1. Introduction
According to United Nations statistics, by 2025, two-thirds of the world’s population may face water shortages [
1]. If discharged without treatment, municipal and industrial wastewater will cause damage to the aquatic ecosystem through eutrophication and fish poisoning, as well as adverse effects on human health due to the discharge of pathogenic organisms in sewage to recreational water bodies [
2]. However, with proper treatment, water pollution can be minimized and wastewater can be used as a resource for nutrients and recycled water. Therefore, wastewater treatment technology plays a vital role in realizing the sustainable future of human society [
3].
Developed countries have experienced the process of “pollution first, treatment later” of the water environment. Domestic sewage treatment developed rapidly after the 1970s. By the end of the 1990s, the average domestic sewage treatment rate in developed countries had reached a relatively high level of over 80% [
4]. Among them, New Zealand, Singapore, and Northern Europe have basically achieved 100% collection and processing rates. In developed countries, sewage treatment facilities are mainly small and medium-sized. For example, there are more than 14,780 sewage treatment plants nationwide in the United States, with an average daily processing water volume of about 6000 tons per plant, of which small and medium-sized sewage treatment plants with a capacity of less than 50,000 tons per day account for about 85% of the total [
5]. The United States is currently the country with the largest number of sewage treatment plants in the world, with an average of 1 in 5000 people, 78% of which are secondary biological treatment plants; the United Kingdom has about 8000 treatment plants with an average of 1 in 7000 people, almost all of which are secondary biological treatment plants. There are about 630 urban wastewater treatment plants in Japan, with an average of 1 in 200,000 people, but secondary treatment plants and advanced treatment plants account for 98.6%; Sweden is currently the country with the most popular sewage treatment facilities, with a sewer penetration rate of over 99%, with an average of 1 in 5000 people, of which 91% are secondary biological treatment plants [
6].
With increasing emphasis on China’s water environment protection, more stringent regulations and standards have been legislated governing the discharge of organic pollutant, nitrogen and phosphorus in effluent of municipal wastewater treatment plants (WWTPs). Therefore, there is a significant need to upgrade the existing municipal WWTPs to meet the strict discharge standards. For example, the Action Plan for Water Pollution Prevention and Control issued in 2015 by China [
7] requests effluent from all WWTPs into the receiving water bodies to meet with requirements of the Grade I-A of the Discharge Standard of Pollutants for Municipal Wastewater Treatment Plant (GB 18918–2002), e.g., chemical oxygen demand (COD), total nitrogen (TN), total phosphorus (TP) concentration limits of effluent should be upgraded to 50 mg/L, 15 mg/L and 0.5 mg/L of the Grade I-A from 60 mg/L, 20 mg/L and 1.0 mg/L of the Grade I-B, respectively. It was reported that the total municipal wastewater discharged in China was 49.24 billion tons in 2017, increased by 30.02% and 5.53% compared with 2010 and 2015, respectively, along with rapid urbanization [
8]. Among the over 4000 municipal WWTPs in China, the small and mid-scale municipal WWTPs with the design capacity less than 50,000 m
3/d accounted for 83.3% in 2017 [
9], most of which need to be upgraded or optimized to meet the discharge requirements of Grade I-A. Generally, the upgrading of the existing sewage treatment plants can be considered from two perspectives: one is only the effluent water quality of WWTPs, the other is both the effluent water quality and the design capacity of WWTPs [
10], and the most important goal of upgrading existing municipal wastewater treatment plants is to ensure high treatment efficiency to meet wastewater quality standards, while keeping investment and operating costs to a minimum by focusing on energy saving and cost reduction methods of bioprocesses, key equipment and hydrodynamics in bioreactors. However one of major challenges of upgrading the existing municipal WWTPs in China, especially those in the small and mid-size cities & towns of the cold region, to meet the Grade I-A standards of GB 18918–2002 is the TN limit of 10 mg/L in the effluent while keeping the investment and operating cost as low as possible [
11,
12].
Wastewater treatment can be divided into physical, chemical and biological treatment methods based on the type of water quality [
13]. According to the degree of treatment, they can also be divided into primary, secondary and tertiary treatment methods [
14]. The physical treatment method of urban sewage is a method that uses physical action to separate and remove pollutants in sewage [
15], including screening and interception, gravity separation, centrifugal separation. Chemical treatment methods are rarely used in urban sewage treatment, and generally involve other chemical methods in urban water supply treatment, such as neutralization, oxidation-reduction, ion exchange, and electrolysis, which are mainly used for industrial wastewater treatment [
16]. Chemical method must be used in conjunction with the physical method. Before chemical treatment, precipitation and filtration are often used as pre-treatment; in some cases, physical means such as precipitation and filtration are required as post-treatment of chemical treatment. Biological treatment is a method that uses the metabolism of microorganisms to remove organic substances in sewage. Commonly used are activated sludge method, biofilm method, as well as oxidation pond and sewage land treatment method [
17].
It is well known that the activated sludge processes are the most applied biological approach for nutrient removal for municipal wastewater treatment in the world. Traditionally, municipal WWTPs have been designed and operated using empirical steady-state equations or ‘rules of thumb’, introducing conservative safety factors that have led to the over-dimensioned, expensive construction and unstable functioning plants [
18]. Different modifications have been thus proposed for the conventional activated sludge processes to meet the strict effluent standards for COD, biological oxygen demand (BOD), nitrogen and phosphorus [
19].
Due to the nonlinear dynamics and uncertainty, the internal process dynamics have multiple time scales and multivariable structures, so activated sludge wastewater treatment plants can be classified as complex systems, and such traditional approaches cannot meet increasing requirements for municipal WWTPs. Process evaluation, design optimization, and cost analysis can be performed by simulation and modeling tools [
20,
21,
22]. Numerical simulation based on activated sludge models is getting an increasing attention, and these models are incorporated in commercial simulation packages such as BioWin (EnviroSim Associates, Flamborough, Ontario, Canada), GPS-X (Hydromantis Inc., Ontario, Canada) and WEST (HEMMIS Inc., East Flanders, Belgium and DHI Inc, Copenhagen, Denmark) for engineering practice [
23,
24]. Numerical simulation has been widely regarded as helpful tools for evaluation, diagnosing and optimizing WWTPs’ design, and operation and control, comparison and selection of the biological treatment processes as well as upgrade of the existing WWTPs [
25,
26,
27]. A series of activated sludge models (i.e., ASM1, ASM2, ASM2d, ASM3) formulated and introduced by IWA (International Water Association) have been considered as indispensable solutions in correlating the complexity of the activated sludge process and the prediction of biological treatment efficiency under dynamic conditions [
28,
29,
30].
The ASM1 is the primary version, which is a structured model based on Monod kinetics to realistically predict the performance of carbon oxidation, nitrification and denitrification in activated sludge systems under aerobic and anoxic conditions. Among these available models, ASM1 has been considered as a reliable reference model due to its most widely applications for the academic and operational fields [
31,
32,
33].
Nevertheless, one of the major limitations for a more widespread application of ASM is to choose a set of related parameters that are essential for achieving a good prediction of the model used [
34]. In the neutral and relatively constant conditions, there are many environmental factors that influence the dynamics and stoichiometric parameters of ASM1 models, but influent characteristics and temperature are the two most common environmental factors [
30,
35]. Most of the parameters are affected by the specific components of the influent [
36,
37,
38], which either promote or inhibit the values of kinetic parameters and stoichiometric coefficients.
However, the characteristics of the influent fractions can change in different seasons during one year [
39]. Different seasons and influent characteristics lead to different kinetic parameters. A small water temperature range (low temperature, medium temperature, high temperature) generally causes the rate coefficient changing (e.g.,
μH,max,
bA or
kh). Moreover, almost all kinetic parameters are affected by water temperature, so the influence of temperature should be considered when setting values (Henze et al., 2000). Although many literatures have proposed different water temperature correction factors [
40,
41,
42,
43], most of them are not universally suitable for all the researches. In a word, it is necessary to simultaneously determine the concentration of influent characteristics and the key kinetic parameters at different temperatures so the models can be accurately applied to the design, operation and upgrade of existing wastewater treatment systems [
30].
Flow and hydrodynamic characteristics play an important role in the stable operation of an activated sludge processes system, especially for the denitrification process within a limited residence time. While in most of the anaerobic or anoxic tanks, the submersible agitators were installed on the basis of general empirical guidelines, operators’ experience or ‘rules of thumb’, which cannot guarantee the evenly mixing of the flow field. Improper installation of the agitators can also cause the damage to the blades. CFD is a powerful tool to simulate the hydrodynamics and mass transfer, and has become increasingly popular in optimizing design and operation of WWTPs [
44,
45,
46,
47,
48], therefore such a study of the agitators’ location and position in the bioreactor has become easier and less expensive.
In China, the Sequencing Batch Reactor (SBR) and its variant processes was the third widely used activated sludge process at 17.19% of municipal WWTPs, while the first two were the oxidation ditch (OD) process at 29.21% and the anaerobic–anoxic–oxic (A
2/O) process at 25.45% [
49]. The CASS process is a variant of the SBR process, in which nitrogen is removed mainly by simultaneous nitrification and denitrification [
50]. The CASS process has been widely used in municipal and industrial WWTPs, e.g., over 400 small and mid-scale municipal WWTPs, especially those in cold regions of northern China because of its configuration flexibility, operational simplicity, low construction and maintenance costs, and simultaneous removal of nitrogen and phosphorus [
51]. It is obvious that treatment performance of small and mid-scale municipal WWTPs in the cold regions of northern China, especially the TN removal, has significant difference in summer and winter, e.g., unstable TN removal and TN concentration of effluent usually does not meet the discharge standard in winter [
52]. Considering key factors such as large fluctuations in water quality and quantity of influent, unstable performance of wastewater treatment in winter, selection of new process or optimization of the existing process, constraints of existing structure and field, and limits of capital & operational costs, it has been suggested that the mathematical ASM model will be a very powerful and cost-effective tool for upgrading, designing, operating and optimizing existing small and medium-scale sewage treatment plants in the cold regions of northern China.
Thus, a hypothesis in this study is that the model-based upgrading of an existing full-scale WWTP is feasible through optimizing the existing biological process to comply with TN of effluent quality criteria while keeping the capital and operational costs as low as possible. A full-scale municipal WWTP with the CASS process at a design capacity of 20,000 m3/day, located in Zhangjiakou, Hebei Province, was selected to upgrade TN removal on the basis of ASM1 modelling and optimization, as well as CFD simulation and optimization for the agitators’ layout. In this study, the modelling, simulation and optimization of this full-scale CASS process WWTP was carried out by ASM1 using a commercial software package GPS-X 8.0 (Hydromantis Inc., Ontario, Canada) to evaluate the treatment performance and diagnose the bottlenecks of operation in summer and winter, determine the influent COD fractions and calibrate the model kinetic parameters. Then, an upgrade solution of this WWTP with different operational strategies and multiple optimization alternatives, as well as the optimized location and position of the agitators in CASS tank with the aid of CFD simulation, were made and carried out to make effluent quality meet the requirements of Grade I-A of the GB 18918–2002. In addition, treatment performance (especially TN removal) and energy consumption of this WWTP were compared before and after the upgrading of this WWTP, and the evolution of the microbial community of the CASS process with seasonal changes was also investigated based on high-throughput 16S rRNA gene sequencing analysis.
2. Materials and Methods
2.1. Full-Scale Wastewater Treatment Plant and CASS Process
In this study, a full-scale municipal WWTP with the CASS process located in northwest of Hebei province (North China) was selected, which design capacity is 20,000 m
3 of domestic wastewater per day [
53]. This WWTP consists of four CASS systems arranged in parallel, mainly including a mechanical treatment unit of primary settling tank to remove floating and settleable solids, a biological treatment unit with activated sludge process for COD and nutrients removals, a sludge treatment unit of dewatering. Before upgrading, the bioreactor of the CASS consisted of a pre-denitrification anoxic zone (L × W × H = 4.3 m × 17.6 m × 5.6 m, with effective volume of 423.8 m
3), in which there are two submersible agitators installed on the opposite corners to make the anoxic tank in good mixing performance. Without submersible agitators, the aeration in the nitrification aerobic zone (L × W × H = 34 m × 17.6 m × 6.5 m, effective depth 5.5 m with working volume of 3,291.2 m
3) was obtained with fine-pore air diffusers located at the bottom of the bioreactor to make the DO concentrations at 2–3 mg·L
−1. A schematic flow diagram of the biological step of the CASS system was shown in
Figure 1a.
The CASS process was operated under two different strategies in summer (June, July and August) and winter (December, January and February), respectively (
Figure 1b). Before upgrading, each CASS cycle was designed for 6 h, while the effluent cannot meet discharge standard of TN limit at 15 mg·L
−1 (GB18918–2002) in winter (
Figure S1). In summer mode, 90 min was allocated to batch feeding without the submersible agitator operation and most of the denitrification process takes place in the front anoxic zone. The nitrification of ammonia nitrogen was achieved by aeration for 210 min. For the last phase, settling and decant time were the same for 60 min, and 30 min for idling phase. The only difference between the summer mode and winter mode was the first two phases. In winter mode, the first two phases consist of 60 min for mixing during the time of 90 min of filling, and the aeration phase was shortened to 150 min. These two kinds of operation strategies depended on the influent characteristics as shown in
Table 1. From the beginning of September 2018, the TN of effluent must meet Grade I-A discharge standard of GB18918–2002, while the residence time of the pre-denitrification anoxic zone is too short to meet the standard nitrogen emission. Because the wastewater filling time of a single CASS system is 90 min, in order to ensure the continuous water supply of the four series as a whole, sometimes the idle period is increased in the operation cycle.
To maintain the biomass, the returned sludge, nearly 50% of the influent flow rate in summer mode, while 100% in winter, from the end of SBR zone was recirculated into the anoxic basin. The MLSS was controlled at about 4000 mg·L−1 in the main SBR zone after filling phase, while approximately 5200 mg·L−1 after decanting with a volumetric exchange ratio of 22.73% for each cycle. The SRT was controlled at nearly 21 d by discharging an appropriate amount of sludge.
2.2. ASM1 Modelling
The ASM1 was used in this study for its good description of the activated sludge process [
41,
54]. Firstly, the influent COD fraction, heterotrophic yield,
YH, and heterotrophic decay rate,
bH, were estimated by the respirometry test in summer (20 °C) and winter (10 ℃), respectively. Secondly, the essential kinetic and stoichiometric model parameters were determined through sensitive analysis. Thirdly, the results of the simulation based on the field investigation data of this WWTP were compared with the actual treated effluent water quality. Fourthly, an upgrade solution with different operational strategies and multiple optimization alternatives, as well as the optimized location of the stirrers in CASS tank with the aid of CFD simulation, were put forward and carried out to achieve better effluent quality.
2.3. Respirometry Test
Respirometry is the useful method in diagnose the COD fractions and some key model parameters [
37,
41]. This method has been widely used in the ASM modelling [
55,
56,
57]. The respirometry experiments were carried out in a 2 L closed vessel which was magnetic stirred to make the proportional liquor well-mixed, and the dissolved oxygen (DO) electrode (Oxi 3205, WTW, Berlin, Germany) was fixed in the vessel to monitor the DO data intermittently. The temperature of these experiments was kept at 20 ℃ for summer samples while 10 ℃ for winter samples throughout the respirometry test period.
The activated sludge samples taken from the CASS system in different seasons (summer and winter in 2018) were aerated for 24 h before use to ensure the endogenous state at the beginning of the experiments. After aeration, in order to remove the external COD, the static sedimentary sludge was then washed for three times with distilled water. Six hundred mL wastewater and 400 mL washed sludge were mixed in a closed vessel according to the actual system operation F/M value. All tests were carried out under the conditions of adding 20 mg·L−1 of nitrification inhibitor (Allylthiourea, ATU, Shanghai, China) to limit the oxygen consumption by nitration. DO concentration in the vessel was recorded intermittently with the dissolved oxygen electrode (Oxi 340i, WTW, Berlin, Germany). The first step was to rapidly increase the DO of the mixture up to 6 mg·L−1 with strong aeration followed by a decrease phase of DO concentration to 2 mg·L−1 by turning off the air pump. The data sampling frequency is once every 10 s, which is relatively high in the initial stage for the high OUR rate due to Ss degradation. The dissolved oxygen in the reactor will drop to 2 mg·L-1 in a short period of time (<5 min). At this time, the data reading will be suspended, the aerator will be used to oxygenate the mixture system in the vessel. The aerator was turned off after the DO was above 6 mg/L again. Then, the change of the DO in the reactor was recorded intermittently. The above steps were continued until the value of the oxygen consumption rate becomes constant. Thus, the OUR of the sludge can be calculated by the slope of the DO concentration reduction during period in which the aeration was turned off.
2.3.1. Heterotrophic Yield YH
The actual heterotrophic yield coefficient,
YH, which must be known before determining the soluble rapidly biodegradable organic,
SS [
37,
58]. It was evaluated by a respirometry test in which four different concentrations of fully biodegradable organic substrate were added to the our system in which the sludge was in endogenous phase [
59]. Sodium acetate (AP-10018818, Sinopharm Chemical Reagent Co., Ltd., Shanghai, China) was used in this test since it was the external carbon source added during the anoxic periods of the case studied [
59,
60].This test was operated at 20 °C and 10 °C, respectively, and a pH value of 7.5 ± 0.1 with a low concentration of MLVSS, which provided a suitable low OUR that improved the
YH assessment.
According to the literature [
60], the heterotrophic yields can be calculated from the following equation:
where OC is the accumulated exogenous oxygen consumption,
SAce is the concentration of the added sodium acetate. The OC in summer and winter was calculated according to the added biodegradable organic substrate concentration, respectively.
2.3.2. Heterotrophic Decay Rate bH
The heterotrophic decay rate,
bH, is critical for the prediction of sludge production and oxygen demand [
30]. Therefore, it must be determined based on the activated sludge used in the OUR test of
bH.
Since heterotrophic bacteria were in the endogenous respiratory stage after 24 h of aeration, OUR can only be caused by the microbial auto-oxidation. There was a linear relationship between ln (OUR) and time
t, the slope of the curve is a negative traditional decay rate which follows the next formula [
41]:
where
fvC is a proportional constant; K
d is global attenuation coefficient, X
0B,H is initial concentration of active heterotrophic biomass (mg/L)
The
bH can be calculated from the next equation which is related to
Kd [
30]:
For each summer and winter, 1500 mL activated sludge taken from the CASS system of this WWTP was put in a batch reactor, then washed with the distilled water after 24 h aeration, and the OUR was further measured for multiple times in 4 days [
41]. During the experiments, pH and temperature were controlled the same as the values in summer and winter, respectively.
2.3.3. COD Fractionation of Influent
In ASM1 model, the constituent elements of total COD were presented by equation:
where:
SI–soluble inert substrates, g O2 m−3;
SS–soluble readily biodegradable substrates, g O2 m−3;
XI–inert particulate organic material, g O2 m−3;
XS–particulate slowly biodegradable substrates, g O2 m−3;
According to the Equation (4), the influent COD fractions can be divided into the above four components and be used as the influent simulation data for the ASM1 model.
The soluble COD value (SCOD) of sewage is usually defined as the COD through a 0.45 μm microfiltration membrane. According to the literature [
61], the floc in the zinc sulfate coagulation filtration method has a small adsorption amount to the SCOD, and the supernatant COD is close to the truly solution COD. Therefore, the steps of the physical and chemical separation method to determine the SCOD were as follows: Firstly, 10 mL of 100 g·L
−1 zinc sulfate was added to 1 L of the wastewater sample to be tested, and the pH was adjusted to about 10.5 with 6 mol·L
−1 NaOH under the rapid stirring of the magnetic stirrer; Secondly, the liquor was mixed at high speed (120 r·min
−1) for 1 min and low speed (60 r·min
−1) for 5 min by magnetic stirrer, and then static settlement for 15 min; At last, conventional disposable filters with pore size of 0.45 μm (Durapore
® Membrane Filter, polyvinylidene fluoride (PVDF), New York, USA) were used for sequential filtration after sedimentation. The COD value of the filtrate is the wastewater SCOD [
36].
The OUR curve of the mixture could be divided into three segments. In S1, t < t1, the OUR value sharply decreased because the substrate was easily biodegradable; in S2, t1 < t < t2, OUR slowly decreases, the rate of which is controlled by the substrate provided by the slow degradation of the substrate XS in the wastewater; in S3, t > t2, OUR is almost constant and is maintained at a relatively low level, which can be considered as the oxygen consumption of endogenous respiration of the original heterotrophic microorganisms before mixing with wastewater. As a result, the SS and XS could be calculated from the partial OUR curves in S1 and S2 individually.
To get a clear description of the OUR curves for COD fractions, before the respiroetric tests, the volume of the wastewater,
VW, and the activated sludge,
VS, should be determined to get a suitable F/M ratio (the ratio between the SCOD of influent value and the MLVSS). The suggested F/M ratio for the OUR tests was between 0.01 and 0.2 mg COD/mg VSS [
38]. For each season’s OUR test, triplicate experiments were performed to get a good repeatability.
SS and
XS could be calculated by the following equations:
where
OURtot is total oxygen uptake rate (mg/(L·d
−1)),
OURXs is oxygen uptake rate of
XS consumption and endogenous respiration (mg/(L·d
−1)),
OURER is oxygen uptake rate of endogenous respiration (mg/ (L·d
−1))
2.4. Model Calibration and Validation Strategy
In this study, the GPS-X software (Hydromantis Inc., Ontario, Canada) with ASM1 model was used to simulate the CASS system. Two key parameters (
YH,
bH) were measured based on the respirometry batch experiment, other model parameters were mainly corrected according to the sensitivity analysis. The input data for the simulations came from the CASS WWTP and comprised the period of June 2018 for the calibration and July 2018 for the validation for summer simulation, December 2018 for calibration and January 2019 for validation for winter simulation, respectively.
Table 1 lists the operating parameters and wastewater characteristic values used for model calibration and verification.
The validated model was used for simulation and optimization to achieve complete denitrification in wastewater, thereby providing upgrade solutions to meet TN emission standards and reduce operating costs. Optimization was completed by running multiple simulations on different operating scenarios, such as reducing aeration time and testing different operation modes in the CASS tanks, increasing anoxic time to perform nitrification-denitrification. The optimized operation strategies were applied to the CASS operation in practice from August 2018 to February 2019, respectively, to acquire the optimized operation data from the WWTP.
2.5. Sensitivity Analysis
Sensitivity analysis can assess to what extent the parameters used in model calibration affect the output of the model [
62]. Sensitivity analysis before model calibration is necessary to evaluate important parameters [
63]. According to EPA guidelines [
64], the sensitivity coefficient (
Si,j) is defined as a ratio of the percentage change in the output variable (
yi) to a 10% change in the input variable (
xi):
In this study, all model coefficients (including kinetic coefficients and stoichiometric coefficients) were changed by 10% in the simulation. The determination of key parameters during calibration depends on the sensitivity of the model output to these parameters. The effect of parameters on the model output can be explained as: (1) < 0.25 means that the parameter has no significant effect on the model output, (2) 0.25 ≤ |
Si,j| <1 means that the parameter has an effect; (3) 1 ≤ |
Si,j| < 2 indicates that the parameters are very influential; (4) |
Si,j| ≥ 2 indicates that the parameters are extremely influential [
65].
2.6. CFD Modelling
In full-scale WWTPs, submerged agitators are always used to control high-flow mixing, and it is generally known that the single performance parameter thrust (F) is the basis of the design of the agitation system and the positioning principle of a series of agitators [
57]. CFD can be used to model the mixer in detail, but it is too complicated to be included in a wide range of plant models, instead, at the same agitators geometrical location, a simile was used to contain the mechanical momentum added by the agitators in the system as the momentum source M (kg·m
−2·s
−2) [
66]. The detail description used in this section was referred to the literature [
67].
In order to evaluate the effect of the two different installation layouts of the agitators on the fluid dynamics and its influence on the kinetic model, a CFD three-dimensional single-phase method was performed by ANSYS® Academic Research Release 18.0 software (ANSYS, Inc., New York, NY, USA) which was used as the CFD modelling in this study. The different simulations described in this section were run on ANSYS-Fluent against traditional and modified configurations to achieve these goals. The second-order upwind and PRESTO scheme, which was suitable for swirling flow, were used for discrete spatial derivatives and discrete pressure, respectively. The semi-implicit method of pressure link equation (SIMPLE) is used to realize the coupling of speed and pressure. At the beginning of the simulation, the under-relaxation factor is reduced to maintain stability and avoid solution divergence. When the proportional residual continuity drops below 1 × 10−4 and the velocity and turbulence drop below 1 × 10−5, the solution is considered to be fully converged.
2.7. High-Throughput 16S rRNA Gene Sequencing
In order to reveal the impact of microbial population changes on the performance of the CASS system, activated sludge samples collected in four seasons were selected as samples for 16S rRNA gene amplicon sequencing. These samples were processed in order to perform DNA extraction and 16S rRNA gene PCR, which was followed by the amplification and purification of PCR products. To extract DNA, 2 mL sludge sample was mixed with the DNA extraction kit according to the manufacturer’s instructions. In order to ensure the accuracy of the extraction, four times of DNA samples were extracted from each sample and then mixed evenly. The DNA extracts were purified and stored at –20 °C until the analysis. For PCR amplification, bacterial 16S rRNA fragments were amplified by adding different eight-base barcodes to the forward primer (5′-GTGCCAGCMGCCGCGGTAA-3′) and reverse primer (5′- GCCAGCMGCCGCGGTAA-3′) of each sample.
After PCR amplification of 16S rRNA gene, the amplicons were purified using SanPrep DNA gel extraction kit. Before transferring the purified PCR product to the sequencing analysis step, it is quantified. Finally, Shanghai Sangong Biotechnology Co., Ltd. (Shanghai, China) performed high-throughput sequencing on the Illumina sequencing platform