Comparison of Two-stage AO and Three-stage AO Processes: An Engineering Perspective
Currently, the majority of wastewater treatment plants (WWTPs) in China adopt activated sludge-based processes for wastewater treatment. Among these, nearly half utilize the Anoxic-Oxic (AO) process. The AO process offers advantages such as stable operation and low cost. However, its total nitrogen (TN) removal efficiency, typically ranging from 60% to 80%, is constrained by internal recycle ratios. With increasingly stringent national requirements for nitrogen removal, conventional single-stage AO processes often struggle to meet the demands for TN treatment. Multi-stage AO processes have thus emerged. By connecting two or more AO stages in series, the nitrate produced in the preceding aerobic stage provides the substrate for denitrification in the subsequent anoxic stage. This achieves the goal of reducing the internal recycle ratio while enhancing overall TN removal. However, excessive stages can also increase operational complexity. Consequently, the most commonly applied configurations in China are currently the two-stage and three-stage AO processes. This paper presents a comparative analysis of two-stage and three-stage AO processes using a WWTP in Southern China as a case study, aiming to provide a reference for the selection of technical routes in similar projects.
1 Project Overview
A WWTP in Southern China covers a total area of 8 hectares. Its original design capacity was 90,000 m³/d, with effluent quality required to meet both the Grade A standard of the "Discharge Standard of Pollutants for Municipal Wastewater Treatment Plants" (GB 18918-2002) and the "Water Pollutant Discharge Limits" of Guangdong Province (DB 44/26-2001) (hereinafter referred to as "Quasi-Class V"). The plant was operating at full capacity. According to relevant planning, an expansion was required. The future effluent standards, based on the current status, needed to consider a long-term requirement of TN ≤ 10 mg/L. Comprehensively considering the site's actual conditions, the civil construction scale for this expansion was set at 70,000 m³/d. The plant would operate at 50,000 m³/d in the near term and reach the 70,000 m³/d scale in the long term, bringing the plant's total treatment capacity to 160,000 m³/d. The designed influent and effluent water quality are shown in Table 1.
Due to site constraints, the preliminary plan for the expansion adopted the process route of "Multi-stage AO + Peripheral-In Peripheral-Out Rectangular Sedimentation Tank + High-Efficiency Sedimentation Tank + Fiber Plate-and-Frame Filter". The civil structures of all major units were constructed for the 70,000 m³/d scale, while equipment was installed for the 50,000 m³/d capacity. The biological tank would employ a multi-stage AO process in the near term. In the long term, adding suspended carriers would create a hybrid biofilm-activated sludge process to meet the 40% capacity expansion demand. For this design, hydraulic conditions were considered for the 70,000 m³/d scale, while biological treatment was designed for the 50,000 m³/d scale. As this project intended to adopt a multi-stage AO process, a comparison between two-stage and three-stage AO was conducted.
2 Comparison of Two-Stage and Three-Stage AO Processes
2.1 Process Flow
The core principle of the multi-stage AO process is to utilize nitrate produced in the preceding aerobic stage for denitrification in the subsequent anoxic stage, thereby reducing the internal recycle ratio. Theoretically, more stages lead to better TN removal, but control becomes more complex. In engineering practice, two-stage and three-stage AO are predominant. Their process flows are shown in Figure 1. For a two-stage AO, internal recycle is typically designed within the first AO stage. For a three-stage AO, internal recycle is generally not employed. WWTPs in Beijing using the two-stage AO process include Qinghe (400,000 m³/d), Xiaohongmen (500,000 m³/d), Gao'antun (400,000 m³/d), Dingfuzhuang (200,000 m³/d), and Huaifang (600,000 m³/d). This process offers advantages such as simple equipment, low operation and maintenance costs, strong resistance to shock loads, and high compatibility with other processes, facilitating future upgrades to meet higher effluent standards. Theoretically, a three-stage AO in series can eliminate the need for internal recycle equipment, allow more rational allocation of carbon sources, and reduce investment and operating costs. This process is primarily applied in scenarios with sufficient carbon sources and high demands for nitrogen removal. Typical cases include the Qujing WWTP in Yunnan (80,000 m³/d), the Ninghe District Urban WWTP in Tianjin (90,000 m³/d), the Zhangguizhuang WWTP in Tianjin (200,000 m³/d), and the Daoxianghu Reclamation Plant in Beijing (80,000 m³/d).
2.2 Process Comparison
Considering that no additional land is available for future upgrades at this site, and that some new local projects are already implementing an effluent TN standard of ≤10 mg/L, the process comparison considered a biological tank effluent TN of ≤10 mg/L to accommodate the possibility of further stringent effluent requirements in the future. Other indicators adhered to the design effluent quality. Based on the layout, for the near-term scale of 50,000 m³/d, the maximum Hydraulic Retention Time (HRT) for the biological tank was 18 hours. Combining the project's actual conditions, BioWin simulation results, and the convenience of coupling with suspended carriers, a comparison between two-stage and three-stage AO processes was conducted.
2.2.1 BioWin Simulation
An initial HRT of 18 hours was set and gradually reduced. The minimum HRT achieving the effluent TN requirement was 14 hours. For the two-stage AO, influent distribution points were the anaerobic zone, the first-stage anoxic zone, and the second-stage anoxic zone. For the three-stage AO, influent points were the anaerobic zone, the second-stage anoxic zone, and the third-stage anoxic zone.
① Study with Fixed Influent Distribution Ratio
Setting the influent distribution ratio at 4:3:3 for both, simulations compared three schemes: two-stage AO (recycle ratio 200%), three-stage AO with a total recycle ratio of 200% (100% recycle within the first AO stage + 100% recycle from the third Oxic zone to the first Anoxic zone), and three-stage AO with a recycle ratio of 100% (recycle only within the first AO stage). The simulation flows are shown in Figure 2.
Table 2 shows the simulation results for a fixed influent ratio at HRT=14 h.
From Table 2, it can be seen that for both two-stage and three-stage AO, it is recommended to set up internal recycle in the first AO stage to maximize denitrification in the first anoxic zone by utilizing the carbon source in the raw influent. For the three-stage AO, setting up internal recycle from the end of the third stage to the first anoxic zone slightly improved TN and TP removal, but organic matter removal efficiency decreased. This is a speculation attributed to the increased overall flow in the biological tank due to the recycle, which carried dissolved oxygen into the anoxic zone, affecting the anoxic environment. Additionally, the actual HRT in each zone shortened, and the transition between operational conditions accelerated, leading to reduced efficiency. For influent characteristics like those in this project in Southern China, where TN concentration is not very high, the two-stage AO can fully meet the effluent requirements, showing no distinct advantage for the three-stage AO. For scenarios with high COD and high TN influent, the three-stage AO might be more suitable.
② Study on Adjusting Influent Distribution Ratios
Both two-stage and three-stage AO were set with a 100% internal recycle ratio in the first AO stage. Studies were conducted on multi-point influent distribution ratios (1:0:0, 3:7:0, 2:4:4). Here, 1:0:0 means all influent enters at the very front; 3:7:0 for the three-stage AO means influent is distributed only to the anaerobic zone and the second AO stage. Simulation results for adjusted distribution ratios are shown in Table 3.
From Table 3, it can be seen that the distribution ratio has a slight impact on effluent quality. The general trend is that as the proportion of influent distributed to later stages increases, effluent TN, NH₃-N, and TP concentrations rise, and aeration demand also gradually increases. When the influent ratio was 3:7:0, the three-stage AO showed slightly better TN removal and a slightly lower air-to-water ratio than the two-stage AO. However, in actual operation, this difference is generally negligible. Moreover, increasing the proportion of influent to later stages, while beneficial for carbon source utilization in denitrification, inevitably increases the load on biochemical reactions due to the input of NH₃-N, organic matter, and TP. Therefore, it is recommended to retain the multi-point influent configuration and make phased adjustments based on actual water quality during operation. It is worth noting that although the three-stage AO showed better TN removal than the two-stage AO at a 2:4:4 influent ratio, as the influent to later stages increased, effluent NH₃-N showed a rising trend, at which point NH₃-N could no longer meet the effluent standard.
③ Treatment Performance of Two-Stage and Three-Stage AO
A three-stage AO configuration was simulated with HRT=14 h, equal volume ratios for each stage (1:1:1), 100% internal recycle set in the first AO stage, and an influent ratio of 4:3:3, under two conditions: with 100% recycle and with recycle closed. A two-stage AO configuration was simulated with HRT=14 h, 100% internal recycle set, and an influent ratio of 4:3:3. Results showed that the two-stage AO achieved the optimal effluent TN at 6.29 mg/L; the three-stage AO with 100% internal recycle at the front achieved the next best at 7.51 mg/L; the three-stage AO without internal recycle performed worse at 8.52 mg/L. All three scenarios could meet the effluent verification requirement (TN ≤ 10 mg/L).
Table 4 shows the design parameter comparison between two-stage and three-stage AO. It can be seen that for both processes, the HRT required to achieve the effluent TN requirement is less than 18 hours. The main differences between the two processes are as follows:
a. Theoretically, the three-stage AO has a higher upper limit; i.e., if operated properly, both investment and operating costs can be lower. The two-stage AO has fewer equipment items and stages, resulting in lower equipment costs and lower operational management difficulty.
b. For this specific project, since the long term was considered and the tank volume was designed for an 18-hour HRT, the civil investment would be identical whether adopting the two-stage or three-stage AO. The equipment cost for the three-stage AO is higher. Therefore, from an investment perspective, adopting the two-stage AO is more economical.
c. Regarding operating costs, the three-stage AO could save approximately 0.002 CNY/m³ by eliminating the 100% mixed liquor recycle energy cost. Considering the potential decrease in carbon source utilization efficiency in actual operation due to alternating anoxic/oxic conditions in the three-stage AO, the actual difference in operating costs would likely be even smaller.
2.2.2 Analysis of Long-term Suspended Carrier Scenario
Due to the unique requirements of this project, the biological tank needed to consider the feasibility and convenience of the long-term capacity expansion plan, i.e., the impact of adding suspended carriers.
The core of the MBBR process is to increase the biomass in the reactor by adding suspended carriers. These can be added to aerobic, anoxic, or anaerobic tanks. However, considering carrier fluidization, adding them to anaerobic or anoxic tanks would significantly increase mixing power requirements. Therefore, addition to aerobic tanks is preferentially recommended. The volume for anaerobic/anoxic zones can be supplemented by partitioning from the aerobic zone, while the deficiency in aerobic volume is compensated for by the added carriers. In other words, the insufficient aerobic volume is borne by the increased surface area of the suspended carriers, which is calculated based on pollutant load conversion to determine the required carrier quantity, controlling a certain fill ratio to obtain the added volume.
Based on calculations, if adopting the two-stage AO process and adding all suspended carriers to the first-stage aerobic zone in the long term, the required MBBR carrier surface area would be 2,597,708 m², costing 12.99 million CNY. Other related fixed equipment costs (including MBBR fluidization systems, dedicated mixers, screening systems, and intelligent control systems) would be 6.15 million CNY. If adopting the three-stage AO process, due to more dispersed zones, the MBBR zone would need to be divided into 2 sections (first-stage and second-stage aerobic zones). Consequently, the cost for installing corresponding MBBR fixed equipment (excluding the carriers themselves) would increase slightly to 7.77 million CNY, while the carrier cost remains the same. This means adopting the three-stage AO would increase future retrofit investment by 1.62 million CNY and also increase retrofit complexity. Furthermore, the screening system is the area most prone to issues after carrier addition. The three-stage AO adds an extra section of screens, increasing operational difficulty.
From the above comparison, due to the excessive partitioning in the three-stage AO, with each partition having a similar volume, its retrofit difficulty is higher than that of the two-stage AO. Construction, operational complexity, and the addition of screening equipment also result in higher investment than the two-stage AO. Therefore, adopting the two-stage AO is more conducive to future coupling with suspended carriers.
2.3 Comparison Result
Based on the above analysis, both two-stage and three-stage AO processes can achieve the target of effluent TN ≤ 10 mg/L. Under the boundary conditions of this project-limited space, the need to maximize near-term tank volume, and the long-term plan to add suspended carriers-the two-stage AO holds advantages in terms of near-term investment and equipment management/maintenance convenience. It also offers higher compatibility for future retrofit with suspended carriers, resulting in lower overall investment and reduced retrofit and operational difficulty. Therefore, after comprehensive consideration, the two-stage AO process was recommended for this design.
3 Operational Performance
The total estimated investment for this project is 304.5721 million CNY, with construction costs of 243.6019 million CNY, translating to a unit construction cost of 3,480.03 CNY/m³. The treatment cost is 1.95 CNY/m³, and the operating cost is 1.20 CNY/m³.
For this project, the biological tank has a total HRT of 18 hours (comprising: anaerobic zone 2 h, first-stage anoxic zone 3.5 h, first-stage aerobic zone 7.5 h, degas zone 0.5 h, second-stage anoxic zone 2.5 h, second-stage aerobic zone 2 h), with an effective water depth of 8.6 m. Adjustable sectional water intake is implemented, allowing adjustments in the influent distribution ratio in 20% increments as needed. In actual operation, the Mixed Liquor Suspended Solids (MLSS) concentration in the biological tank ranges from 3,500 to 4,000 mg/L, the sludge return ratio ranges from 40% to 100%, and the mixed liquor internal recycle ratio ranges from 100% to 200%. Actual influent and effluent quality are shown in Table 5, which aligns basically consistent with the simulation results.
4 Conclusion
Using a WWTP in Southern China as a case study, a technical and economic comparison between two-stage and three-stage AO processes was conducted with the aid of BioWin simulation. The two-stage AO, with fewer equipment items and stages, lower equipment costs, and lower operational management difficulty, is more suitable for conditions in Southern China where influent TN is not very high. For the three-stage AO, setting up internal recycle from the end of the third stage to the first anoxic zone negatively affected TN removal efficiency, increased operational management difficulty, and raised investment costs. The design simultaneously meets the near-term treatment requirements of 50,000 m³/d and TN ≤ 10 mg/L, while the long-term scale of 70,000 m³/d can be achieved by coupling with suspended carriers. Actual operational results are largely consistent with the BioWin simulation results, with an average effluent TN of 6.86 mg/L, meeting the design requirements.