Abstract

A micro-pressure swirl reactor (MPSR) was developed for carbon and nitrogen removal of wastewater, in which dissolved oxygen (DO) gradient and internal circulation could be created by setting the aerators along one side of the reactor, and micro-pressure could be realized by sealing most of the top cap and increasing the outlet water level. In this study, velocity and DO distribution in the reactor was measured, removal performance treating high-concentration wastewater was investigated, and the main functional microorganisms were analyzed. The experiment results indicated that there was stable swirl flow and spatial DO gradient in MPSR. Operated in sequencing batch reactor mode, distinct biological environments spatially and temporally were created. Under the average influent condition of chemical oxygen demand (COD) concentration of 2,884 mg/L and total nitrogen (TN) of 184 mg/L, COD removal efficiency and removal loading was 98% and 1.8 kgCOD/(m3·d) respectively, and TN removal efficiency and removal loading reached up to 90% and 0.11 kgTN/(m3·d) respectively. With efficient utilization of DO and simpler configuration for simultaneous nitrification and denitrification, the MPSR has the potential of treating high-concentration wastewater at lower cost.

HIGHLIGHTS

  • The experiment designed a new type of micro-pressure swirl reactor (MPSR), which can realize the coexistence of different dissolved oxygen environments (anaerobic/anoxic/aerobic) in a single aeration tank, which improves the efficiency of oxygen utilization.

  • MPSR achieves high-efficiency simultaneous removal of high-concentration chemical oxygen demand and nitrogen, and has the potential to treat high-concentration wastewater at a lower cost than a conventional sequencing batch reactor.

  • MPSR achieves short-term simultaneous nitrification and denitrification in the treatment of high-concentration wastewater. MRSR's unique aeration method is the key to its realization.

Graphical Abstract

Graphical Abstract
Graphical Abstract

INTRODUCTION

Energy consumption represents a significant part of the operative costs of a wastewater treatment plant (WWTP). In a conventional WWTP, about 25–40% of the operating costs is ascribable to energy consumption; the main contributor to energy consumption is the aeration of mixed liquor, which typically accounts for 55–70% of the energy consumed (Venkatesh & Bratteb 2011,;Elias-Maxil et al. 2014; Panepinto et al. 2016). Much effort has been made to find ways to reduce aeration energy, such as developing efficient blowers, applying in situ feedback control strategy for aeration and improving management level of the WWTP (Serralta et al. 2002; Panepinto et al. 2016; Sun et al. 2016).

On the other hand, to comply with the more restrictive discharge standards, advanced biological treatment processes are needed to improve nutrient removal efficiency. For conventional biological nitrogen removal, it involves nitrification of ammonia nitrogen (-N) to nitrate nitrogen (-N) followed by denitrification of nitrate nitrogen to nitrogen gas (N2). In recent years, simultaneous nitrification and denitrification (SND) has been described in the literature for various systems (He et al. 2008; Walters et al. 2009; Zheng & Cui 2012). Such processes eliminate the need for two separate reactors or intermittent aeration, and thus simplify the treatment system. The other advantage is that the organic compounds in the wastewater can be utilized efficiently as hydrogen donors for denitrification. Also, neutral pH level and less demand for alkalinity can be accomplished.

In this study, a micro-pressure swirl reactor (MPSR) was developed, which had two main characteristics, namely high efficient utilization of dissolved oxygen (DO) and simpler configuration for SND. Consequently it is expected that carbon and nitrogen in the wastewater could be more efficiently removed at lower cost. In our previous study (Bian et al. 2015; Ren et al. 2017), it was found that chemical oxygen demand (COD) and total nitrogen (TN) removal efficiency of the MPSR operating at sequencing batch reactor (SBR) mode was 10–20% and 15–20% higher respectively than that of a conventional SBR. In this paper, removal performance of the MPSR treating higher concentration organic carbon and ammonia wastewater was investigated; velocity and DO distribution and main functional microorganisms were analyzed, to better understand the removal capacity and mechanisms of the novel reactor.

MATERIAL AND METHODS

Concept of the MPSR

Figure 1(a) shows the basic structure and nitrogen removal concept of the MPSR. The reactor comprised two parts, the main reaction part and water level-lifting part. By setting the aeration diffuser along one side at the bottom, internal circulation could be created with liquid flow velocity and DO concentration difference. By increasing the height of the reactor outlet, micro-pressure could be created and higher saturated DO concentration could be obtained. Moreover, because of the top sealing structure, the contact time of air bubbles with wastewater could be prolonged and higher oxygen transfer efficiency could be achieved. At proper aeration rate, wastewater and activated sludge could get mixed and circulated, DO concentration decreased gradually from the outer part to the center of the reactor, SND was supposed to be achieved with nitrification occurring in the outer aerobic zone and denitrification occurring in the inner anoxic zone, and simultaneous carbon and nitrogen removal could be realized with higher DO utilization efficiency.

Figure 1

(a) Schematic diagram of the micro-pressure swirl reactor (MPSR) under ideal conditions; (b) photograph of the MPSR model.

Figure 1

(a) Schematic diagram of the micro-pressure swirl reactor (MPSR) under ideal conditions; (b) photograph of the MPSR model.

As shown in Figure 1(b), a bench scale reactor was established to investigate the velocity and DO distribution in the main reaction part and high-concentration wastewater treatment performance of the MPSR. The reactor was made of plexiglass with the effective volume of 54 L. The main reaction part was 800 mm long, 100 mm wide and 600 mm high; the upper water-lifting part was 100 mm long, 80 mm wide and 800 mm high. There were 30 holes in the side wall of the reactor for installing the device for detecting flow velocity and DO. The diameter of the holes was 30 mm. The aeration tube was installed on the bottom of the side opposite to the water level-lifting part. The distance from its center to the bottom and side wall of the reactor was 50 mm. The aeration tube was a plexiglass tube with a diameter of 10 mm. The part of it extending into the reactor was 100 mm long, and 40 small holes with a diameter of 0.8 mm were evenly distributed on this part.

Flow velocity distribution test

The flow velocity distribution in MPSR was monitored with a flow meter. The water temperature was 20 ± 1 °C and the reactor was filled with tap water; the velocity distribution of the main reaction part of MPSR was studied. The test had showed that the flow velocity distribution of the reactor in the width direction was uniform under the influence of uniformed aeration of the aeration tube. The cross-section of the reactor at a width of 50 mm was used as coordinate plane. The apex of the right angle at the bottom of the main reaction part without the aeration device was the origin, the length direction was the X axis, and the height direction was the Y axis. The velocity sensors of the liquid velocity meter were also placed on this plane. The data collection points for the velocity distribution are shown in Figure 2(a). Thirty flow velocity collection points were set in the test. The multi-machine non-constant flow function of the multi-functional intelligent flow meter (LGY-III, Nanjing Institute of Water Resources Research, China) was selected to measure the internal flow velocity. The device installation is shown in Figure 2(b). The principle of the LGY-III type multi-function intelligent flow meter was to collect the rotation speed of the rotor of the front end of the flow sensor within a certain time, and obtain the average flow rate of the liquid at the monitoring point in this period through conversion calculation. The parameters of the instrument were set as follows: the flow rate scale was 1, the acquisition time was 5 seconds, the acquisition interval was 2 seconds, and 20 sets of data were collected. The multi-channel and multi-address method was used to measure the flow rate, so as to ensure that the data collection time of each point was consistent. The following steps were used in the test.

  • 1.

    The reactor was filled with tap water to the predetermined water level, and the water temperature was controlled at 20 ± 1 °C . The test was performed in a constant temperature controlled room, and the water temperature was adjusted to 20 ± 1 °C by controlling the room temperature.

  • 2.

    An air compressor (ACO-12, Sun Sun, China) was used to aerate the reactor, and the aeration amount was controlled by an air flow meter (LZB-DK600-4F, Cheng feng, China) to 0.2 m3/h. The rise in water level after adjusting to the set value was recorded, and velocity data were collected. Because the rise of the liquid level would cause the display value of the flow meter to decrease, the test stipulated that the set aeration amount was based on the actual display value of the flow meter after inflation.

  • 3.

    The average velocity of each point was taken and the contour map of the internal velocity distribution of MPSR was drawn by using a commercial 3D visualization software package (Surfer 8.0, Golden Software, USA).

Figure 2

(a) Monitoring point layout of the reactor for flow velocity and DO test (unit of distance between detection points: mm); (b) photograph of the velocity meter on-site installation.

Figure 2

(a) Monitoring point layout of the reactor for flow velocity and DO test (unit of distance between detection points: mm); (b) photograph of the velocity meter on-site installation.

MPSR treatment of high-concentration wastewater

Inoculated with surplus sludge from a local wastewater treatment plant, the reactor was operated in a sequencing batch mode and fed with synthetic wastewater. One single operation cycle of the reactor comprised 1 h filling, 10 h aeration, 45 min setting and 15 min draining. The volumetric exchange ratio was 35%, the sludge residence time was controlled at 25–30 days by regularly discharging biomass, and the mixed liquid suspended solids (MLSS) concentration was maintained at around 4,000–4,500 mg/L. The influent synthetic wastewater was prepared by mixing beer, NH4Cl and KH2PO4 with other required nutrients in tap water. COD concentration of the influent was in the range of 713–2,979 mg/L and TN was 46–215 mg/L. During the operation, aeration rate was kept at 0.2 m3/h. Temperature of the influent was maintained at 20 ± 1 °C.

Routine tests were carried out to determine the overall treatment performance of the MPSR treating high-concentration wastewater. Additionally, cycle tests were conducted to probe removal characteristics of the reactor by monitoring COD and nitrogen concentration variation over one single cycle. COD, TN and MLSS were measured according to standard methods (APHA 2007), -N and -N were analyzed by ion chromatography (Metrohm883, Switzerland).

Dissolved oxygen distribution test

The DO distribution test in the main reaction part of MPSR was conducted after the test of MPSR treatment of high-concentration wastewater was completed. When the MPSR was in an idle state, the activated sludge was transferred out of the reactor, and the DO detection devices were installed on the side wall of the reactor. The test used eight four-channel DO detectors (DTD-9600, Shanghai Nuobo Environmental Protection Technology Co., Ltd, China) to measure the DO distribution. After the electrolyte calibration, the detection accuracy and error between each instrument and each channel of each instrument had reached the test requirements. The detection points of the DO distribution were shown in Figure 2(a). The detection points were consistent with the flow velocity distribution detection. The difference compared with the flow velocity distribution detection was that the diameter of the DO detection probe was larger and the length of the probes extending into the reactor were 10 mm to reduce the influence of the probe on the flow of the mixed liquid in the reactor. That means the DO distribution in the cross-section at 10 mm of the reactor width was tested in this study. After the DO detection devices were installed, the activated sludge was backfilled into the reactor. When it was confirmed that the performance of system wastewater treatment was not affected, the DO distribution study of the typical cycle of the main reaction part was started.

During the test, the control parameters such as reactor operating cycle, aeration, drainage ratio, and sludge age were consistent with the wastewater treatment test. The influent COD and TN concentrations were 2,800 and 180 mg/L, respectively. The difference in influent COD and TN concentration between cycles was less than 50 and 10 mg/L, respectively, to reduce the influence of influent water quality fluctuations on the DO distribution. DO data collection started at the beginning of aeration, and the DO detection devices continuously detected for 600 minutes in each cycle, and recorded data every 15 minutes. The test was carried out five times and completed in five cycles. The average value of the five tests was recorded as the duration data of DO at each point under the typical cycle of MPSR.

Microbial community analysis

Biomass samples were collected during the operation period. The samples were washed with phosphate-buffered saline, and then DNA was extracted using a DNA isolation kit (MO BIO PowerSo) (Amann et al. 1995); GC-clamp-F357 primer (5′–CGCCCGCCGCGCCC CGCGCCCGGCCCGCCGCCCCCGCCCCCCTACGGGAGGCAGCAG-3′) and R517 primer (5′-ATTACCGCGGCTGCTGG-3′) were used to amplify the bacterial 16SrRNA gene. The polymerase chain reaction (PCR) products were separated by using denaturing gradient gel electrophoresis (DGGE) with a D-Code universal mutation detection system (Bio-Rad Laboratories, Hercules, CA, USA) according to the instruction manual. The dominant bands in the DGGE gel were excised and used as direct templates for PCR, agarose gel was recovered and purified after electrophoresis and DNA fragments were sequenced.

RESULTS AND DISCUSSION

Research on velocity distribution in MPSR

Figure 3(a) shows the contour map of the velocity distribution of the MPSR main reaction part when the aeration was 0.2 m3/h. There was relatively stable swirl flow in the main reaction part, and the flow velocity gradually decreased from the periphery to the center. The outer peripheral velocity of liquid was greater than 0.05 m/s, center velocity was in the range of 0.02–0.04 m/s and the core velocity liquid was approximating 0 m/s. There was a relatively stable low-speed circulation area that formed in the center of the main reaction part. Meanwhile, the center of it did not coincide with the center of the main reaction part. The reason was that on the right side of the area around the main reaction part, the flow velocity in this area was faster while the water flow cross-section, which was affected by single-side aeration, was smaller. When the water flow was driven to the left side of the main reaction part by aeration, due to the influence of the wall and the gas–liquid separation, the flow velocity decreased and the water flow cross-section diffused. However, the water flow velocity was still higher than the central circulation area, so that the center of the low-speed circulation area was biased to the right of the main reaction part. The flow velocity distribution in the main reaction part provided the basic conditions for the formation of different DO zones in actual operation. In addition, due to the impact and reflection of the side plate on the water flow, the water flow velocity decreased in both the upper right corner and the lower left corner of the main reaction part.

Figure 3

(a) Contour map of water flow velocity distribution in the main reaction zone when the aeration was 0.2 m3/h (unit of water flow velocity: m/s); (b) schematic diagram of the MPSR flow partition.

Figure 3

(a) Contour map of water flow velocity distribution in the main reaction zone when the aeration was 0.2 m3/h (unit of water flow velocity: m/s); (b) schematic diagram of the MPSR flow partition.

Through the study of the flow velocity distribution, the results showed that the MPSR main reaction part could be divided into five zones according to the different directions and modes of water flow. The regional distribution diagram is shown in Figure 3(b). They were up-flow zone (I), upper horizontal acceleration zone (II), down-flow zone (III), lower horizontal acceleration zone (IV) and central low-speed circulation zone (V). The up-flow zone was located directly above the aeration diffusion device. Under the action of aeration, the bubbles accelerated upward due to buoyancy, and the water current also accelerated upward under the influence of the bubbles. This zone was the main source of power for the water circulation in the reactor, and it was also the zone with the fastest water flow velocity. The top horizontal accelerated flow zone was located near the top cover of the reactor. When water and air bubbles accelerating in the up-flow zone met the top cover, the flow rate dropped rapidly. Some bubbles started to move horizontally with the inertia of the water flow, but some bubbles would gather near the cover. As the volume of the bubbles increased, the water flow began to accelerate horizontally under the continuous squeezing of the bubbles. The down-flow zone was located on the opposite side of the up-flow zone, and the flow direction was also opposite to that of the up-flow zone. This zone would wrap some of the bubbles in the upper horizontal acceleration zone, but due to buoyancy, the number of bubbles that were wrapped was small and the volume of them was small too. The down-flow zone was also in an accelerated motion state, and its power was mainly derived from the gravity acceleration of the water. The lower horizontal acceleration zone was located in the near-wall area at the bottom of the reactor. There were almost no air bubbles in this area. The water flow was accelerated horizontally under the suction effect of the aeration area. Due to the limited scope of aeration, when the amount of aeration was small and the water flow energy on the side near the down-flow zone was insufficient, activated sludge precipitation in this area was prone to occur in actual operation. The central low-speed circulation zone was located in the central area of the main reaction zone. There were no air bubbles in this zone. Due to the viscosity of the water, the kinetic energy of the water flow in the four surrounding zones was transferred to the central area, thereby ensuring the continuity and stability of the cycle. Although there were different acceleration forces in the four surrounding zones, it was difficult to continuously accelerate the water flow due to the blockage of the reactor side wall. At the same time, turbulence existed in the surrounding zones due to aeration disturbance, gas–liquid separation, and side wall reflection of the water flow. The existence of these turbulences promoted the material exchange between the surrounding zones and the central circulation zone, which had a certain promotion effect on the actual sewage treatment of MPSR.

Removal performance and characteristics

In order to explore the ability of MPSR to remove organic matter and denitrify simultaneously, this study comprised three stages of experiments. Firstly, the influent water concentration range of COD and TN in the cultivation phase was 713–1,500 mg/L and 46–70 mg/L, respectively, and the comparison proved that the sewage treatment capacity of MPSR was superior to the traditional SBR process (Bian et al. 2015). To further explore the treatment capacity of MPSR, the influent concentration of COD and TN were gradually increased. A buffer training phase was added between the cultivation phase and the formal operation phase to prevent the irreversible impact of high-concentration influent on the biochemical system. At this phase, the COD and TN influent concentration ranges were 1,500–2,000 mg/L and 100–120 mg/L, respectively. When the reactor ran to the 61st day and the removal rate of TN was close to 80%, the concentration of COD and TN in the influent was further increased and the formal operation phase of high-concentration wastewater treatment was started. At this time, the influent concentration range of COD and TN was 1,800–3,000 mg/L and 170–200 mg/L respectively.

Figure 4(a) shows the reactor had satisfying COD removal performance. When the influent COD concentration was increased from about 1,000 mg/L to nearly 3,000 mg/L, the effluent COD was always less than 100 mg/L. During the formal operation phase when the influent COD was in the range of 2,782–2,978 mg/L (day 88–day 96), the effluent COD was 60–81 mg/L, and COD removal efficiency and removal loading was up to 98% and 1.8 kgCOD/(m3·d) respectively.

Figure 4

COD and TN removal performance of MPSR. (a) COD; (b) TN.

Figure 4

COD and TN removal performance of MPSR. (a) COD; (b) TN.

Figure 4(b) shows the TN removal effect of MPSR in the buffer training phase and the formal operation phase. It can be seen from the figure that at the beginning of the buffer training phase (the reactor ran to day 50), the removal rate of the system TN was only 26.65%. The main reason was the denitrifying bacteria group in the reactor had not been trained and their number was small. When the reactor was operated to the 58th day, the TN removal rate reached 78.56%. It showed that after a period of training, MPSR could enrich the denitrifying bacteria group, which improved the nitrogen removal efficiency of the system. From the continuous training of the reaction to the 61st day, the denitrification efficiency was stable at about 80%. After that, the formal operation phase was carried out and the TN concentration in the influent increased from 112 mg/L to 182 mg/L. At this time, MPSR had undergone the same process as the buffer training phase. At the beginning of the formal operation phase(the reactor ran to day 64), the activated sludge in the system had not yet been adapted to the high concentration of TN in influent, and the nitrogen removal rate had dropped to only about 55.08%. At the 81st day of operation, the TN removal rate of the system reached 79.29% and the TN removal efficiency continued to stabilize at around 80%.

Influent COD concentration is considered to be an important factor affecting the nitrogen removal process. In this study, the effect of influent COD concentration on MPSR nitrogen removal was embodied in two aspects. First, organic matter directly participated in the denitrification process, providing electrons and energy for the reduction of nitrate nitrogen. Secondly, because the biological denitrification process was closely related to the DO environment, the organic matter could change the DO distribution in the MPSR through respiration, thereby affecting the system denitrification process. At the end of the official operation phase (day 88–day 96), the average COD concentration of the influent was increased to 2,884 mg/L, while the TN concentration of the influent was unchanged. It can be seen from Figure 4 that the COD concentration of the effluent was stable at this time, and the average effluent concentration was below 100 mg/L. Meanwhile, the TN effluent concentration further decreased, and the average removal rate of TN increased from about 80% to more than 90%. The increase in the COD concentration of the influent provided sufficient carbon source for denitrification, and prolonged the existence time of the anoxic environment in the MPSR, thereby promoting further denitrification of the system.

Figures 5 and 6 show COD and nitrogen variation in one single operating cycle at day 58, day 75 and day 95. It can be noticed from Figure 5 that COD concentration decreased sharply during the first hour filling stage. When the influent COD was 2,012, 1,803 and 2,782 mg/L, COD concentrations at the end of the filling stage were only 274, 307 and 755 mg/L respectively. One explanation for the drop was the dilution effect of the residual mixed liquid of the last cycle, and the other reason was the absorption of the activated sludge. It was also noticed that COD in the mixed liquid was almost removed during the first 2 aeration hours, decreasing little during the latter 8 hours, which indicated that the operating cycle could probably be further shortened if only organic carbon was to be removed.

Figure 5

Typical COD variation in a single operating cycle of MPSR.

Figure 5

Typical COD variation in a single operating cycle of MPSR.

Figure 6

Typical nitrogen variation in a single operating cycle of MPSR: (a) day 58; (b) day 75; (c) day 95.

Figure 6

Typical nitrogen variation in a single operating cycle of MPSR: (a) day 58; (b) day 75; (c) day 95.

To explore the law of denitrification by MPSR in the treatment of high-concentration wastewater, the experiment included a diachronic detection of TN, -N, -N and -N in a single operating cycle. Figure 6 shows the changes of TN, -N, -N and -N with the aeration time in the single operation cycle on the 58th day (the cultivation period of the buffer training phase), 75th day (the cultivation period of the formal operation phase) and 95th day (the stable period of the formal operation phase) of the reactor operation. It can be seen from the figure that on the 58th day of the reactor operation, it was the nitrification process that restricted the improvement of the nitrogen removal efficiency of the system, and the -N in the effluent TN accounted for 76.44%. At this time, the nitrification effect of the system was poor. The reason was that the system did not accumulate a sufficient amount of nitrifying bacteria in the cultivation phase, resulting in a low nitrification rate of the system. In addition, the TN concentration increased by this research was mainly composed of -N. When the concentration of -N in the influent increased, the concentration of free ammonia in the influent also increased, which would inhibit the activity of nitrifying bacteria. This was also one of the important reasons for the poor nitrification effect of the system at this time.

Figure 6(b) shows the change of various forms of nitrogen with the aeration time on the 75th day of reactor operation. It can be seen from the change over time of -N that there was also a phenomenon of insufficient nitrification in the reactor at this time. From the figure, the TN concentration decreased only within 0–2 h of aeration, and remained almost unchanged at other times. At this time, the -N concentration and the TN concentration decreased synchronously, which indicated that the decrease of the TN concentration during this period was not only due to the dilution of the influent, but also the SND. The smooth progress of SND was attributed to the structural design of MPSR. During MPSR aeration, a number of bubbles gather in the upper right corner and under the semi-closed top cover of the reactor. These areas obtained higher DO from the beginning of aeration, which facilitated the nitrification reaction. Meanwhile, most areas of the reactor were in an anoxic or anaerobic environment. Figure 5 shows that there was free organic matter in the reactor at this time, which was advantageous for denitrification. As the aeration progressed, the TN concentration basically did not decrease. Within 2–10 h of aeration, the drop in concentration of -N was almost equal to the concentration of -N accumulated, and -N did not accumulate throughout the reaction. This indicated that under the experimental conditions, the removal of nitrogen in this system was mainly through short-term SND. Because the denitrification bacteria in the system were not domesticated and matured at this time, the denitrification effect was poor.

Figure 6(c) shows the change of various forms of nitrogen with the aeration time on the 95th day of reactor operation. With the continuous operation of the reactor, the denitrification effect of the final system was good, and the TN removal rate was 90.36% at this time. During the aeration period of 0–5 h, the -N concentration and TN concentration decreased synchronously with a faster speed, which indicated that the system was performing efficient SND. Through continuous culturing, MPSR enriched the nitrosating bacteria, which enabled the system to carry out high-efficiency nitrosation. At the same time, with sufficient substrates such as nitrite nitrogen, organic matter and free ammonia, the denitrification of the system was not limited. During the aeration period of 5–7 h, the rate of decrease of -N and TN suddenly became slower. Combined with the change of DO over time (Figure 7(a)), it could be found that the DO in each part of the reactor was gradually increased. This indicated that the organic matter in the system was gradually consumed completely, and the decrease in the concentration of organic matter and -N reduced the rate of short-term SND. Another proof was that the accumulation rate of -N gradually increased during this period. During the aeration period of 7–8 h, the -N and TN concentrations showed a period of rapid decline again. Similarly, comparing the DO duration, it could be found that within 7–8 h of aeration, DO in other areas began to rise rapidly except for the central area of the main reaction part of MPSR, while the central area was still in anoxic state. The expansion of the aerobic area and the rise of DO were conducive to the progress of nitrosation. The gradual accumulation of nitrite nitrogen and the presence of an anoxic environment made it possible for bacteria to use their own carbon source to complete short-term SND. During the aeration period of 8–10 h, -N was further converted into nitrite nitrogen, and the TN concentration remained basically unchanged. Meanwhile, the substrates such as -N and organic matter were completely consumed; the system DO continued to rise, and the anoxic environment in the central area of MPSR gradually disappeared. These factors inhibited the system from further removing TN. At this time, the accumulation of -N further confirmed that the main nitrogen removal process of the system was short-term SND.

Figure 7

(a) Changes of DO at monitoring points 1–1, 3–4, and 5–6 with aeration time under the typical cycle of MPSR; (b) DO distribution in the main reaction part at 30 min of aeration (unit of DO: mg/L); (c) DO distribution in the main reaction part at 375 min of aeration; (d) DO distribution in the main reaction part at 600 min of aeration.

Figure 7

(a) Changes of DO at monitoring points 1–1, 3–4, and 5–6 with aeration time under the typical cycle of MPSR; (b) DO distribution in the main reaction part at 30 min of aeration (unit of DO: mg/L); (c) DO distribution in the main reaction part at 375 min of aeration; (d) DO distribution in the main reaction part at 600 min of aeration.

In conventional SBR reactors, TN removal efficiency through SND was limited (Gieseke et al. 2002; Yang et al. 2010). In MPSR, high TN removal efficiency was mainly attributed to the unique structure of the system. It is also worth noting that, under the test conditions, MPSR achieved short-term SND. Compared with the conventional nitrogen removal process, the shortcut nitrification and denitrification process reduced the demand of oxygen and organic carbon by 25 and 40% respectively. The test confirmed that the MPSR has the potential of treating high-concentration wastewater at lower cost.

DO duration and distribution of MPSR in a typical cycle

Figure 7(a) shows the change of DO with aeration time at three DO monitoring points (1–1, 3–4 and 5–6) in the main reaction part of MPSR in a typical cycle. Figure 7(b)–7(d) show the contour maps of DO distribution in the main reaction part at 30 minutes aeration, 375 minutes aeration and 600 minutes aeration, respectively. It can be seen that the duration of the aeration cycle was 600 min (10 h). During the aeration period of 0–345 min, the average DO values at the three monitoring points were 0.045, 0.021, and 0.54 mg/L, respectively. The DO at points 5–6 was significantly higher than the other two points. With reference to Figure 7(b), it can be seen that the DO in most areas of the reactor during this period was below 0.1 mg/L. The DO value was higher in the upper right corner and the area under the semi-closed top cover of the reactor, which laid the foundation for the domestication culture of the reactor's short-term nitrification. During the aeration period of 345–420 min, the average DO values at the three monitoring points were 0.32, 0.04, and 1.65 mg/L, respectively. With reference to Figure 7(c), it can be seen that the overall DO in the reactor was higher than the previous period. The aerobic area in the reactor increased, and the DO in the aerobic area also increased, while the central area of the reactor was still in anoxic state at this time. Compared with the conventional SBR aeration tank, this change rule of DO in MPSR prolonged the existence time of the anoxic/anaerobic zone in the reactor, thus avoiding the contradiction between denitrification and nitrification time to a certain extent, and improving the denitrification efficiency of the system. During the aeration period of 420–600 min, the average DO values at the three monitoring points were 2.13, 0.23, and 30.5 mg/L, respectively. With reference to Figure 7(d), it can be seen that the DO in the reactor increased rapidly during this period, and the main reaction zone of the reactor was in aerobic state by the end of the aeration period (600 min).

Looking back at the description of the MPSR, it could be seen that the MPSR improved the oxygen transfer efficiency through the liquid level-lifting part and the semi-closed top cover plate. In the experiment of treating high-concentration wastewater, through the study of the duration and distribution of DO, it was found that the unique aeration method of MPSR provided the basic conditions for the domestication and cultivation of short-term nitrification which enabled MPSR to save energy while ensuring an effective treatment.

Microbial community analysis

DNA bands from the DGGE profiles were cloned and sequenced, and the sequences were compared with the data in the GenBank for homology analysis. As shown in Table 1, most bacteria in the reactor were classed as ‘uncultured bacterium’. Among the 16 DNA bands, eight bands had the highest similarity with Proteobacteria, three bands had the highest similarity with Bacteroidetes, three bands had the highest similarity with Firmicutes and one band had the highest similarity with Acidobacteria. Among them, the Proteobacteria accounted for 50%. Studies have confirmed that most microorganisms that play an important role in the process of biological nitrogen removal, biological phosphorus removal and degradation of many pollutants belong to the Proteobacteria phylum (Nguyen et al. 2011). The emergence of more types of Proteobacteria proved that the system has a strong ability to remove organic matter, nitrogen and phosphorus. Among all the DNA bands, Nitrosomonas sp. has been the focus of attention. Its appearance indicated that the reactor was enriched with higher abundance of nitrifying bacteria, thus confirming that the denitrification of the reactor was mainly through short-term SND.

Table 1

fragment sequencing analysis results

BandLength (bp)Highest similarity bacteria (No.)Similarity (%)
169 Uncultured Acidobacterium sp. (JQ065958100 
174 Uncultured bacterium (CU925702100 
189 Uncultured Bacteroidetes bacterium (EU24682292 
195 Desulfobulbaceae bacterium PR2_D12 (HE60089097 
194 Uncultured Nitrosomonas sp. (GQ22764499 
189 Uncultured Flavobacteriales bacterium (GU92937798 
189 Uncultured Sphingobacteriales bacterium (AY69447797 
169 Eubacterium ventriosum (L34421100 
194 Erysipelotrichi bacterium (JF345332100 
10 169 Uncultured alpha proteobacterium (FN67923298 
11 169 Uncultured proteobacterium (EU30054798 
12 195 Acinetobacter sp. SH825131.2 (AJ63363698 
13 169 Arcobacter sp. GCDN6_III (JQ072063100 
14 195 Acinetobacter sp. AHJ6 (JQ837283100 
15 195 Uncultured Clostridia bacterium (FJ535170100 
16 169 Uncultured proteobacterium (GQ35500596 
BandLength (bp)Highest similarity bacteria (No.)Similarity (%)
169 Uncultured Acidobacterium sp. (JQ065958100 
174 Uncultured bacterium (CU925702100 
189 Uncultured Bacteroidetes bacterium (EU24682292 
195 Desulfobulbaceae bacterium PR2_D12 (HE60089097 
194 Uncultured Nitrosomonas sp. (GQ22764499 
189 Uncultured Flavobacteriales bacterium (GU92937798 
189 Uncultured Sphingobacteriales bacterium (AY69447797 
169 Eubacterium ventriosum (L34421100 
194 Erysipelotrichi bacterium (JF345332100 
10 169 Uncultured alpha proteobacterium (FN67923298 
11 169 Uncultured proteobacterium (EU30054798 
12 195 Acinetobacter sp. SH825131.2 (AJ63363698 
13 169 Arcobacter sp. GCDN6_III (JQ072063100 
14 195 Acinetobacter sp. AHJ6 (JQ837283100 
15 195 Uncultured Clostridia bacterium (FJ535170100 
16 169 Uncultured proteobacterium (GQ35500596 

CONCLUSIONS

The MPSR showed good performance treating high-concentration wastewater. COD and TN removal efficiencies reached 98 and 90% respectively under the influent loading of 1.8 kgCOD/(m3·d) and 0.11 kgTN/(m3·d), and nitrogen was removed via shortcut nitrification and denitrification throughout the aeration phase. There was relatively stable circular flow and DO gradient in MPSR. Operated in SBR mode, MPSR provided distinct biological environments spatially and temporally, which ensured diverse microbial communities and the occurrence of microbial reactions. With its simple structure for simultaneous organic carbon and nitrogen removal and higher efficient utilization of DO, the novel MPSR is considered very promising for practical wastewater treatment, with smaller footprint and less energy consumption compared with other systems such as anoxic/aerobic, anaerobic/anoxic/aerobic and oxidation ditch processes that require a large area for aeration.

ACKNOWLEDGEMENT

The research was funded by the National Natural Science Foundation of China (51878067) and the Jilin Province Science and Technology Development Program (20170101082JC; 20180201020SF).

DATA AVAILABILITY STATEMENT

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

REFERENCES

REFERENCES
APHA
2007
Standard Methods for the Examination of Water and Wastewater
, 22nd edn.
American Public Health Association, American Water Works Association, and Water Environment Federation
,
Washington, DC, USA
.
Bian
D. J.
Zhou
D. D.
Huo
M. X.
Ren
Q. K.
Tian
X.
Wan
L. G.
Zhu
S. Y.
Ai
S. S.
2015
Improving oxygen dissolution and distribution in a bioreactor with enhanced simultaneous COD and nitrogen removal by simply introducing micro-pressure and swirl
.
Appl. Microbiol. Biotechnol.
199
,
8741
8874
.
Elias-Maxil
J. A.
Peter Van der Hoek
J.
Hofman
J.
Rietveld
L.
2014
Energy in the urban water cycle: actions to reduce the total expenditure of fossil fuels with emphasis on heat reclamation from urban water
.
Renewable Sustainable Energy Rev.
30
,
808
820
.
Nguyen
H. T. T.
Le
V. Q.
Hansen
A. A.
Nielsen
J. L.
Nielsen
P. H.
2011
High diversity and abundance of putative polyphosphate-accumulating Tetrasphaera-related bacteria in activated sludge systems
.
FEMS Microbiol. Ecol.
76
(
2
),
256
267
.
Panepinto
D.
Fiore
S.
Zappone
M.
Genon
G.
Meucci
L.
2016
Evaluation of the energy efficiency of a large wastewater treatment plant in Italy
.
Appl. Energy
161
,
404
411
.