With rapid urbanization, great strains are not only being placed on the production of clean water; there is also an equal, if not, more important need to treat the increasing quantities of wastewaters being produced. With urbanization driving up land prices drastically, it makes good sense to leverage on technologies with smaller plant footprints, like the membrane bioreactor (MBR), as well as to adopt an innovative underground installation of the sewage treatment plant (STP), allowing a non-obnoxious co-existence with nearby residences. Being one of the largest underground STPs, the Jingxi Underground STP in Guangzhou is a 100,000 m3/day MBR-based treatment plant that was commissioned in 2010. The case study described here demonstrates the operational excellence of the STP by illustrating how an optimized internal recirculation for the A2O-MBR process reduced aeration energy demands by 18% while simultaneously improving Total Nitrogen (TN) removal by 24%. Furthermore, the success of a plant trial elucidating conditions that stabilized higher flux operations (to handle surges in water production demands) were also discussed.

Guangzhou city is the third largest city in China (after Beijing and Shanghai), with a population of approximately 14 million (Bureau 2014a, 2014b). With the city flourishing along and extending out from the Pearl River, water resources have been, and will continue to be, an important factor that places Guangzhou as one of China's leading commercial hubs (per capita GDP was USD$ 19,264 in 2013) (Bureau 2014a, 2014b). As the population grows, increasingly greater strains will be placed upon infrastructures and public services. Consequently, the demand for high quality water to support the industries and urban population is accompanied by an equally important, if not greater, requirement on sewage treatment capabilities. In conjunction with the need to handle the increased sewage production accompanying the 2010 Asian Games (hosted by Guangzhou), the municipality had the impetus to develop one of the world's largest underground sewage treatment plants (STPs). Undertaken by CITIC Envirotech Ltd (CEL) (then United Envirotech Ltd.), the 100,000 m3/day STP is based on membrane bioreactor (MBR) technology, allowing the Jingxi Underground STP (JXUSTP) to stand at just 1.82 hectares and in close proximity to the existing light industries and residential areas. Such pre-existing circumstances necessitate advanced wastewater treatment technologies with smaller plant footprints to combat land constraints, such as the MBR technology (Bureau 2014a, 2014b). The MBR system is an improvement over the conventional activated sludge process (ASP) where biological wastewater treatment is combined with membrane technologies for solid-liquid separation. The advantages include:

  • (a)

    Smaller footprints,

  • (b)

    Better effluent quality, and

  • (c)

    Independent control of hydraulic retention time and solids retention time (Judd 2010).

Furthermore, by adopting an underground installation, unsightly architecture and unpleasant odours typical of STPs are eliminated. The synergy between smaller footprints from MBR technology and underground placement of the plant significantly reduces CAPEX (capital expenditures) for land investments as the phenomenon of rapid urbanization has driven up land prices in Guangzhou significantly. The project cost the Guangzhou municipality RMB 580 million and the project was completed in time (began in September 2009 and commissioned by August 2010) for the 2010 Asian Games. The challenges of operating a membrane-based STP in densely populated cities like Guangzhou involve the handling of sporadic spikes in wastewater inflows (above design flows) and trying to further reduce the operational costs of such large scale plants; both of which being pertinent concerns of operators. Henceforth, this paper aims to share some of the innovative protocols that have been successfully implemented to address the aforementioned issues.

Treatment train

A schematic detailing the treatment train design for the JXUSTP is presented in Figure 1.

Figure 1

Treatment train design for Jingxi Underground STP.

Figure 1

Treatment train design for Jingxi Underground STP.

Close modal

Influent qualities are characteristic of domestic sewage (Metcalf & Eddy 2003) and given the stringency placed on effluent suspended solids (SS), total nitrogen (TN) and total phosphorus (TP) levels (refer to Table 1), the A2O-MBR process was selected as a core component of the design.

Table 1

Influent wastewater quality and regulated effluent levels

BOD5CODSSNH4+-NTNTP
Influent (mg/L) 160 270 220 30 35 4.5 
Effluent regulation (mg/L)a 10 50 10 15 0.5 
BOD5CODSSNH4+-NTNTP
Influent (mg/L) 160 270 220 30 35 4.5 
Effluent regulation (mg/L)a 10 50 10 15 0.5 

aBased on China's Class 1A effluent water quality standards (refer to Table 3 for a simplified standard) (China 2007).

Energy-saving and performance optimization for A2O-MBR operations

While MBR systems provided good quality effluent, it is a double-edged sword because air scouring requirements for membrane fouling control (Yamamoto et al. 1989) resulted in under-utilized, high dissolved oxygen (DO) levels within the mixed liquor. As shown in Figure 2, where mixed liquors were originally recirculated into the anoxic tank from the membrane tank, great energy savings were achieved by recirculating into the aerobic tank instead, leveraging on excess DO to reduce overall aeration demands. In detail, the surplus DO can be consumed along the length of the oxic tanks by the aerobic and facultative bacteria to reduce aeration needs directly. Additionally, when nitrate-rich mixed liquors with lower oxygen levels are recirculated from the end of the oxic zones to the anoxic tanks, denitrification performances can be enhanced and exhibit higher TN removals.

Figure 2

A2O-MBR internal recirculation configurations.

Figure 2

A2O-MBR internal recirculation configurations.

Close modal

Protocol for stable operation at higher fluxes

Rapid growth and expansion of the Guangzhou city resulted in spikes of water demand that the original baseline flux of 15 litres per m2 per hour (LMH) was unable to cater sufficiently for. Hence in order to increase water production without impairing membrane longevity and operational stability, appropriate changes have to be made to the accompanying membrane maintenance cleaning (MC) protocols to counteract the accelerated membrane fouling at higher fluxes. 3 MC protocols with incremental intensities have been studied in a plant trial test and great successes were found at stabilizing MBR fouling for fluxes up to 25 LMH (temperature corrected to 25 °C).

Aeration energy reduction and denitrification enhancements

Aeration demands and TN removals for conventional A2O and A2O-MBR processes have been compared against an optimized version used at JXUSTP in Table 2. While MBRs consume more energy than conventional systems due to membrane scouring needs, an optimized internal recirculation path (as shown in Figure 2) effected an 18% reduction in blower energy demands (from 0.22 kWh/m3 to 0.18 kWh/m3). An accompanying benefit of lowered oxygen levels within the mixed liquor is a more conducive environment for the denitrifying bacterial consortium to consume nitrates as the terminal electron acceptor (TEA) of their anoxic respiration (Oh & Silverstein 1999), where higher oxygen levels can be inhibitory to nitrate reduction by repressing the nitrate reduction enzyme (Metcalf & Eddy 2003). This resulted in an approximate 10% improvement in TN removals (refer to Table 2) and in comparison to a conventional A2O-based STP in the Guangzhou municipality, JXUSTP has even exhibited better TN removals (73% compared to 71% for conventional A2O). This is most likely a result of the longer sludge retention time (SRT) applied in the A2O-MBR system (SRT of 20–30 d as compared to SRT of 15–20 d for conventional A2O), which allowed for a larger denitrifying microbial population.

Table 2

comparison of aeration demands and nitrogen removals for different A2O configurations

Operational parametersCapacity (m³/d)Oxic Tank Aeration Demands (Nm3/m3)Membrane Aeration Demands (Nm3/m3)Blower Energy Consumption (KWh/m³)Influent TN (mg/L)Effluent TN (mg/L)
Conventional A2O (Guangzhou) HRT: 10–11 h
MLSS: 3,500 mg/L
SRT: 15–20 d 
560,000 0.09 31 9 (71% removal) 
Conventional A2O-MBR HRT: 7 h
MLSS: 7,000 mg/L
SRT: 20–30 d 
100,000 0.22 31 11 (64% removal) 
Optimized A2O-MBR HRT: 7 h
MLSS: 7,000 mg/L
SRT: 20–30 d 
100,000 0.8 0.18 31 8.3 (73% removal) 
Operational parametersCapacity (m³/d)Oxic Tank Aeration Demands (Nm3/m3)Membrane Aeration Demands (Nm3/m3)Blower Energy Consumption (KWh/m³)Influent TN (mg/L)Effluent TN (mg/L)
Conventional A2O (Guangzhou) HRT: 10–11 h
MLSS: 3,500 mg/L
SRT: 15–20 d 
560,000 0.09 31 9 (71% removal) 
Conventional A2O-MBR HRT: 7 h
MLSS: 7,000 mg/L
SRT: 20–30 d 
100,000 0.22 31 11 (64% removal) 
Optimized A2O-MBR HRT: 7 h
MLSS: 7,000 mg/L
SRT: 20–30 d 
100,000 0.8 0.18 31 8.3 (73% removal) 

HRT: hydraulic retention time; MLSS: mixed liquor suspended solids; SRT: sludge retention time

Table 3

Chinese discharge standards stipulated (extracted from GB18918-2002)

ParametersUnitsClass 1 Standards
Class 2 StandardsClass 3 standards
Class 1AClass 1B
COD mg/L 50 60 100 120 
BOD mg/L 10 20 30 60 
Suspended solids mg/L 10 20 30 50 
Total Nitrogen mg/L 15 20 – – 
Ammonia mg/L 25 – 
Total Phosphorus mg/L 0.5 
pH – 6–9 
CFU  103 104 104 – 
ParametersUnitsClass 1 Standards
Class 2 StandardsClass 3 standards
Class 1AClass 1B
COD mg/L 50 60 100 120 
BOD mg/L 10 20 30 60 
Suspended solids mg/L 10 20 30 50 
Total Nitrogen mg/L 15 20 – – 
Ammonia mg/L 25 – 
Total Phosphorus mg/L 0.5 
pH – 6–9 
CFU  103 104 104 – 

COD: chemical oxygen demand; BOD: biological oxygen demand; CFU: colony forming unit

Stable high flux operations

With advanced membrane technologies forming the pillar of excellence, JXUSTP has been demonstrating stable operation for more than 5 years, meeting discharge requirements consistently. As membranes deliver higher effluent qualities through superior solid-liquid separation, high membrane qualities will help to amplify this advantage (manifesting as system longevity and fidelity). Characteristics of the membrane system that made the reliability possible are tabulated in Table 4.

Table 4

Summary of membrane system parameters

ParametersUnitsValueComments
Membrane brand – Memstar  
Model – SMM1520 Double-deck skid 
Pore size μ<0.1  
Total membrane area m2 352,000  
Average flux LMH 15  
SADm Nm3/h·m2 0.08  
SADp Nm3/m3 4.3  
Years of operation years 5.5 As of 2016 
Membrane replacement – No  
ParametersUnitsValueComments
Membrane brand – Memstar  
Model – SMM1520 Double-deck skid 
Pore size μ<0.1  
Total membrane area m2 352,000  
Average flux LMH 15  
SADm Nm3/h·m2 0.08  
SADp Nm3/m3 4.3  
Years of operation years 5.5 As of 2016 
Membrane replacement – No  

SADm: specific aeration demand - membrane area; SADp: specific aeration demand - permeate volume

With reliability forming a solid foundation, it is feasible for the JXUSTP to operate at higher fluxes to meet water demand surges of the growing Guangzhou municipality. In order to counteract accelerated fouling under higher membrane fluxes, stronger MC protocols were applied to mitigate membrane fouling rates, measured as daily increments in transmembrane pressure (TMP ). Evidently, as can be seen in Figure 3 and Table 5, fouling rates generally increased as fluxes increased from 15 to 25 LMH, but the application of stronger MC protocols allowed for a controlled TMP rise even at higher fluxes.

Figure 3

Graphical representation of the TMP response curve for different fluxes and MC protocols.

Figure 3

Graphical representation of the TMP response curve for different fluxes and MC protocols.

Close modal
Table 5

Tabulated parameters for fouling rates and MC protocols at different fluxes.

Flux (25 °C)MC ProtocolCleaning intensity (% wrt Protocol 1)Fouling rate (kPa/day)Costs of MC protocol (¥/ton)
15 LMH 100% 0.0587 0.006 
20 LMH 100% 0.169 0.004 
20 LMH 200% 0.0402 0.008 
25 LMH 200% 0.194 0.008 
25 LMH 400% 0.0912 0.016 
Flux (25 °C)MC ProtocolCleaning intensity (% wrt Protocol 1)Fouling rate (kPa/day)Costs of MC protocol (¥/ton)
15 LMH 100% 0.0587 0.006 
20 LMH 100% 0.169 0.004 
20 LMH 200% 0.0402 0.008 
25 LMH 200% 0.194 0.008 
25 LMH 400% 0.0912 0.016 

However, as costs escalated with MC intensities (as shown on Table 5, MC costs have risen more than 1.5 times as protocols have intensified to enable stable performances when fluxes were raised to 25 LMH from 15 LMH), there exists a diminishing return where MC intensification is no longer economically viable and high-flux operational stability should be tackled from other perspectives – pre-treatment, aeration optimization, membrane performance improvements are some of the possible ways forward. Additionally, the MC protocols can be applied to future projects so that stable operations for higher fluxes can be achieved and reduce CAPEX for membrane systems.

As worldwide urbanization progresses into overdrive, land pricing will become increasingly significant in the CAPEX component for future STP projects. The innovative configurations and operational methodologies that the JXUSTP has illustrated should gain an increased amount of attention to help tackle wastewater treatment in tight urban spaces at lowered energy consumptions, improved nutrient removals and stable MBR operations.

Bureau
G. S.
2014a
Guangzhou Economic and Social Development Statistics Bulletin 2014
.
Guangzhou Daily, Guangzhou Statistics Bureau: A10
.
Bureau
G. S.
2014b
广州信息手册 Guangzhou Statistics Report. Guangzhou Statistics Bureau, Guangzhou Statistics Bureau
.
China M. o. E. P. o. t. P. s. R. o.
2007
GB18918—2002: Discharge standard of pollutants for municipal wastewater treatment plant. M. o. E. P. o. t. P. s. R. o. China, Ministry of Environmental Protection of the People's Republic of China
.
Judd
S.
2010
The MBR Book: Principles and Applications of Membrane Bioreactors for Water and Wastewater Treatment
.
Oxford: Elsevier
.
Metcalf
I.
&
Eddy
H.
2003
Wastewater Engineering; Treatment and Reuse
.
New York
:
McGraw-Hill
.
Oh
J.
&
Silverstein
J.
1999
Oxygen inhibition of activated sludge denitrification
.
Water Research
33
(
8
),
1925
1937
.
Yamamoto
K.
,
Hiasa
M.
,
Mahmood
T.
&
Matsuo
T.
1989
Direct solid-liquid separation using hollow fiber membrane in an activated sludge aeration tank
.
Water Science & Technology
21
(
4–5
),
43
54
.