TP. WWTPC Actiflo process
needed three types of materials, namely PAC, microsand and PAM,
among which the consumption of PAC accounted for 98%. The
chemical applied for other five coagulation/filtration processes was
only PAC. Therefore, PAC was the main chemical for six tertiary
treatment processes. To clarify the main factors affecting the
removal of TP and TSS, the stepwise multiple linear regression
analysis was adopted, which was realized by the statistical software
SPSS (Statistical Package for the Social Science, version 19.0).
Stepwise multiple linear regression is a mathematical statistical
method to quantify the functional relationship between a depen-
dent variable and multiple independent variables, and it also
benefits to clarify the independent variables which significantly
affect a dependent variable. For the stepwise multiple linear
regression establishment, the independent variables were entered
into the regression model one by one, and the significance of in-
dependent variables that were already in the regression model was
tested after each new independent variable to be included in
regression model. If the independent variables were tested to be
non-significant, they were excluded from the model. Selection and
exclusion steps were repeated until there was no significant inde-
pendent variable to be removed or entered, then the final linear
regression equation was obtained (Zhang et al., 2016). The inde-
pendent variables added or removed from the regression model
were generally based on the significance test of a ¼ 0.05 (95%
confidence level). The dependent variables were TP and TSS
removal efficiency (RTP, RTSS) in the present study. The independent
variables included temperature (T), pH, filtration rate of D/V type
filter (VD, VV), hydraulic retention time (HRT) of coagulation tank,
influent concentration of TP/TSS (TPin, TSSin) and the consumption
of PAC (PACTSS, PACTP).
2.4. Economic cost evaluation
The economic cost of a WWTP generally includes investment,
maintenance and operation costs. The investment costs can be
subdivided into construction costs, machinery and equipment
costs, land costs, etc. Operation and maintenance costs of waste-
water treatment process include energy consumption costs,
chemical consumption costs, sludge disposal costs and personnel
salaries. For the present study, the investment cost during con-
struction phase was not considered, mainly considering the oper-
ation cost during the operation stage, due to their relationship with
the operation of tertiary treatment processes. Costs of chemical and
electricity were mainly considered as the operation cost of tertiary
treatment processes in this study. The sludge disposal and
manpower expenditure were excluded because of lacking relevant
practical data. The unit of operation cost is USD/t CODeq removed.
Since the statistical data came from the year of 2014, the CNY was
converted into USD by the annual average exchange rate (1
USD ¼ 6.143 CNY) in 2014 (National Bureau of Statistics of China
(NBSC), 2015). The actual unit prices of electricity, PAC, PAM and
microsand were 0.106 USD/kWh, 0.109 USD/kg, 3.419 USD/kg, and
0.342 USD/kg, respectively.
2.5. Environmental impact evaluation
2.5.1. Greenhouse gas emission
Electricity and chemical were two main sources of GHG emis-
sions for tertiary treatment processes. GHG emissions are
 D. Wang et al. / Journal of Cleaner Production 170 (2018) 1185e1194
commonly expressed as carbon dioxide (CO2) equivalent. The GHG
emissions generated by electricity consumption was calculated as
follows:

.
GHGelectricity ¼ EFgrid  Celectricity DCODeq
where, GHGelectricity is the GHG emission generated from electricity
(kg CO2eq/t CODeq removed). EFgrid is average power grid emission
factor in southwest of China, with the value of 0.66 kg CO2eq/kWh
(Song et al., 2013); Celectricity is electricity consumption per year
(kWh/y); DCODeq is the removed COD equivalent (t CODeq
removed/y).
The GHG produced by chemical mainly came from PAC. Apart
from WWTPC, the tertiary chemical consumption of all WWTPs
was PAC. Although there were three types of chemicals for the
Actiflo process, PAC was considered as the main source of GHG due
to its largest consumption (98%). The GHG generated from micro-
sand and PAM were ignorable, because of small consumption (2%).
The formula of GHG emission from PAC consumption was as
follows:
GHGchemical ¼ ðfchemical  Wchemical Þ=DCODeq
where, GHGchemical is the GHG emission generated from PAC (kg
CO2eq/t CODeq removed); fchemical is PAC emission factor, with the
value of 1.182 kg CO2eq/kg PAC (de Haas et al., 2008); Wchemical is
PAC consumption per year (kg/y).
2.5.2. Effluent eutrophication
The environmental impact of eutrophication associated with
tertiary effluent of WWTPs was evaluated. Different pollutants that
could potentially cause eutrophication were qualified according to
the environmental impact assessment method (Guine e et al.,
2002). Eutrophication potential (EP) was used to measure the de-
gree of eutrophication caused by the discharge of 1 kg of pollutant.
The EP of different pollutants varies, with the EP of TP, NH4-N,
Nitrate-nitrogen (NO3-N) and COD of 3.06 kg PO4-Peq/kg TP,
0.33 kg PO4-Peq/kg NH4-N, 0.1 kg PO4-Peq/kg NO3-N and 0.022 kg
PO4-Peq/kg COD, respectively. The formula of eutrophication
caused by different pollutants was as follows:
Eutrophicationi ¼ ðEFi  Mi Þ=DCODeq
where, Eutrophicationi is the eutrophication caused by the tertiary
effluent of pollutant i (kg PO4-Peq/t CODeq removed); EPi is the
eutrophication potential of pollutant i (kg PO4-Peq/kg pollutant i);
Mi is the emission of pollutant i (kg pollutant i/y); The concentra-
tion of NO3-N was obtained by the subtraction between TN and
NH4-N, due to that NO3-N was not monitored in the tertiary
effluent and the nitrite nitrogen was negligible.
2.6. Comprehensive evaluation
Combined with the green tax (Wu et al., 2005), environmental
impact from GHG and eutrophication could be converted into
environmental costs. According to Wu et al. (2005), the environ-
mental impact costs of eutrophication and GHG were 0.094 USD/kg
NO3-Neq and 0.036 USD/kg Ceq, respectively. After the conversion
of unit, the environmental costs of eutrophication and GHG were
0.944 USD/kg PO4-Peq and 0.010 USD/kg CO2eq, respectively.
The comprehensive evaluation result was characterized by the
composite cost index. The calculated formula of composite cost
index was below:
1189
Composite cos t index ¼ ðOPEX þ GHGtax þ Eutrophicationtax Þ
 =DCODeq
where, composite cost index is used to evaluate different tertiary
treatment processes comprehensively (USD/t CODeq removed);
OPEX is the operational cost per year (USD/y); GHGtax and Eutro-
phicationtax are the green tax levied on the emission of GHG and
eutrophication (USD/y), respectively.
3. Results and discussion
3.1. Technical performance evaluation
3.1.1. Removal efficiency and technology performance statistic
The influent, effluent and removal efficiency of six tertiary
treatment processes are given in Table 2. There was no significant
difference for the effluent COD concentration and COD fluctuation
in six tertiary treatment processes, and the COD removal percent-
ages were in the range of 24.5%e35.2%. The TSS removal percent-
ages of six different coagulation/filtration processes ranged from
50.0% to 74.0%. The TN removal percentages of six tertiary treat-
ment processes were only 3.8%e19.2%. All coagulation/filtration
processes removed TP efficiently, with the removal percentages of
55.0%e80.0%. Six coagulation/filtration processes had the best
removal efficiency for TP and TSS, with TP and TSS removal per-
centages all above 50%. Üstün et al. (2011) and Friedler et al. (2008)
also demonstrated that the main purpose of coagulation/floccula-
tion tertiary treatment processes was to remove TP, TSS and other
suspended pollutants. Thus, it is essential to examine factors
affecting TP and TSS removal of six coagulation/filtration processes.
The TPS analysis of TP in six tertiary treatment processes is
shown in Fig. 2(a). The tertiary effluent TP concentration was be-
tween 0.03 and 0.4 mg/L (Fig. A.1), reaching the Class 1A level
(0.5 mg/L) of the national discharge standard of China (GB18918-
2002) (MOEP, 2002). Compared with the effluent TP concentration
of different tertiary processes, WWTPF achieved the worst effluent
quality because of the highest value of TPS-3.84%, TPS-50% and TPS-
95%, followed by WWTPC Actiflo þ D type filter. The ideal (TPS-
3.84%) values of the effluent TP concentration in six tertiary treat-
ment processes were less than 0.1 mg/L, while the median (TPS-
50%) and the reliably achievable (TPS-95%) values were all less than
0.2 mg/L and 0.35 mg/L, respectively. Sun et al. (2016) analyzed the
TP effluent quality of 3115 WWTPs in China by the QUARTILE
function. The lower quartile, the median, the upper quartile in the
QUARTILE function are equal to the TPS-25%, TPS-50% and TPS-75%
in the PERCENTILE function. The statistical analysis results from
3115 WWTPs in China showed that the lower quartile (25%), the
median (50%) and the upper quartile (75%) of the effluent TP con-
centrations were 0.3 mg/L, 0.5 mg/L and 0.8 mg/L, respectively (Sun
et al., 2016). The median of TP (0.5 mg/L) in 3115 WWTPs was
higher than that of TPS-95% (0.35 mg/L) in this study, indicating
that the TP effluent quality in six tertiary treatment processes was
much better than that of the national average level.
The TPS analysis of TSS in six tertiary treatment processes is
presented in Fig. 2(b). The tertiary effluent TSS concentration
ranged from 4 mg/L to 10 mg/L (Fig. A.2). The minimum value of TSS
was 4 mg/L due to the detection limit of 4 mg/L. The TPS results of
TSS demonstrated that the ideal (TPS-3.84%) and median (TPS-50%)
values of the effluent TSS were both 4 mg/L, indicating that at least
half of the samples of the effluent TSS were tested with the same
detection result (4 mg/L). The reliably achievable (TPS-95%) values
of effluent TSS in six tertiary treatment processes were all less than
or equal to 9 mg/L, reaching the Class 1A level (10 mg/L) of the
1190 D. Wang et al. / Journal of Cleaner Production 170 (2018) 1185e1194

Table 2
The technical performance of six tertiary treatment processes.

WWTP WWTPA WWTPB WWTPC WWTPD WWTPE WWTPF

COD Influent (mg/L) 18.2 ± 4.4 18.5 ± 4.5 21.9 ± 7.0 15.5 ± 4.0 19.2 ± 5.3 16.5 ± 3.6
Effluent (mg/L) 13.2 ± 2.3 12.3 ± 2.1 14.2 ± 3.0 11.7 ± 2.3 13.9 ± 2.9 11 ± 1.4
Removal efficiency 27.5% 33.5% 35.2% 24.5% 27.6% 33.3%
TSS Influent (mg/L) 12.3 ± 5.3 17 ± 10.4 13.3 ± 9.4 9.8 ± 3.1 11.0 ± 4.6 17.7 ± 8.7
Effluent (mg/L) 5.7 ± 2.7 5.0 ± 2.0 6.0 ± 5.8 4.9 ± 1.7 5.3 ± 2.0 4.6 ± 1.2
Removal efficiency 53.7% 70.6% 54.9% 50.0% 51.8% 74.0%
TN Influent (mg/L) 10.5 ± 4.8 11.3 ± 2.0 17.2 ± 4.5 9.5 ± 2.1 11.3 ± 3.2 14.1 ± 2.4
Effluent (mg/L) 10.1 ± 3.4 10.6 ± 1.9 13.9 ± 2.3 8.7 ± 2.1 10.0 ± 2.7 11.6 ± 2.3
Removal efficiency 3.8% 6.2% 19.2% 8.4% 11.5% 17.7%
TP Influent (mg/L) 0.2 ± 0.1 0.5 ± 0.2 0.6 ± 0.2 0.3 ± 0.2 0.3 ± 0.3 0.7 ± 0.2
Effluent (mg/L) 0.09 ± 0.07 0.10 ± 0.05 0.15 ± 0.10 0.13 ± 0.12 0.13 ± 0.09 0.22 ± 0.09
Removal efficiency 55.0% 80.0% 75.0% 56.7% 56.7% 68.6%
NH4-N Influent (mg/L) 0.9 ± 1.6 0.3 ± 0.4 2.4 ± 2.7 0.9 ± 1.5 0.5 ± 1.2 1.3 ± 1
Effluent (mg/L) 0.6 ± 1.2 0.2 ± 0.2 1.6 ± 1.5 0.6 ± 1.1 0.46 ± 1 0.5 ± 0.7
Removal efficiency 33.3% 33.3% 33.3% 33.3% 8.0% 61.5%

Fig. 2. The technology performance statistics (TPS) of tertiary effluent TP (a) and TSS (b) concentration.

national discharge standard (GB18918-2002) (MOEP, 2002). The
effluent TSS concentration of WWTPC Actiflo þ D type filter was the
lowest, with the minimum TPS-95% value of 6.6 mg/L. While the
tertiary treatment process of WWTPA achieved the worst effluent
TSS quality, with the maximum TPS-95% value of 9 mg/L. Sun et al.
(2016) showed that the lower quartile (25%) of TSS in 3194 WWTPs
was 7.4 mg/L, which was higher than that of TPS-50% (4 mg/L) in
this study. The median was 10.4 mg/L, which was higher than TPS-
95% of 9 mg/L. Thus, the tertiary effluent TSS concentration of six
coagulation/filtration processes were also lower than the national
average value.
3.1.2. Factors affecting the removal of TP and TSS
The influence factors on TP and TSS removal are listed in Table 3.
The stepwise multiple linear regression analysis of TP and TSS are
presented in Tables 4 and 5. The variance inflation factors (VIF) of all
independent variables were not listed in this paper. In this study,
the VIF values of the independent variables in all regression
equations were less than 2, indicating that multicollinearity did not
exist for each of independent variables (Chatterjee and Hadi, 2013).
The independent variables in the regression equation can influence
the dependent variable significantly (P < 0.05), while the inde-
pendent variables which failed to pass the significance test
(P > 0.05) cannot be retained in the regression equation. The
contribution of different independent variables to a dependent
variable could be determined by comparing absolute values of
standardized coefficients of different independent variables in the
regression equation. The greater the absolute value of the stan-
dardized coefficient of independent variables, the greater influence
of independent variables on dependent variables (Chatterjee and

Table 3
The influence factors of TP and TSS removal for coagulation/filtration processes.

Influence factors Unit WWTPA WWTPB WWTPC WWTPD WWTPE WWTPF

T C 18.1 ± 3.6 18.2 ± 3.6 18.1 ± 3.6 18.5 ± 2.5 19.8 ± 3.2 18.0 ± 3.7
pH e 7.4 ± 0.1 7.5 ± 0.1 7.6 ± 0.1 7.5 ± 0.1 7.5 ± 0.2 7.5 ± 0.1
HRT of coagulation tank min e e 10.3 ± 1.8 2.2 ± 0.3 16.8 ± 3.5 5.8 ± 0.5
Filtration rate of D/V filter m/h 14.2 ± 3.9 5.0 ± 1.1 18.0 ± 2.7 9.5 ± 1.3 14.0 ± 3.1 12.8 ± 1.3
TSSin mg/L 12.8 ± 5 16.0 ± 7.7 12.2 ± 5.2 10.1 ± 2.8 10.9 ± 3.8 18.0 ± 8.5
TPin mg/L 0.2 ± 0.1 0.5 ± 0.2 0.6 ± 0.2 0.3 ± 0.2 0.3 ± 0.3 0.7 ± 0.2
PACTP g PAC/g TP removed 27.3 ± 20.7 7.9 ± 7.0 23.7 ± 18.5 38.5 ± 66.1 36.1 ± 42.4 12.0 ± 26.9
PACTSS g PAC/g TSS removed 0.5 ± 0.5 0.3 ± 0.2 1.7 ± 2.0 0.8 ± 0.8 0.9 ± 0.9 0.5 ± 0.8

Note: PAC in the table refers to the effective component of Al2O3.

 D. Wang et al. / Journal of Cleaner Production 170 (2018) 1185e1194 1191

Table 4
The stepwise multiple linear regression analysis of TP removal effect in six tertiary treatment processes.

WWTP Regression equation Standardized coefficient R2 Independent variables

WWTPA RTP ¼ 0.828e0.005PACTPe0.428TPin PACTP ¼ 0.734, TPin ¼ 0.359 0.369 T, pH, VD, TPin, PACTP
WWTPB RTP ¼ 0.832e0.012PACTPþ0.008Te0.024VV PACTP ¼ 0.733, T ¼ 0.261, VV ¼ 0.229 0.678 T, pH, VV, TPin, PACTP
WWTPC RTP ¼ 0.828e0.004PACTP PACTP ¼ 0.58 0.336 T, pH, VD, TPin, PACTP, HRT
WWTPD RTP ¼ 0.653e0.002PACTP PACTP ¼ 0.578 0.334 T, pH, VV, TPin, PACTP, HRT
WWTPE RTP ¼ 0.59e0.003PACTPþ0.198TPin PACTP ¼ 0.571, TPin ¼ 0.286 0.559 T, pH, VD, TPin, PACTP, HRT
WWTPF RTP ¼ 3.715e0.003PACTPe0.398 pH PACTP ¼ 0.663, pH ¼ 0.306 0.557 T, pH, VD, TPin, PACTP, HRT

Table 5
The stepwise multiple linear regression analysis of TSS removal effect in six tertiary treatment processes.

WWTP Regression equation Standardized coefficient R2 Independent variables

WWTPA RTSS ¼ 0.53e0.176PACTSSþ0.009TSSin PACTSS ¼ 0.608, TSSin ¼ 0.294 0.633 T, pH, VD, TSSin, PAC TSS
WWTPB RTSS ¼ 0.849e0.665PACTSS PACTSS ¼ 0.869 0.756 T, pH, VV, TSSin, PACTSS
WWTPC RTSS ¼ 0.479e0.051PACTSSþ0.014TSSin PACTSS ¼ 0.588, TSSin ¼ 0.436 0.799 T,pH,VD,TSSin, PACTSS, HRT
WWTPD RTSS ¼ 0.029e0.089PACTSSþ0.038VVþ0.017TSSin PACTSS ¼ 0.437, VV ¼ 0.292, TSSin ¼ 0.286 0.66 T, pH, VV, TSSin, PACTSS, HRT
WWTPE RTSS ¼ 0.393e0.123PACTSSþ0.019TSSin PACTSS ¼ 0.571, TSSin ¼ 0.383 0.706 T,pH, VD, TSSin, PACTSS, HRT
WWTPF RTSS ¼ 0.63e0.148PACTSSþ0.007TSSin PACTSS ¼ 0.678, TSSin ¼ 0.377 0.838 T,pH,VD, TSSin, PACTSS, HRT

Hadi, 2013). The positive standardized coefficient and the positive
coefficient in regression equation indicate that the independent
variable is positively related to the dependent variable. On the
contrary, the negative coefficient denotes that the dependent var-
iable decreases with the increase of the independent variable.
The multiple linear regression result of TP is listed in Table 4.
PACTP was the most important factor influencing TP removal for six
tertiary treatment processes, according to the judgement of stan-
dardized coefficient of independent variables. The multiple
regression result also clarified that the lower PACTP would be
beneficial to TP removal. Based on the coagulant of FeCl3 for TP
removal, Smith et al. (2008) found that the optimal iron dosage
existed when the removal efficiency of TP reached the highest.
Based on random collisions theory, the iron particles were more
likely to be associated and interacted with other iron particles
rather than phosphate molecules when iron dosage was higher
than the optimal iron dosage, thus decreasing the efficiency of TP
removal. According to Smith et al. (2008), it could be speculated
that PACTP was higher than the optimal dosage for the six coagu-
lation processes. The PACTP of six tertiary treatment processes was
7.9e38.5 g PAC/g TP removed (Table 3). The great range of PAC
consumption demonstrated overdosing of PAC. Although multiple
regression analysis showed that PAC dosage was negatively corre-
lated with the TP removal percentage, the proposed method could
not determine the range of PAC dosage that would lead to a
decrease for TP removal percentage. The optimal PAC dosage for TP
removal could be determined by the jar test in the laboratory, based
on the influent of each tertiary treatment process.
TSS removal efficiencies of six tertiary treatment process had
great relationship with the PAC dosage (Table 5). The smaller PAC
dosage benefited TSS removal. The PAC dosage for TSS removal of
tertiary treatment processes was quite different, ranging from 0.3 g
PAC/g TSS removed to 1.7 g PAC/g TSS removed. Similar to TP
removal, when PACTSS was higher than the optimal dosage, the
aluminum atoms were more likely to interact with other aluminum
atoms rather than TSS, thus reducing the removal efficiency of TSS.
Aparting from the PAC dosage, the tertiary influent TSS also had
significant effect on TSS removal. In this study, PAC dosage had the
greatest effect on TSS removal, followed by the tertiary influent TSS.
For these two types of influence factors, the tertiary influent TSS
concentration was not easily controlled, while the PCA dosage was
an artificially controllable factor. The optimum PAC consumption
for TSS removal could also be determined though laboratory tests.
TP and TSS multiple linear regression results indicated that
removal efficiencies of TP and TSS were greatly influenced by the
PAC dosage, which was consistent with the opinion of Ebeling et al.
(2003) and Clark and Stephenson (1999) that chemical dosage
rather than hydraulic conditions and other factors had the greatest
effect on suspended solid pollutants removal in the coagulation/
flocculation process.
Figs. A.3 and A.4 are the comparison of measured and calculated
removal efficiencies of TP and TSS. The calculated removal effi-
ciency was obtained by the regression equation, and the measured
removal percentage was obtained from the monitored influent and
effluent concentrations in each tertiary treatment process. The
value of R2 above 0.50 indicated that there was a strong linear
relationship between the independent variables and a dependent
variable (Avella et al., 2011). The difference existed between the
calculated and measured results when R2 was less than 0.5,
nevertheless, the independent variables which had significant ef-
fects on the dependent variables could be determined by the
regression equation (Hijosa-Valsero et al., 2011). For WWTPB,
WWTPE and WWTPF, the R2 values for TP removal (Fig. A.3) were
above 0.5, and the R2 values for TSS removal (Fig. A.4) in all tertiary
treatment processes were greater than 0.5. The regression equation
of TP removal in WWTPA, WWTPC and WWTPD confirmed that
PACTP had greater effect on TP removal than other independent
variables, but the linear regression equation could not well reflect
the actual TP removal percentage (R2 < 0.5). In the case of R2 less
than 0.5, the exact functional relationship between the dependent
variable and the significant independent variable needs to be ob-
tained by other statistical methods (Hijosa-Valsero et al., 2011),
which is beyond the scope of this research.
3.2. Economic cost evaluation
The electricity and chemical consumption of tertiary treatment
processes are listed in Table 6. WWTPA D type filter consumed the
largest amount of chemical, which was 691.2 kg PAC/t CODeq
removed and followed by WWTPC Actiflo þ D type filter process.
The tertiary treatment process of WWTPF consumed the minimum
chemical (262.3 kg PAC/t CODeq removed). There was significant
difference in power consumption of six tertiary treatment pro-
cesses. WWTPA D type filter consumed the maximum electricity,
with the value of 952.6 kWh/t CODeq removed. While the elec-
tricity consumption of WWTPF was the least, with the electricity
1192 D. Wang et al. / Journal of Cleaner Production 170 (2018) 1185e1194

Table 6
Chemical and electricity consumption of six tertiary treatment processes.

Unit WWTPA WWTPB WWTPC WWTPD WWTPE WWTPF

Chemical consumption PAC kg/t CODeq removed 691.2 264.1 561.6 495.2 499.9 262.3
Microsand kg/t CODeq removed 5.8
PAM kg/t CODeq removed 4.9
Electricity consumption kWh/t CODeq removed 952.6 368.4 240.2 686.1 566.1 109.6

consumption of 109.6 kWh/t CODeq removed.

The economic costs of six tertiary treatment processes were
obviously different (Fig. 3). The economic costs of electricity and
PAC were proportional to the consumption of them, because of the
same unit price of electricity and PAC in different tertiary treatment
processes. The tertiary treatment processes with the highest and
lowest economic cost were WWTPA and WWTPF, respectively.
From the analysis of influence factors on TP and TSS removal, the
PAC dosage was the most significant influence factor. The closer the
PACTP and PACTSS to the optimal values, the higher the removal
efficiency of TP and TSS. The PACTP and PACSS of six coagulation/
filtration processes were quite different, with the value of
7.9e38.5 g PAC/g TP removed and 0.3e1.7 g PAC/g TSS removed,
respectively, indicating that there was potential waste of PAC. The
reduction of the PAC dosage could not only improve the removal
efficiency of TP and TSS, but also save the economic cost.
Fig. 4. GHG emissions of six coagulation/filtration processes.

3.3. Environmental impact evaluation

The GHG emissions from six tertiary treatment processes are
presented in Fig. 4. The GHG emissions generated from electricity in
six tertiary treatment processes were far greater than that of
chemical, accounting for more than 61% of total GHG emissions. The
difference of GHG emissions between the six tertiary treatment
processes was affected by the removal efficiency of pollutants,
because of the functional unit of COD equivalent removed. The GHG
emissions of six tertiary treatment processes were quite different
(Fig. 4). WWTPA D type filter generated the largest amount of GHG
emissions (740.6 kg CO2/t CODeq removed), because of the highly
intensive electricity and chemical consumption (Table 6), but minor
removal of pollutants (Fig. A.6). The GHG emissions based on the
functional unit of 1 m3 treated wastewater indicated that the ter-
tiary treatment process of WWTPF produced the least GHG, while
there was no significant difference between other five coagulation/
flocculation processes (Fig. A.5). Therefore, compared with the
functional unit of 1 m3 treated wastewater, it was better to reflect
the removal efficiency of tertiary treatment processes if the
functional unit was 1 t CODeq removed.
The tertiary effluent NO3-N was the main indicator that caused
eutrophication (Fig. 5), accounting for 78% of overall eutrophica-
tion. Because coagulation/filtration processes had no significant
effect on TN removal, the final effluent concentration of NO3-N was
mainly affected by the secondary biological treatment process. The
key to reduce the effluent eutrophication of the whole WWTP was
to improve the efficiency of biological nitrogen removal. The pro-
portion of eutrophication caused by TP effluent was only 6%e11%.
Phosphorus was the main limiting factor causing eutrophication
and coagulation/filtration processes contributed greatly to TP
removal by 55%e80%, therefore, coagulation/filtration processes
had positive effect on reducing eutrophication. The effluent
eutrophication of tertiary treatment processes was also affected by
the removal efficiency of pollutants. As shown in Fig. 5, WWTPA D
type filter contributed the most to the eutrophication, with eutro-
phication of 150 kg PO4-Peq/t CODeq removed, due to the worst
removal efficiency, especially for TN and TP removal.

Fig. 3. Economic costs of six coagulation/filtration processes. Fig. 5. Eutrophication of six coagulation/filtration processes.
 D. Wang et al. / Journal of Cleaner Production 170 (2018) 1185e1194
Fig. 6. Comprehensive evaluation of six coagulation/filtration processes.
3.4. Comprehensive evaluation
The comprehensive evaluation results of tertiary treatment
processes are shown in Fig. 6. The GHG costs were the lowest in six
tertiary treatment processes, accounting for less than 3% of total
costs. The economic costs and eutrophication costs of the six ter-
tiary treatment processes were quite different, leading to the
different overall costs. The overall cost of WWTPA D type filter was
the largest, with the comprehensive cost of 324.6 USD/t CODeq
removed, due to the highest economic cost (Fig. 3) but the small
removal amount of pollutants (33 g CODeq/m3). Therefore, WWTPA
had great potential for energy saving and chemical reduction.
WWTPD coagulation tank þ V type filter ranked the second largest
overall cost, which was 223.0 USD/t CODeq removed. For WWTPF
tertiary treatment process, it achieved the greatest improvement in
pollutant removal (126 g CODeq/m3) (Fig. A.6) with the minimal
cost of electricity and chemical (Fig. 3). Therefore, the compre-
hensive cost of WWTPF tertiary treatment process was the least,
with the total cost of 84.8 USD/t CODeq removed.
The results from above analysis were mainly based on the
assumption for the specific conditions. In addition, it should be
noted that the green tax could be modified according to the specific
local preference (Wu et al., 2005). The costs of GHG emissions and
eutrophication would be changed if different green tax was adop-
ted. Considering the green tax as weighting factors, Li et al. (2013)
evaluated a WWTP (in Suzhou, China) by the LCA method, with the
green tax of GHG emissions and eutrophication of 0.03 USD/kg
CO2eq and 0.25 USD/kg NO3eq, respectively, which were about
three times higher than that of green tax in this study. Neverthe-
less, what we are more concerned about is the method itself rather
than the value of green tax. Therefore, the relativity of the evalu-
ation results of six tertiary treatment processes would not be
changed in spite of the change of green tax.
4. Conclusions
The six coagulation/filtration tertiary treatment processes
removed TP and TSS effectively. The PAC dosage was the most
important influencing factor on TP and TSS removal and the PAC
dosage closed to the optimal value would be beneficial to remove
TP and TSS. The PACTP and PACSS of six coagulation/filtration pro-
cesses were 7.9e38.5 g PAC/g TP removed and 0.3e1.7 g PAC/g TSS
removed, respectively. The proper reduction of the PAC dosage
could not only help improve the removal efficiency of TP and TSS,
but also save the economic cost. The coagulation/filtration pro-
cesses had positive effect on reducing eutrophication. The
1193
contribution of six tertiary treatment processes to the reduction of
eutrophication was mainly from TP removal. The tertiary treatment
process of WWTPF achieved the best comprehensive evaluation
result due to the lowest composite cost index. While the composite
cost index showed that D type filter in WWTPA had great potential
for energy saving and chemical reduction.
The present study showed that the multiple linear regression
method is an effective tool to clarify the significant factors affecting
the removal of TP and TSS in tertiary treatment processes, and it
helps the managers or decision makers in WWTPs to improve the
operating performance by adjusting the PAC dosage to the optimal
value. But the optimal PAC dosage should be determined by labo-
ratory test. The proposed composite cost index method for tertiary
treatment process in technical, economic, and environmental as-
pects is an integrated evaluation method and it is also applicable to
other wastewater treatment processes or WWTPs. The sustain-
ability of a WWTP usually involves technology, economy, envi-
ronment, society and so on, and the composite cost index method
provides a feasible evaluation approach for realizing the sustain-
ability of technology, economy and environment.
Acknowledgement
This research was supported by the Major Science and Tech-
nology Program for Water Pollution Control and Treatment of China
(2012ZX07302002).
Appendix A. Supplementary data
Supplementary data related to this article can be found at
https://doi.org/10.1016/j.jclepro.2017.09.231.
References
€ rner, T., Yvon, J., Chappeb, P., Guinot-Thomasb, P., de Donato, P., 2011.
Avella, A.C., Go
A combined approach for a better understanding of wastewater treatment
plants operation: statistical analysis of monitoring database and sludge
physico-chemical characterization. Water Res. 45 (3), 981e992.
Benedetti, L., Dirckx, G., Bixio, D., Thoeye, C., Vanrolleghem, P.A., 2008. Environ-
mental and economic performance assessment of the integrated urban
wastewater system. J. Environ. Manage 88 (4), 1262e1272.
Bott, C.B., Parker, D.S., Jimenez, J., Miller, M.W., Neethling, J.B., 2012. WEF/WERF
study of BNR plants achieving very low N and P limits: evaluation of technology
performance and process reliability. Water Sci. Technol. 65 (5), 808e815.
Chatterjee, S., Hadi, A.S., 2013. Regression Analysis by Example. John Wiley & Sons.
Clark, T., Stephenson, T., 1999. Development of a jar testing protocol for chemical
phosphorus removal in activated sludge using statistical experimental design.
Water Res. 33 (7), 1730e1734.
Copp, J.B., Spanjers, H., Vanrolleghem, P.A., 2002. Respirometry in Control of the
Activated Sludge Process: Benchmarking Control Strategies. IWA Publishing,
Conwall, United Kingdom.
Corominas, L., Foley, J., Guest, J.S., Hospido, A., Larsen, H.F., Morera, S., Shaw, A.,
2013. Life cycle assessment applied to wastewater treatment: state of the art.
Water Res. 47 (15), 5480e5492.
de Haas, D., Foley, J., Barr, K., 2008. Greenhouse gas inventories from WWTPs-the
trade-off with nutrient removal. In: Water Environment Federation Sustain-
ability Conference, National Harbour, MD.
Dong, X., Zhang, X.Y., Zeng, S.Y., 2017. Measuring and explaining eco-efficiencies of
wastewater treatment plants in China: an uncertainty analysis perspective.
Water Res. 112, 195e207.
Ebeling, J.M., Sibrell, P.L., Ogden, S.R., Summerfelt, S.T., 2003. Evaluation of chemical
coagulationeflocculation aids for the removal of suspended solids and phos-
phorus from intensive recirculating aquaculture effluent discharge. Aquac. Eng.
29 (1), 23e42.
Friedler, E., Katz, I., Dosoretz, C.G., 2008. Chlorination and coagulation as pre-
treatments for greywater desalination. Desalination 222 (1), 38e49.
Guida, M., Mattei, M., Della Rocca, C., Melluso, G., Meriç, S., 2007. Optimization of
alum-coagulation/flocculation for COD and TSS removal from five municipal
wastewater. Desalination 211 (1e3), 113e127.
Guinee, J.B., Gorree, M., Heijungs, R., Huppes, G., Kleijn, R., de Koning, A., van
Oers, L., Weneger Sleeswijk, A., Suh, S., Udo de Haes, H.A., de Bruijn, H., van
Duin, R., Huijbregts, M.A.J., 2002. Handbook on life cycle assessment: opera-
tional guide to the ISO standards. Int. J. Life Cycle Assess. 7 (5), 311e313.
Hijosa-Valsero, M., Sidrach-Cardona, R., Martin-Villacorta, J., Valsero-Blanco, M.C.,
1194 D. Wang et al. / Journal of Cleaner Production 170 (2018) 1185e1194
Bayona, J.M., Becares, E., 2011. Statistical modelling of organic matter and
emerging pollutants removal in constructed wetlands. Bioresour. Technol. 102
(8), 4981e4988.
Illueca-Mun ~ oz, J., Mendoza-Roca, J.A., Iborra-Clar, A., Bes-Pia, A., Fajardo-
Montanana, V., Martinez-Francisco, F.J., Bernacer-Bonora, I., 2008. Study of
different alternatives of tertiary treatments for wastewater reclamation to
optimize the water quality for irrigation reuse. Desalination 222 (1e3),
222e229.
Li, X.D., Zhu, Y.M., Zhang, Z.H., 2010. An LCA-based environmental impact assess-
ment model for construction processes. Build. Environ. 45 (3), 766e775.
Li, Y., Luo, X.Y., Huang, X.W., Wang, D.W., Zhang, W.L., 2013. Life Cycle Assessment of
a municipal wastewater treatment plant: a case study in Suzhou, China. J. Clean.
Prod. 57, 221e227.
Lorenzo-Toja, Y., Vazquez-Rowe, I., Amores, M.J., Termes-Rife, M., Marin-Navarro, D.,
Moreira, M.T., Feijoo, G., 2016. Benchmarking wastewater treatment plants
under an eco-efficiency perspective. Sci. Total Environ. 566, 468e479.
Mannina, G., Ekama, G., Caniani, D., Cosenza, A., Esposito, G., Gori, R., Garrido-
Baserba, M., Rosso, D., Olsson, G., 2016. Greenhouse gases from wastewater
treatment-A review of modelling tools. Sci. Total Environ. 551, 254e270.
Ministry of Environmental Protection of the People’s Republic of China (MOEP),
2002. Discharge Standard of Pollutants for Municipal Wastewater Treatment
Plant. GB18918e2002. Beijing, China (in Chinese).
Molinos-Senante, M., Go mez, T., Garrido-Baserba, M., Caballero, R., Sala-Garrido, R.,
2014. Assessing the sustainability of small wastewater treatment systems: a
composite indicator approach. Sci. Total Environ. 497, 607e617.
Molinos-Senante, M., Gemar, G., Gomez, T., Caballero, R., Sala-Garrido, R., 2016. Eco-
efficiency assessment of wastewater treatment plants using a weighted Russell
directional distance model. J. Clean. Prod. 137, 1066e1075.
Nardo, M., Saisana, M., Saltelli, A., Tarantola, S., 2008. Handbook on Constructing
Composite Indicators: Methodology and User Guide. OECD.
National Bureau of Statistics of China (NBSC), 2015. China Statistical Yearbook.
China Statistics Press, Beijing, China available from: http://www.stats.gov.cn/
tjsj/ndsj/2015/indexch.htm. (Accessed 15 August 2017). in Chinese.
Norton-Brand~ ao, D., Scherrenberg, S.M., van Lier, J.B., 2013. Reclamation of used
urban waters for irrigation purposes-a review of treatment technologies.
J. Environ. Manage 122, 85e98.
Plakas, K.V., Georgiadis, A.A., Karabelas, A.J., 2016. Sustainability assessment of
tertiary wastewater treatment technologies: a multi-criteria analysis. Water Sci.
Technol. 73 (7), 1532e1540.
Pophali, G.R., Chelani, A.B., Dhodapkar, R.S., 2011. Optimal selection of full scale
tannery effluent treatment alternative using integrated AHP and GRA approach.
Expert Syst. Appl. 38 (9), 10889e10895.
Quadros, S., Rosa, M.J., Alegre, H., Silva, C., 2010. A performance indicators system
for urban wastewater treatment plants. Water Sci. Technol. 62 (10), 2398e2407.
Rahman, S.M., Eckelman, M.J., Onnis-Hayden, A., Gu, A.Z., 2016. Life-cycle assess-
ment of advanced nutrient removal technologies for wastewater treatment.
Environ. Sci. Technol. 50 (6), 3020e3030.
Remy, C., Miehe, U., Lesjean, B., Bartholomaus, C., 2014. Comparing environmental
impacts of tertiary wastewater treatment technologies for advanced phos-
phorus removal and disinfection with life cycle assessment. Water Sci. Technol.
69 (8), 1742e1750.
Rodriguez-Garcia, G., Molinos-Senante, M., Hospido, A., Hernandez-Sancho, F.,
Moreira, M.T., Feijoo, G., 2011. Environmental and economic profile of six ty-
pologies of wastewater treatment plants. Water Res. 45 (18), 5997e6010.
Smith, S., Takacs, I., Murthy, S., Daigger, G.T., Szabo, A., 2008. Phosphate complex-
ation model and its implications for chemical phosphorus removal. Water En-
viron. Res. 80 (5), 428e438.
Song, R.P., Zhu, J.J., Hou, P., Wang, H.T., 2013. Getting Every Ton of Emissions Right:
an Analysis of Emission Factors for Purchased Electricity in China. World Re-
sources Institute, Beijing. Full report available at: http://www.wri.org/
publication/analysis-of-emission-factors-forpurchased-electricity-in-china (in
Chinese).
Sun, Y., Chen, Z., Wu, G.X., Wu, Q.Y., Zhang, F., Niu, Z.B., Hu, H.Y., 2016. Character-
istics of water quality of municipal wastewater treatment plants in China:
implications for resources utilization and management. J. Clean. Prod. 131, 1e9.
Üstün, G.E., Solmaz, S.K.A., Çiner, F., Baskaya, H.S., 2011. Tertiary treatment of a
secondary effluent by the coupling of coagulation-flocculation-disinfection for
irrigation reuse. Desalination 277 (1), 207e212.
Wu, X., Zhang, Z., Chen, Y., 2005. Study of the environmental impacts based on the
“green tax” -applied to several types of building materials. Build. Environ. 40
(2), 227e237.
Zeng, G.M., Jiang, R., Huang, G.H., Xu, M., Li, J.B., 2007. Optimization of wastewater
treatment alternative selection by hierarchy grey relational analysis. J. Environ.
Manage 82 (2), 250e259.
Zhang, Y.Y., Ma, H.L., Wang, B., Qu, W.J., Wali, A., Zhou, C.S., 2016. Relationships
between the structure of wheat gluten and ACE inhibitory activity of hydroly-
sate: stepwise multiple linear regression analysis. J. Sci. Food Agric. 96 (10),
3313e3320.