Modification of a plant-scale semi-centralized wastewater treatment system to enhance nitrogen and phosphorus removal from black water

Rapid worldwide urbanization in recent years has been accompanied by the widespread construction of traditional centralized wastewater treatment systems to facilitate effective and reliable collection and treatment of municipal wastewater, through the extensive practical experience gained from the management of pipeline networks and wastewater treatment plants (Libralato et al. 2012; Tang et al. 2020). However, centralized wastewater treatment systems have certain disadvantages, including high operational costs, high energy consumption, and low resource recovery, limiting sustainable development of the wastewater industry. Consequently, these systems tend to be inefficient in recycling water and pollutant resources, and the system operates as a resource-consuming model (Wilderer & Schreff 2000; Bieker et al. 2010; Libralato et al. 2012; Edi et al. 2016).

With a view toward more efficient regional wastewater collection, treatment, and recycling, German scholars have proposed the concept of ‘semi-centralized wastewater treatment systems,’ which classifies domestic wastewater into two types based on the wastewater source, namely black water (from toilets) and gray water (from bathrooms, washing machines, and kitchens) (Huelgas et al. 2009; De Graaff et al. 2010; Hocaoglu et al. 2013; Lam et al. 2015). The separate collection of black and gray water enables their treatment in separate specified systems. Hence, compared with traditional centralized wastewater treatment systems, semi-centralized systems have higher water-reuse efficiency, greater flexibility, and lower capital investment and operational costs (Bieker et al. 2010) and are thus promising for urban application. In Germany, based on the DEUS21 research project concept, vacuum toilets and food waste shredders were installed in a residential area built in 2004, resulting in the generation of black water with high organic content that could be treated using an anaerobic process converting organic pollutants to biogas, yielding a nutrient-containing effluent that could be reused for agricultural irrigation (Kotz et al. 2006). In China, a semi-centralized wastewater treatment system has been installed for an approximate population equivalent of 12,000 in the vicinity of the International Horticultural Expo, Qingdao. This system collects 700 m³/day of gray water, 800 m³/day of black water, and 50 m³/day of sludge and kitchen waste for separate treatment, recycles treated water for toilet flushing landscape and agricultural irrigation, and recovers bioenergy for electricity production (approximately 200 kWh/day; Tolksdorf & Cornel 2017).

The current wastewater treatment technologies are primarily based on activated sludge processes, including anaerobic–anoxic–aerobic (A2O), sequencing batch reactor (SBR), and membrane bioreactor (MBR) processes (Zhang et al. 2020; Huang et al. 2023). Among them, the A2O process and its modifications are the most widely used in China, accounting for approximately 33% of the statistical data (CUWA 2020). In a conventional A2O process with alternating anaerobic, anoxic, and aerobic zones, the nitrogen and phosphorus are removed via a series of biological reactions. Nitrogen removal is achieved via nitrification-denitrification; i.e., ammonia is oxidized in the aerobic zone, and then the generated nitrate is recycled to the anoxic zone and reduced to nitrogen gas using a carbon source as an electron donor (Zhang 2015; Sedlak 2018). The phosphorus removal is realized by phosphate-accumulating organisms (PAOs), which take up volatile fatty acids (VFAs) and obtain the energy by cleaving phosphate molecules from the internal polyphosphate granule under anaerobic conditions, and excessively absorbs phosphate from wastewater and stores it in granules within the cells under aerobic conditions (Zhang 2015). However, certain amounts of carbon sources have been utilized in the anaerobic zone of the conventional A2O process, which may lead to less carbon sources available for denitrification in the anoxic zone, resulting in unsatisfactory nitrogen removal efficiency (Zhao et al. 2017). And the anaerobic environment in the anaerobic zone of the conventional A2O process may be destroyed by recirculating nitrate-containing return sludge to the anaerobic zone, thus reducing the phosphorus uptake potential of PAOs (Ye et al. 2018). By switching the sequence of anaerobic and anoxic zones in the conventional A2O process (i.e., anoxic–anaerobic–aerobic process or the so-called reversed A2O process), nitrogen removal efficiency can be improved with larger amounts of carbon sources in the influent available for anoxic denitrification preferentially (Hu et al. 2014; Wang 2015; Yin et al. 2021). And the anaerobic environment can be also maintained in the anaerobic zone of the reversed A2O process, which follows after the anoxic zone. Consequently, it has been confirmed that the reversed A2O process is effective in improving nitrogen and phosphorus removal by reducing the disadvantages inherent in the conventional A2O process, such as uncoordinated sludge age for nitrogen and phosphorus removal and competition between denitrifying bacteria and PAOs for substrates (Ye et al. 2014). For example, Miao et al. (2004) found that with a 0.20–0.35 kg COD/kg MLSS·d organic loading rate in the influent wastewater, the COD, -N, and TP removal efficiencies of the reversed A2O process could reach more than 90% in a municipal wastewater treatment plant in Changzhou, China (Miao et al. 2004). In the semi-centralized wastewater treatment system in the International Horticultural Expo (Qingdao, China), the reversed A2O process followed by a post-anoxic zone and a MBR has been applied for black water treatment. MBR technology combines membrane separation and biological treatment, which can maintain high MLSS concentration in the biological system and avoid activated sludge loss in case of sludge expansion, thus improving effluent quality (Zhang et al. 2009; Song et al. 2010; Falahti-Marvast & Karimi-Jashni 2015; Nam et al. 2021). Even though MBR has high energy consumption and susceptibility to membrane contamination, MBR is generally considered to be an effective technology due to its advantages of short hydraulic retention time (HRT), simple operation, and low sludge production (Nam et al. 2021). However, considering the low C/N ratio of black water, it remains difficult to effectively remove nitrogen and phosphorus (removal efficiencies 79.4 and 28.8%, respectively). And the prolongation of HRT may increase the volume of tanks and thus increase the investment and operation cost of the system. Consequently, further study is required to optimize the system by developing efficient and economical operating methods. However, few experimental studies have focused on the modification of black water treatment system, and there remains a lack of data regarding the practical application of plant-scale systems.

Given the aforementioned considerations, in this present study, we assessed a series of operational strategies in a plant-scale modified AS-MBR system designed to enhance the removal of organic and nutrient pollutants from black water, including optimization of the coagulant dosage for chemical-enhanced primary treatment, and evaluated different operational strategies, such as adjusting dissolved oxygen (DO) levels in the aerobic zone, shifting the aerobic zone to the anaerobic zone, and determining the appropriate site for carbon source supplementation.

In this study, black water was treated to remove organics and nutrients. The black water to be treated was derived from a residential area near the International Horticultural Expo Qingdao, China, which was collected by a separate municipal network and delivered to a plant-scale black water treatment system. The characteristics of the two types of substrates used in the batch tests are shown in Table 1. The influent of the primary settling tank of the plant-scale black water system was collected as the substrate for pretreatment tests, and the chemically pre-treated wastewater was used for further enhancement of nutrient removal. During chemical pretreatment, a certain amount of organic material is removed from the wastewater, whereas most of the nitrogen remains, resulting in a low SCOD/TN ratio of 3–4.

Table 1

The characteristics of wastewater

Parameter .	Influent of the primary setting tank .		Influent of the biochemical system .
	Range .	Average .	Range .	Average .
TCOD	570–886	717	476–665	515
SCOD	280–424	329	347–423	375
-N	64–102	85	76–108	87.4
-N	0.7–2.8	1.9	0.9–3.2	2.41
TN	94–111	104	93–128	107.2
-P	4.2–8.2	6.1	6.6–14.6	9.6
TP	6.3–11.3	8.0	7.8–15.4	11.5
SS	255–407	340	195–387	343
pH	7.1–8.2	7.5	7.2–8.3	7.5

Parameter .	Influent of the primary setting tank .		Influent of the biochemical system .
	Range .	Average .	Range .	Average .
TCOD	570–886	717	476–665	515
SCOD	280–424	329	347–423	375
-N	64–102	85	76–108	87.4
-N	0.7–2.8	1.9	0.9–3.2	2.41
TN	94–111	104	93–128	107.2
-P	4.2–8.2	6.1	6.6–14.6	9.6
TP	6.3–11.3	8.0	7.8–15.4	11.5
SS	255–407	340	195–387	343
pH	7.1–8.2	7.5	7.2–8.3	7.5

The units of the parameters are mg/L, except pH.

View Large

Laboratory-scale tests were conducted in cylindrical batch reactors with an effective volume of 5 L. The reactors could be operated independently or in series to simulate the operation of the settling and biochemical tanks by controlling stirring and aeration devices.

To optimize pollutant removal, we performed two series of tests in which we modified the pretreatment and biochemical treatment processes (Table 2).

To optimize the coagulant dosage of chemical pretreatment, different amounts of chemical coagulant (FeCl₃; 0–50 mg Fe³⁺/L) were added to 5 L of black water in each laboratory-scale batch reactor, which was operated as primary settling tanks (Test 1.1). The initial pH was adjusted to 7.5. After mixing evenly, the contents were allowed to settle for 50 min; thereafter, the supernatants in the reactors were collected for analysis.

According to the results obtained in batch tests, the established optimal dosage of coagulant (FeCl₃) was introduced into the inlet weir of the plant-scale primary settling tank for 30 d (Test 1.2). The coagulant was first dissolved in a 0.5 m³ of tank to produce a high concentration of solution, and then, it was pumped to the inlet weir of the primary settling tank by a dosing pump with a maximum flow rate of 32 L/h. Then, the coagulant solution is mixed with the influent quickly by the hydraulic flow. Twenty-four-hour-mixed samples of the effluent were collected for analyses to compare pollutant removal efficiencies of the conventional primary and chemical-enhanced primary treatments.

To assess nutrient removal performance using the chemical-enhanced primary treatment, seven laboratory-scale serial reactors were operated as non-aeration or aeration zones to simulate the plant-scale biochemical processes (Figure 1). The main operational parameters, such as HRT and recirculation ratio, were controlled using the same procedures as those for the plant-scale system. To further confirm the nutrient removal efficiency with low energy, in addition to reducing the internal mixed liquid recirculation ratio from 400 to 300%, two operating strategies were proposed: (1) reducing the DO level in aerobic zones O3, O4, and O5 from 1.0–1.5 to 0.7 mg/L, and (2) converting aerobic zone O3 to anaerobic zone A3 by regulating the aeration device (Test 2). The chemically pre-treated effluent was used as the influent, and the liquid samples from each reactor were collected and analyzed to examine nitrogen and phosphorus transformation.

In Test 3.1, sodium acetate (50 mg COD/L) was introduced to pre-anoxic zone A1 or post-anoxic zone A6 in the plant-scale system as a carbon source for denitrification. To compare the system performance with/without the addition of a carbon source to the pre/post-anoxic zone, we continuously monitored the concentrations of nitrogen and phosphorus in the biochemical system's influents and effluents.

In order to examine the conversions of nitrogen and phosphorus in the biochemical reaction system when a carbon source was added to the pre/post-anoxic zones, laboratory-scale reactors were operated as SBRs, with sequential non-aeration (3.5 h), aeration (2.3 h), non-aeration (1.2 h), and aeration (1 h) (Test 3.2). To simulate anoxic pre-denitrification or post-denitrification processes, sodium acetate (50 mg COD/L) was added to the batch reactor at 0 or 5.83 h, respectively.

Conventional wastewater parameters such as SS, TCOD, SCOD, TN, -N, -N, TP, -P, and sulfide contents were determined according to the standard methods (APHA 2005). DO and pH were measured using a portable DO and pH analyzer (HACH hq40d, USA), respectively. All the parameters of the samples are detected at least three times.

The statistical analyses were conducted using statistical software Origin 9.0. Polynomial regression (Test 1.1) and linear regression (Tests 1.2 and 3.1) were performed using experimental data of concentrations of effluent pollutants or removal efficiencies with the confidence level (p = 0.05). And the goodness-of-fit (R2) was used as model determination criteria.

The shift of aerobic zone O3 to anaerobic zone A3 not only lengthened the pre-denitrification phase but also prolonged the anaerobic phosphorus release time. However, high concentrations of nitrate recirculated with the aerated liquid from the aerobic zone O5 of the system slightly inhibited the release of phosphorus, thereby resulting in the release of only 2.63 mg P/L compared with the 3.4 mg P/L released without prolonged pre-denitrification (Hai et al. 2015). Under such conditions, an average effluent phosphate concentration of 1.61 mg/L could be achieved, which is lower than that prior to regulating the aeration device in zone 3; this result might be attributed to a more stable influent phosphorus concentration following pretreatment in the primary process.

As mentioned in section 2.1, more than 70% of phosphorus was removed following chemical-enhanced primary treatment, thereby significantly reducing a load of biological phosphorus removal, resulting in a high average system TP removal efficiency of 93.6% (Figure 5(c)). Furthermore, we achieved slightly higher and more stable phosphorus removal efficiencies (at approximately 96%) when a carbon source was added in post-anoxic zone A6 compared to when it was not. We also observed that after adding a carbon source to the pre-anoxic zone A1, the TP removal efficiency fluctuated and then stabilized in 5 days.

Upon initial addition of sodium acetate (50 mg/L) as a carbon source (0 min), we observed that nitrate was reduced at a conversion rate of 10.76 mg N/(L·h) during the first non-aeration phase; this conversion was more rapid than that without carbon source addition (Figure 6(b)). However, residual organic matter was mostly oxidized during the aeration phase, thereby reserving limited carbon for denitrification during the second non-aeration phase, resulting in no enhancement in nitrogen removal efficiency. These findings also indicate that the addition of a carbon source to the pre-anoxic zone is wasteful, as it only accelerates the reduction rate of nitrate in the influent, whereas a certain amount of ammonia remains in the system, which would subsequently be converted to nitrate, necessitating further removal in the post-anoxic zone. However, some amount of this nitrate would be reduced in the post-anoxic zone, as most of the additional carbon source would be consumed prior to this stage (Zhang et al. 2019).

The addition of the carbon source at the beginning of the second non-aeration phase (350 min) (Figure 6(c)) resulted in a difference in the nitrogen removal performance. The supply of carbon source significantly improved the nitrogen removal efficiency to 70.5%, with only a residual amount (11.96 mg TN/L) remaining in the system effluent (Figure 6(c)). The ratio of the amount of supplemented carbon source to reduced nitrate was approximately 3.8, further revealing that a carbon source added to the post-anoxic zone could be fully utilized by denitrification. Consistently, the utilization of an additional carbon source was more efficient for the enhancement of denitrification in the post-anoxic stage than in the pre-anoxic stage, reducing effluent nitrate concentrations from 8.95 ± 0.19 to 3.60 ± 0.15 mg/L (James & Vijayanandan 2022).

The effect of carbon source dosage on phosphorus transformation primarily enhanced phosphate release (Isaacs et al. 1993). As shown in Figure 6(b), when the carbon source was added during the pre-anoxic phase, 5.4 mg P/L was released, which was 4.5-fold higher than the amount released without carbon source addition, whereas the uptake of excess phosphate (7.1 mg P/L) that occurred in the aeration phase led to a reduction in phosphate concentrations, consistent with the findings of Dai et al. (2021). Contrastingly, carbon source addition during the post-anoxic phase did not promote a significant phosphate release (Figure 6(c)), which may be due to the poor competitiveness of PAOs for carbon source with denitrifying bacteria and the inhibition caused by the residual nitrate, thereby confirming the prior utilization of the added carbon source during denitrification.

In this study, we established that appropriate modification to a plant-scale AS-MBR system can enhance nitrogen and phosphorus removal from black wastewater. Our main conclusions are as follows.

• Phosphorus removal could be significantly promoted by chemical-enhanced primary treatment, and nitrogen removal was primarily enhanced by modifying the reversed A2O system.

• Chemical-enhanced primary treatment of black water with a 21 mg Fe³⁺/L dose could increase organic and phosphorus removal from 23.8 to 51.1% and 15.1 to 74.1%, respectively. However, this addition had little effect on the nitrogen removal efficiency.

• A reduction in the aeration intensity of the aerobic reactors is a potential energy-saving method, and under such conditions, simultaneous nitrification and denitrification can occur. However, a prolongation of the pre-denitrification phase does not significantly enhance the removal of nitrogen and phosphorus.

• Compared with the system without carbon source addition, the addition of a carbon source in the post-anoxic zone significantly enhanced the nitrogen removal efficiency from 41.5 to 70.5%, along with an increased phosphorus removal performance. However, adding a carbon source in the pre-anoxic zone failed to enhance nitrogen removal and resulted in fluctuations in TP removal efficiency. The release and uptake of phosphate were confirmed to be influenced by carbon source addition in the pre-anoxic zone, whereas addition in the post-anoxic zone had no effect.

This study was supported financially by the National Key Research and Development Plan of China (2020YFD1100303).

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.