A BAFFLED MEMBRANE BIOREACTOR NEW MBR FOR EFFICIENT NITROGEN REMOVAL

J. Environ. Eng. Manage., 16(6), 435-439 (2006) A BAFFLED MEMBRANE BIOREACTOR─ NEW MBR FOR EFFICIENT NITROGEN REMOVAL Yoshimasa Watanabe* and Katsuki...
Author: Julie Carr
1 downloads 0 Views 277KB Size
J. Environ. Eng. Manage., 16(6), 435-439 (2006)

A BAFFLED MEMBRANE BIOREACTOR─ NEW MBR FOR EFFICIENT NITROGEN REMOVAL Yoshimasa Watanabe* and Katsuki Kimura Department of Urban and Environmental Engineering Hokkaido University Sapporo 060-8628, Japan

Key Words: Denitrification, membrane bioreactors, nitrification, wastewater treatment ABSTRACT Submerged membrane bioreactors (MBRs) have been gaining in popularity in various types of wastewater treatment. One drawback of submerged MBRs is the difficulty in removing nitrogen because intensive aeration is usually carried out in the tank and the MBRs must therefore be operated under aerobic condition. In this paper, feasibility of a baffled MBR (BMBR), treating municipal wastewater particularly in terms of nitrogen removal, was examined on the basis of pilotscale experiments. Simultaneous nitrification/denitrification in a single reactor was possible by inserting baffles into a single submerged MBR equipped with polyvinylidenfluoride flat sheet membranes as long as the wastewater was fed in the appropriate way. When operating conditions of the BMBR were appropriately set, the nitrogen removal efficiency exceeded over 95% without adding any external organic carbon. Total nitrogen concentration in the treated water could be reduced to about 1 mg L-1 with the hydraulic retention time of 5.3 h. Very efficient and stable removal of organic carbon and phosphorus was also found. Average concentration of total organic carbon and total phosphorus in the treated water was 3.6 and 0.17 mg L-1, respectively. INTRODUCTION As an efficient technology for municipal wastewater treatment, membrane bioreactors (MBRs) have gained significant popularity in the past decade. MBRs, in which biomass is strictly separated by a membrane, offer several advantages over the conventional activated sludge process, including a high biomass concentration, reduced footprint, low sludge production, and better performance quality [1]. MBRs can be generally classified into two categories: recirculated MBRs and submerged MBRs. Recently, submerged type of MBRs have been preferred since energy consumption can be significantly reduced [2]. Generally, intensive aeration is carried out in a submerged MBR to supply oxygen to microorganisms and clean the membrane. As a result of the intensive aeration, one weak point in the use of submerged MBRs becomes obvious: poor removal of nitrogen. Regarding elimination of NH4+ (i.e., nitrification), submerged MBRs generally show good performances due to dense populations of nitrifiers and the aerobic condition provided by the intensive aeration. However, anoxic condition that is indispensable for the promo*Corresponding author Email: [email protected]

tion of biological denirification cannot be created under such a high intensity of aeration. One possible approach to overcome this drawback is to install an additional anoxic reactor in which denitrification occurs followed by the aerobic MBR [3-5]. Mixed liquor is circulated between the two reactors at a fixed ratio. With this configuration, nitrogen removal in submerged MBRs can be possible. Removal of up to 80% of total nitrogen (TN) in municipal wastewater has been reported. This type of MBR, however, apparently impairs the advantages of the submerged MBRs such as small footprint or ease of operation. To address the problems stated above, insertion of baffles into the membrane chamber was proposed by the authors. Although this approach proposed herein is rather simple, in combination with an appropriate way of the feed water addition, it can significantly improve the performance of submerged MBRs by promoting simultaneous nitrification/denitrification. Preliminary experiments [6] demonstrated the proposed baffled MBR (BMBR) certainly worked well. This paper will describe the results obtained in pilotscale experiments with a real wastewater and show the feasibilities of the proposed reactor.

J. Environ. Eng. Manage., 16(6), 435-439 (2006)

436

MATERIALS AND METHODS 1. Concept of the BMBR

Figure 1 shows the concept of the BMBR. Air diffuser is placed inside the inserted baffles. In the operation of the BMBR, membrane filtration is carried out in the constant flow rate mode and the flow rate of the raw wastewater feed is larger than that of filtration. When the water level is higher than the top of the baffles (Fig. 1a), the entire reactor is vigorously mixed by the aeration and kept aerobic. Addition of the raw wastewater is stopped when the water level reaches the set highest level. Then, the water level goes down due to membrane filtration and is eventually lowered below the top of the inserted baffles. From this point, the reactor is separated by the baffles and discriminated into two zones. The outside of the baffles should become anoxic due to oxygen consumption by biomass while the inside of the baffles should be kept aerobic due to the aeration (Fig. 1b). When the water level reaches the set lowest level, addition of the raw water is restarted and the water level is allowed to rise. The addition of raw wastewater must be done to the outside of the baffles so that organic carbon contained in the raw wastewater can be utilized for denitrification. Eventually, the water level exceeds the top of the inserted baffles and the outer zone should become aerobic again due to vigorous mixing provided by the aeration (Fig. 1a). Thus, in the operation of the BMBR, aerobic and anoxic conditions are alternatively created in the outer zone at a constant interval and therefore the improvement in nitrogen removal is expected in comparison to normal submerged MBRs, which are principally aerobic. 2. Experimental Apparatus

Continuous operation of a pilot-scale BMBR was conducted at an existing municipal wastewater treatment facility (Soseigawa Wastewater Treatment Plant, Sapporo, Japan). Temperature of the raw wastewater sometimes decreased below 10 °C during winter season, making biological treatment (especially nitrification) difficult. The BMBR used in this study equipped 6.0 m2 of flat-sheet type of micro-filtration membranes (Toray, Japan). Nominal pore size and material of the membrane were 0.1 µm and polyvinylidenfluoride, respectively. Effective volume of the BMBR was 500 L, in which the ratio of the outer zone to the inner zone was approximately 2.2 when the water level was at the top of the inserted baffles. Aeration rate was fixed at 110 L min-1. As a result, dissolved oxygen concentration inside the baffles was always maintained above 5 mg L-1. The authors have been proposing “the hybrid MBR” which is composed of pre-coagulation/sedimentation and a MBR [7]. By carrying out the pre-treatment, enhanced

Fig. 1. Concept of the baffled membrane bioreactor BMBR

removal of organic matter and phosphorus, and mitigation of membrane fouling can be achieved. Firstly, the BMBR was examined as a hybrid MBR. Namely, the wastewater treated by coagulation and sedimentation processes was introduced to the BMBR as the feed water. Iron-based coagulant, poly-silicato iron (PSI) [8], was used as a coagulant. Dose of PSI was fixed at 10 mg-Fe L-1. Subsequently, the pre-treatment was abandoned and the BMBR directly treated the raw wastewater. 3. Operating Conditions

In this study, the constant flow rate mode of filtration was employed. Therefore, required transmembrane pressure (TMP) difference increased as the operation period became longer. Membrane flux was fixed at 0.4 m3 m-2 d-1 in all operations. Hydraulic retention time (HRT) in the reactor was 5.3 h. Intermittent operation of the suction pump (15 min operation and 1 min pause) was also carried out. The first continuous operation was initiated on Nov. 4, 2003 after acclimatization of biomass was confirmed. This operation was continued for approximately three months. The experimental period was divided into two phases (Runs 1 and 2). In Run 1, from Nov. 4, 2003 to Dec. 5, 2003, existence of the optimum operation time per cycle (OTPC) was explored with a fixed mixed liquor suspended solids (MLSS) concentration of 10,000 mg L-1. In Run 2, from Dec. 5, 2003, MLSS concentration was allowed to rise by abandoning sludge extraction. As a result, MLSS concentration exceeded 20,000 mg L-1 in Run 2. The next run (Run 3) was carried out without the pre-treatment, as stated above. The operation of Run 3 was initiated on Sep. 13, 2005 after acclimatization of biomass. MLSS concentration was fixed at 20,000 mg L-1 in Run 3 and the optimum OTPC for the different condition was again explored. RESULTS AND DISCUSSION 1. Optimum OTPC for the BMBR

In the operation of the proposed BMBR, the length of time for which the outside of the baffles is

Watanabe and Kimura: A Baffled MBR for N-Removal

Fig. 2. Relationship between the OTPC and nitrogen concentration in the permeate.

anoxic (aerobic) is obviously influential on the performance of the reactor. OTPC is defined herein as the time length that is required to complete one operation cycle. In the first phase of the continuous operation (Run 1), the optimum OTPC that could maximize nitrogen removal efficiency was explored. TN concentration was obviously reduced by the BMBR process in the continuous operation of Run 1, demonstrating that denitrification in the reactor was significant. Figure 2 shows the relationship between the OTPC and concentrations of TN and NH4+-N in the permeate. Concentration of NH4+-N in the permeate was always maintained below 0.5 mg L-1 regardless of the length of the OTPC while TN concentration in the permeate was apparently affected by the OTPC. Under the tested condition, the optimum OTPC seemed to be around 30 min. During Run 1, increase in TMP was minimal (data not shown) and consequently membrane filtration could be stably carried out.

437

sent study, increase in TMP was not significant even after the abandonment of sludge extraction and membrane filtration could be continued stably (data not shown). This stable membrane filtration could be partly attributed to the use of flat sheet type of membrane, which is free from inter-fiber clogging. Another possible explanation might be that creation of alternative aerobic/anoxic conditions would be beneficial to maintain good filterability of mixed liquor. As shown before, in Run 1, concentration of TN in the permeate could be reduced to around 10 mg-N L-1 under the best operational condition (i.e., OTPC of 29 min). In Run 2, where MLSS concentration was allowed to increase above 20,000 mg L-1, however, better reactor performance compared to that seen in Run 1 was found. Figure 3 shows change in TN concentration observed in Run 2. Although the water temperature was lower in Run 2 than was recorded in Run 1 (i.e., around 15 °C), TN concentration could be decreased as low as 6 mg-N L-1 in Run 2. Regarding organic carbon and phosphorus removal, the BMBR showed a very good performance throughout Run 2. Table 1 shows average water quality of the feed water and the permeate observed in Run 2 in terms of TOC, total phosphorus (T-P) and NH4+-N. As shown in Table 1, nitrification was almost completed in Run 2. 3. Performance of the BMBR without pre-treatment

In the experiments described above, the raw wastewater was delivered to the BMBR after implementation of coagulation and sedimentation. By implementing pre-coagulation/sedimentation, membrane fouling

2. Performance of the BMBR with High MLSS Concentration

In operations of MBRs, maintaining extremely high MLSS concentration is not necessarily good since it often causes severe membrane fouling. To avoid severe membrane fouling, MLSS concentration was kept around 10,000 mg L-1 in Run 1. However, judging from the extremely slow increase in TMP observed in Run 1, it was likely possible to carry out a stable membrane filtration with the BMBR even at a higher MLSS concentration. Therefore, extraction of sludge was abandoned on Dec. 1 in 2003 and MLSS concentration in the membrane chamber was allowed to increase (Run 2). The OTPC was fixed at 29 min in Run 2. As a result of the abandonment of sludge extraction, MLSS concentration started to increase and eventually exceeded 20,000 mg L-1 at which the MBRs with hollow-fiber membrane expressed difficulty in filtering the mixed liquor [9-10]. In the pre-

Fig. 3. Removal of total nitrogen by the BMBR in Run 3. Table 1. Reduction of TOC, T-P and NH4+-N by the BMBR in Run 2* TOC (mg L-1) TP (mg L-1) NH4+-N (mg L-1)

* Averaged values.

Feed 46.2 0.67 16.1

Permeate 3.9 0.07 0.57

438

J. Environ. Eng. Manage., 16(6), 435-439 (2006)

Fig. 4. Removal of total nitrogen by the BMBR in Run 3.

in the subsequent MBR could be mitigated and removal of phosphorus could be enhanced. As a result of the pre-treatment, however, a significant amount of organic matter was removed from the raw wastewater and consequently the amount of organic carbon available for denitrification in the BMBR was limited. In Run 3, the performance of the BMBR was examined using the raw municipal wastewater (primary clarifier influent) without pre-treatment. Changes in concentrations of TN in the raw water and treated water in Run 3 are shown in Fig. 4. In Run 3, the optimum OTPC was likely to differ from the one found in the previous runs because the composition of the feed water of the BMBR was considerably different. To find out the optimum OTPC without the pre-treatment, the OTPC was increased from 10 to 50 min by increments of 10 min. On day 69, the OTPC was returned to 10 min. As shown in Fig. 4, a significant difference in TN concentrations between the raw water and treated water was found again. In Run 3, TN concentration in the raw water gradually declined and that in the treated water gradually increased when the OTPC was increased in the step-wise manner. With OTPC of 50 min, the rate of removal of TN by the reactor was limited to 65%. After the OTPC was shortened from 50 to 10 min (day 69), the concentration of TN in the treated water was reduced to < 1 mg L-1. At that time, nitrification was almost completed. These -observations clearly demonstrate that the treatment performance of the BMBR was influenced by the length of the OTPC. After day 69, measurement of total phosphorus (TP) was carried out for the raw wastewater and the treated water. Averaged concentrations of TP in the raw wastewater and the treated water (n = 4) were 4.6 and 0.17 mg L-1, respectively. Biologically enhanced phosphorus removal might occur in the operation of the BMBR, but this was not confirmed in this study. Another possible explanation for the good removal of phosphorus is precipitation with inorganic substances. Aggregates of phosphorus and inorganic substances

would settle in dead zones of the bottom of MBRs and could not be found by effluent analysis [5]. TOC concentration in the treated water was fairly constant (average: 3.6 mg L-1; standard deviation: 0.51 mg L-1; n = 38) throughout the operation, although that in the raw wastewater considerably fluctuated (average: 43 mg L-1; standard deviation: 19 mg L-1; n = 38). Removal of organic carbon was not affected by the OTPC. Thus, it was demonstrated that the proposed BMBR could carry out very efficient removal of organic carbon, nitrogen, and phosphorus as long as appropriate operating conditions were provided. Table 2 presents a comparison of the performance of the proposed BMBR with previously reported performances of other MBRs with which an additional anoxic tank was combined. All of these study used real municipal wastewaters. It is apparent that the performance of the proposed BMBR is superior to the performances of other MBRs despite a much shorter HRT. Figure 5 shows time course change in TMP observed in Run 3. Despite the fact that the pretreatment was abandoned in Run 3, increase in TMP was very slow again except for the period when the OTPC was set at 50 min. TMP suddenly started to riseafter the OTPC was set at 50 min, and then the rate of increase in TMP became low immediately after the OTPC was set at 10 min. When the OTPC was set at 50 min, inside of the BMBR became excessively reTable 2. Performance comparison with other reported MBRs

Ueda and Hata [4] Côté et al. [3] Ahn et al. [11] This study*** * Removal rate. ** Values of COD. *** OTPC of 10 min.

HRT (h) 13.4 9 8 5.3

TOC* (%) 93 98** ~95** 92

TN* (%) 79 80 60 95

TP* (%) 74 15 93 96

Fig. 5. Time course change in TMP observed in Run 3.

Watanabe and Kimura: A Baffled MBR for N-Removal

439

REFERENCES

Fig. 6. Changes in concentrations of colloidal protein/ carbohydrates in the BMBR in Run 3.

ductive and therefore the decay of biomass was very likely. Figure 6 shows changes in concentrations of colloidal protein/carbohydrates in the reactor. As seen in Fig. 6, concentrations of colloidal (i.e., supernatant after centrifuging at 3,000 rpm for 5 min) carbohydrate and protein in the mixed liquor suspension suddenly increased when the OTPC was changed to 50 min, implying that the decay of biomass occurred at that time. The period when an increase in colloidal protein/carbohydrate was observed corresponds to the period when increase in TMP was rapid. To avoid membrane fouling in the operation of the BMBR, it is important to set operating conditions so as not to cause decay of biomass. SUMMARY Insertion of baffles into a submerged MBR can significantly improve the performance of the reactor. In this study, the performance of the proposed BMBR was examined based on a pilot-scale study. Important findings obtained in this study can be summarized as follows: The optimum OTPC that maximizes denitrification performance of the BMBR does exist. With the optimum OTPC, anoxic condition can efficiently be created outside the inserted baffles, which leads to good denitrification. Under the conditions examined in this study, TN concentration in real municipal wastewater could be reduced to < 1 mg L-1 without any external organic carbon with HRT of 5.3 h. Very efficient removal of organic carbon and phosphorus was also found. Increase in TMP in the operation of the BMBR was generally very slow. Alternative creation of aerobic/anoxic conditions and/or the use of a flat-sheet type of membrane might be beneficial for maintaining high the membrane permeability.

1. Stephenson, T., S. Judd, B. Jefferson and K. Brindle, Membrane Bioreactors for Wastewater Treatment. IWA Publishing, London, UK (2000). 2. Yamamoto, K., M. Hiasa, T. Mahood and T. Matsuo, Direct solid-liquid separation using hollow fiber membrane in an activated sludge aeration tank. Water Sci. Technol., 21(4-5), 43-54 (1989). 3. Côté, P., H. Buisson, C. Pound and G. Arakaki, Immersed membrane activated sludge for the reuse of municipal wastewater. Desalination, 113(2-3), 189-196 (1997). 4. Ueda, T. and K. Hata, Domestic wastewater treatment by a submerged membrane bioreactor with gravitational filtration. Water Res., 33(12), 2888-2892 (1999). 5. Rosenberger, S., U. Kruger, R. Witzig, W. Manz, U. Szewzyk and M. Krume, Performance of a bioreactor with submerged membranes for aerobic treatment of municipal waste water. Water Res., 36(2), 413-420 (2002). 6. Kimura, K. and Y. Watanabe, Baffled membrane bioreactor (BMBR) for advanced wastewater treatment: Easy modification of existing MBRs for efficient nutrient removal. Water Sci. Technol., 52(10-11), 427-434 (2005). 7. Watanabe, Y. and K. Kimura, Hybrid membrane bioreactor for water recycling and phosphorus recovery. Water Sci. Technol., 53(7), 17-24 (2006). 8. Hasegawa, T., K. Hashimoto and N. Tambo, Characteristics of metal-polysilicate coagulants. Water Sci. Technol., 23(7-9), 1713-1722 (1991). 9. Itonaga, T. and Y. Watanabe, Performance of membrane bioreactor combined with precoagulation/sedimentation. Water Sci. Technol., 4(1), 143-149 (2004). 10. Itonaga, T., K. Kimura and Y. Watanabe, Influence of suspension viscosity and colloidal particles on permeability of membrane used in membrane bioreactor (MBR). Water Sci. Technol., 50(12), 301-309 (2004). 11. Ahn, K.H., K.G. Song, E. Cho, J. Cho, H. Yun, S. Lee and J. Kim, Enhanced biological phosphorus and nitrogen removal using a sequencing anoxic/anaerobic membrane bioreactor (SAM) process. Desalination, 157(1-3), 345-352 (2003). Discussions of this paper may appear in the discussion section of a future issue. All discussions should be submitted to the Editor-in-Chief within six months of publication. Manuscript Received: September 12, 2006 Revision Received: October 29, 2006 and Accepted: November 7, 2006

Suggest Documents