3.1 Predicted hyper-quantity functional performance of the process
Considering more stringent effluent limits, it is essential to emulate the performance of the process to recommend optimal hyper-quantity operations. Thus, in this section the performance of the process under eight different conditions in winter and summer is described in Fig. 3.
The average effluent COD concentration in simulated scenarios 1 and 2 (Fig. 3 (a)) were found to be 13.50 mg/L, 13.65 mg/L, 13.45 mg/L and 13.64 mg/L respectively, with corresponding SDs were 2.56, 2.56, 2.59 and 2.58. Notably, the effluent COD concentration under scenarios 1 and 2 showed significant overlap when the water quantity exceeded 1.1 and 1.2 times than that of the design capability during winter, thereby hardly influencing on effluent COD concentration. In addition, the average effluent concentration of NH3-N in simulated scenarios 1 and 2 (Fig. 3 (b)) were observed to be 0.43mg/L, 0.60 mg/L, 0.34 mg/L and 0.58 mg/L, respectively. Importantly, the NH3-N concentration was found to achieve the discharge limit even under 1.2 times the hyper-quantity (4.8×104 t/d). Conversely, the emulated effluent quality in scenarios 2 and 4 following the addition of P removal agents (mainly Fe ions) showed higher fluctuations. The effluent TP concentration was noticeably higher compared to scenarios without the dosing phosphorus removal agent, indicating the negative effect of P removal agent on the NH3-N effluent concentration (Fig. 3 (b)). Furthermore, the average effluent TN concentration in simulated process were observed to be 11.20mg/L, 11.34 mg/L, 11.76 mg/L and 11.94 mg/L, respectively. The TN concentration remained below the limit of 15 mg/L for effluent concentration, when the influent quality was increased to 1.1 and 1.2 times the design scale. However, the risk of exceeding the discharge limits existed to a certain extent. Therefore, there was still a certain level of risk of exceeding the discharge limits, and special attention should begiven to the effluent concentration of NH3-N and TN during hyper-quantity operation in winter (Fig. 3 (c)). The average simulated effluent TP concentration were observed to be 0.41 mg/L, 0.0066 mg/L, 0.75 mg/L and 0.0065 mg/L, respectively. It is worth noting that the effluent TP concentration without any TP removal agents dosing exceeded the discharge standard by a margin. When the P removal agents were added, the effluent TP concentration showed an expectingly reached discharge standard (Fig. 3 (d)). In summary, 1.2 times the influent quantity is recommended, and the effluent NH3 and TN concentration should be given to special attention in winter.
Under emulating conditions 5 to 8 (Fig. 3 (e)), the average effluent COD concentration was measured to be 14.03 mg/L, 13.80 mg/L, 13.94 mg/L and 13.79 mg/L, respectively. It was observed that the effluent COD concentration remained relatively unchanged when the water quantity exceeded 1.3 times and 1.5 times effluent quality during summer. However, the average emulating effluent COD concentration were slightly higher compared to emulating conditions 1 to 4. The average NH3-N effluent concentration for 5 ~ 8 scenarios (Fig. 3 (f)) was found to be 0.43mg/L, 0.60 mg/L, 0.34 mg/L and 0.58 mg/L, respectively. It can be seen that the NH3-N concentration easily achieved the discharge limit even during 1.3 and 1.5 times of hyper-quantity operation in summer. Nevertheless, the simulated effluent TN concentration under scenarios 6 and 8 following the addition of P removal agents (mainly Fe ions) exhibited relatively higher fluctuations. The average effluent TN concentration under these simulating conditions was measured to be 11.20mg/L, 11.34 mg/L, 11.76 mg/L and 11.94 mg/L, respectively. The effluent TN concentration remained below the discharge limit of 15 mg/L effluent concentration, when the influent water quantity was increased to 1.3 and 1.5 times (Fig. 3 (f)). However, that effluent TN concentration frequently exceeded the discharge limits, especially when 1.5 times the dosage of the P removal agent was added. Thus, 1.3 times the influent quantity is recommended to operate, and special attention should be given to the effluent COD and TN concentration during hyper-quantity operation in summer. Additionally, the average effluent concentration of TP in the emulated water were found to be 0.41 mg/L, 0.0066 mg/L, 0.75 mg/L and 0.0065 mg/L, respectively, for scenarios 5 to 8. It is noticed that the effluent TP concentration was significantly higher when P removal agent was added compared to the scenarios without their addition, indicating the negative effect of these agents on effluent NH3-N concentration (Fig. 3 (g)) (Jia et al. 2016). Meanwhile, the concentration of effluent TP exceeded the standard considerably, when the P removal agent was not enhanced at the same multiple as the treated quantity. In contrast, the effluent quality reached the standard completely, as the dosing of P removal agents was increased (Fig. 3 (g)).
3.2 Performance of the process
The daily average treated quantity in 2022 reached around 5.1×104 t/d, demonstrating a 50% increase (Fig. 4 (a))) compared to the 2018 level of 3.4×104 t/d. The average treated quantity in January, February and December was nearly 4.88×104 t/d, which was close to the emulated typical winter scenarios 4.8×104 t/d. Furthermore, the average treated quantity in May, June and July was 5.27×104 t/d, which was 1.3 times the designed water quantity of the typical summer project set to 5.2×104 t/d.
Table 3 presented the statistical data regarding the performance of the A2O + MBR process during the hyper-quantity operational periods.
Table 3
Statistics data of the A2O + MBR process performance
| BOD | COD | NH3-N | TN | TP |
Average concentration influent (mg/L) | 121.33 | 242.73 | 43.39 | 48.99 | 6.80 |
Average concentration effluent (mg/L) | 3.13 | 11.86 | 0.26 | 10.99 | 0.20 |
Discharge limits (mg/L) | 6 | 30 | 1.5 | 15 | 0.3 |
Average removal efficiency (%) | 97.4 | 95.07 | 99.4 | 77.5 | 97.06 |
Table 4
Correlation of the main influent water quality
| BOD | COD | NH3-N | TN | TP |
BOD | 1 | 0.999** | 0.314 | 0.303 | 0.481 |
COD | | 1 | 0.306 | 0.294 | 0.502 |
NH3-N | | | 1 | 0.997** | 0.029 |
TN | | | | 1 | 0.041 |
TP | | | | | 1 |
Table 5
SDs of monthly average concentration for effluent water quality (mg/L)
Years | BOD | COD | NH3-N | TN | TP |
2018 | 24.03 | 45.75 | 5.21 | 5.36 | 0.30 |
2022 | 16.31 | 33.10 | 2.14 | 2.21 | 0.30 |
The influent concentrations of BOD and COD exhibited a similar trend in 2022, gradually decreasing from the first half of the year to the second half of the year. On the other hand, the influent concentration of NH3-N, TN and TP remained relatively stable throughout the year. The main discharge limits were consistently reached during this period. The A2/O + MBR process proved to be effective during the hyper-quantity operational periods. The real treated water quantity is closely aligned with emulating predictions, indicating the reliability of the ASM model played a particularly important role in predicting the quantity and quality.
3.3 Impact on the GHG emission
3.3.1 Variations in total GHG emission and intensity
The total GHG emissions form the WWTP was 23.33 ktCO2e in 2022, which represented a 42.69% increase, compared to the 2018 level (16.35 ktCO2e). this increase can be attributed to the higher treated quantity of the WWTP. The GHG emissions and intensity showed minor fluctuation over the analyzed period (Fig. 5 (b)). The monthly total GHG emissions were projected to be relatively constant throughout the year (1.9 ktCO2e), except in May. The lowest monthly emissions of 1.60 ktCO2e were observed in August, while the highest emissions of 2.21 ktCO2e occurred in December (Fig. 5 (a)). Furthermore, the annual GHG emissions from CH4, N2O, electricity consumption and chemical agent consumption GHG emissions were 4.75 ktCO2e, 7.44 ktCO2e, 11.12 ktCO2e, 13.11ktCO2e, respectively (Fig. 5 (b)). In terms of the overall contribution of GHG emissions, CH4 accounted for 20.38%, N2O accounted for 31.89%, electricity accounted for 47.67% and chemical agent consumption accounted for 0.06%, as illustrated in Fig. 6.
The total GHG emission intensity, measure in kg CO2e/m3 of wastewater) was 1.24 kgCO2e/m3 (Fig. 5 (b)) in 2022. This represented a decreased of 5.3% compared to the 2018 level. Moreover, the average GHG emission intensity for CH4, N2O, energy consumption and material consumption were found to be 0.25 kgCO2e/m3, 0.39 kgCO2e/m3, 0.59 kgCO2e/m3 and 0.00070 kgCO2e/m3, respectively (Fig. 5 (b)), with their increasement of 7.4%, 7.1%, 4.9%, 12.5%, respectively.
Further, among the total GHG emissions, CH4, N2O, consumption of electricity and chemical agents accounted for 20.38%, 31.89%, 47.67% and 0.06% of the total GHG emissions (Fig. 6), respectively, suggesting that electricity consumption remained the largest contributor to the total GHG emissions. In comparison, the respective proportions in 2018 were 20.91%, 32.19%, 48.84%, and 32.19% (Hu et al. 2021). The proportion of the four types of GHG emissions nearly unaltered during the hyper-quantity operational period, showing minimal changes compared to those in 2018. However, the total GHG emissions increased as the treated quantity of wastewater increased. Notably, there were a 1.17% reduction in proportion of energy-related GHG emissions, ascribing to the improvement of the aeration efficiency within the activated sludge system. In general, achieving carbon mitigation target often requires carbon- and energy-intensive industries to reduce production, resulting in economic loss for these industries. Nevertheless, moderate hyper-quantity operation can offer dual benefits. It not only helps to reduce GHG emission intensity to some extent but also provides advantages for the operating sectors.
3.3.2 Variations in the direct and indirect GHG Emission
Characteristics of the indirect and direct GHG emissions are shown in Fig. 7. The highest monthly emissions of CH4 and N2O emissions were recorded in January (470.41 tCO2e) and April (685.75 tCO2e), while the lowest emissions were observed in May (376.23 tCO2e) and December (556.75 tCO2e), respectively (Fig. 7 (a)). The average monthly direct CH4 and N2O emissions were estimated at 396.41 tCO2e and 620.16 tCO2e, showing an increase of 28.11% and 41.41%, respectively, compared to the levels recorded in 2018.
Besides, the highest monthly indirect emissions from electricity consumption and chemical consumption emissions were observed in January. (1070.81 tCO2e) and November (1.21 tCO2e) (Fig. 7 (b)), whereas the lowest values occurred in May (575.28 tCO2e) and February (0.88 tCO2e), respectively (Fig. 7 (b)). The average monthly indirect GHG emissions for energy consumption and chemical consumption were found to be 927.22 tCO2e and 1.09 tCO2e, respectively (Fig. 7 (b)), demonstrating an increment of 45.3% and 29.76%, respectively, compared to the levels in 2018. Overall, the growth rate of carbon emissions from N2O and energy consumption were consistent with the growth rate of treated water (50%).
3.4 Sensitivity analysis for GHG emission intensity at different quantities
The sensitivity analysis conducted using the running data from the H-reclaimed WWTP in 2018 and 2022 revealed interesting findings. There was a weakly linear trend observed in the total GHG emission as the percentage change in treated water inclined increase until the quantity achieved the designed treated capacity (Fig. 8 (a)). Further, the percentage change for the GHG emissions varied from − 20–20% at the lower treated quantities (Fig. 8 (a)). However, when the process operated at hyper-quantity conditions, exceeding the design capability 22.5–37.5%, the percentage change in the GHG emissions intensity (-12.5–12.5%) showed higher fluctuation than the total GHG emissions (35–55%) (Fig. 8 (b)). Similar trends were observed in the types of GHG emissions (Fig. 8 (c, d and e)). Additionally, weakly negative correlations were manifested between influent wastewater quantity and different types of GHG emissions (Hu et al. 2021). This trend also supported by sensitivity analysis, illustrating a similar of pattern in the intensity of different types of GHG emissions, except for the N2O emission intensity.
The sensitivity analysis suggested that increasing the percentage change in the treated quantity of WWTP does not always result in an equivalent percentage decrease in GHG emissions intensity. The marginal benefit of this measure appeared to diminish, particularly, when the treated quantity exceeded approximately 125% of the design capability. Moreover, the ASM emulating results also displayed that when the treated quantity exceeds the nearly 125% of the design capability, the risk of effluent quality exceeding the discharge limits increased dramatically especially during winter. Thus, it becomes imperative to explore other technologies and strategies to reduce the GHG emission intensity. In this regard, Su et al. (Su et al. 2023) revealed a potential 29% reduction in GHG emission and intensity by implementing a 0 to 10% reduction in GHG emission in imported electricity. However, it should be noted that the sensitivities of GHG emissions, GHG emission intensity, and water quality to other engineering technologies were ignored in this paper, highlighting the need for a more accurate model and cross-agency collaboration to implementation for future implementation.