- Email: [email protected]
Performance evaluation of a field-scale pilot bioreactor for anaerobic treatment of palm oil mill effluent
International Biodeterioration & Biodegradation xxx (2014) 1e4
Contents lists available at ScienceDirect
International Biodeterioration & Biodegradation journal homepage: www.elsevier.com/locate/ibiod
Performance evaluation of a field-scale pilot bioreactor for anaerobic treatment of palm oil mill effluent Ja-Won Shin, Jun-Hyeon Pyeon, Sung-Min Son, Joo-Young Jeong, Joo-Yang Park* Department of Civil and Environmental Engineering, Hanyang University, 17 Haengdang-dong, Seongdong-gu, Seoul 133-791, Republic of Korea
a r t i c l e i n f o
a b s t r a c t
Article history: Received 15 January 2014 Received in revised form 1 May 2014 Accepted 3 May 2014 Available online xxx
A combined system of anaerobic reactors was applied in a field-scale pilot bioreactor to treat palm oil mill effluent (POME) from the Amagra palm plantation in Sumatra Island, Indonesia. The initial start-up of this system failed due to shock loads of suspended solids, organics, and flow rates. After adjusting for influent conditions and operation variables, the second start-up successfully proceeded, where 70% of chemical oxygen demand (COD) was removed at 6 kg m3 day1 of the organic loading rate. Nutrients such as N or P were not added. An addition of alkalinity was made at the initial start-up and later found to be not necessary because amino and fatty acids were rapidly removed in this high rate system. Ó 2014 Elsevier Ltd. All rights reserved.
Keywords: Anaerobic treatment Hybrid reactor Combined system Start-up Palm oil mill effluent Field-scale
1. Introduction Palm oil is a valuable resource in tropical areas, particularly in Indonesia and Malaysia. Approximately 80% of palm oil is used for food purposes. Over the past three decades, there has been a significant growth in the global consumption of vegetable oil. Palm oil production involves the generation of a huge amount of highly concentrated organic wastewater called palm oil mill effluent (POME). Treatment of POME still depends greatly upon a conventional pond system that consists of anaerobic and facultative ponds followed by polishing ponds. Open digesting tanks are also frequently adopted due to their convenience in removing scum and bottom sludge (Najafpour et al., 2006). In both methods, methane gas is released directly to the atmosphere, which produces serious concerns for global warming due to POME (Yacob et al., 2005). In addition, the infiltration of untreated POME of high organic content into surrounding water has caused serious water pollution problems (Wu et al., 2010). Closed anaerobic digesters are very effective in preventing greenhouse gas emissions. There has been great improvement in developing high rate anaerobic reactors. The most successful processes are upflow anaerobic sludge blanket (UASB) reactors and anaerobic filters (AF). The disadvantages of these reactors are washout of sludge in UASB and clogging in AF (Wu et al., 2000; Bodkhe, * Corresponding author. Tel.: þ82 2 2220 0411; fax: þ82 2 2220 1945. E-mail address: [email protected] (J.-Y. Park).
2008). Hybrid reactors have also been developed, combining the advantages of both reactors. Recently, a combined system of an anaerobic hybrid reactor (AHR) with an anaerobic baffled filter (ABF) was developed. A laboratory pilot of this system was used to investigate its applicability for POME treatment and proved its effectiveness and stability in treating highly concentrated organic waste (Choi et al., 2013). One of the drawbacks in anaerobic treatment is the difficulty of start-up (Yacob et al., 2006). Start-up of an anaerobic reactor takes a longer time due to the low growth yield of microorganisms. Recently, research has been done on the anaerobic digestion of POME. However, only a few studies are available for the field application of high rate anaerobic reactors to POME. In this study, a combined system of AHR and ABF was scaled up to a field-scale pilot bioreactor to treat 16 m3 day1 of 55000 mg L1 of POME. The field pilot was constructed in a palm oil mill plant in Sumatra Island, Indonesia. The objective of this study was to investigate the performance of the combined system at a field scale for anaerobic high rate treatment of POME, especially at the start-up period.
2. Materials and methods 2.1. Palm oil mill effluent Raw POME from the Amagra palm plantation in Sumatra Island, Indonesia was examined to contain 4e5% of total solids (TS), 2e4%
http://dx.doi.org/10.1016/j.ibiod.2014.05.005 0964-8305/Ó 2014 Elsevier Ltd. All rights reserved.
Please cite this article in press as: Shin, J.-W., et al., Performance evaluation of a field-scale pilot bioreactor for anaerobic treatment of palm oil mill effluent, International Biodeterioration & Biodegradation (2014), http://dx.doi.org/10.1016/j.ibiod.2014.05.005
2
J.-W. Shin et al. / International Biodeterioration & Biodegradation xxx (2014) 1e4 Table 1 Characteristics of the POME used in this study. Parameters
Value
pH COD (mg L1) TSS (mg L1) VSS (mg L1) T-N (mg L1) T-P (mg L1)
4.0e5.6 32,300e63,200 2,060e24,720 2,040e20,360 380e880 182e1,500
suspended solids (SS), 0.6e0.7% of residual oil and 95e96% of water. High concentrations of SS and oil can cause serious operational problems such as excessive scum build-up and clogging in high rate anaerobic treatment processes (Latif et al., 2011). The raw POME was pretreated using a three-phase screw decanter. Table 1 shows the characteristics of the pretreated POME. Further treatment using air flotation was also performed to prevent scum build-up. When the pH in AHR decreased below 5.5, NaHCO3 was added at the rate of 250 mg L1 at the beginning of start-up. No nutrients such as N or P were added. 2.2. Reactor design and configuration A schematic diagram of the field-scale reactor used in this research is shown in Fig. 1. Anaerobic processes consist of a primary reactor (AHR) and secondary reactor (ABF). The AHR was made of reinforced concrete with internal epoxy coating to have an octagonal pillar shape, of which the volume was 60 m3 (height, 8.6 m; cross sectional area, 7 m2). The upper zone of the AHR was packed with Tri-Pack media up to a depth of 2.6 m from the top (polyethylene, Solmaro Trading & Engineering Company, South Korea; diameter, 3.5 in; specific gravity, 0.95; specific area, 38 ft2 ft3; porosity, 95%). The surge tank was placed just after the primary reactor to recycle the effluent. Recycling for AHR was intended to improve buffering capacities for pH and shock loads. The second reactor was made in the same manner to have half of the volume of the primary reactor (volume, 30 m3; height, 7.64 m; cross sectional area, 4 m2). The ABF was designed to have a longer
Fig. 2. A picture of the field pilot located in Riau province of Sumatra Island, Indonesia.
flow path using a vertical baffle. The ABF was packed with the same Tri-Pack from the top to a depth of 5.32 m (70%, H H1). Generated biogas flowed through water traps to prevent backflow of gas. The height of the trap was about 50 cm, and then biogas flowed to a dry desulfurizer. The desulfurizer was filled with a powdered activated carbon to adsorb H2S gas generated from the reactors. A front view of the high rate anaerobic reactors (HRARs) system in Indonesia is shown in Fig. 2. 2.3. Start-up and experimental methods The reactors were operated at mesophilic temperatures between 32 and 37 C. Organic loading rates (OLR) to the reactors were altered by changing the flow rate (Table 2). A step at the same OLR lasted from 6 to 55 days. The intended OLR at the same step was not maintained due to fluctuations in water quality of the POME from the field. A slurry of mesophilic anaerobic granules from a UASB reactor in a brewery plant in South Korea was seeded as 5 m3 for AHR and 1 m3 for ABF. Chemical oxygen demand (COD) concentrations were measured using the reactor digestion method (HACH, method 8000, range: 20e1500 mg L1). Total suspended solids (TSS) and volatile suspended solids (VSS) were measured
Fig. 1. Schematic of the field pilot, showing the anaerobic reactors combined with AHR and ABF to treat POME.
Please cite this article in press as: Shin, J.-W., et al., Performance evaluation of a field-scale pilot bioreactor for anaerobic treatment of palm oil mill effluent, International Biodeterioration & Biodegradation (2014), http://dx.doi.org/10.1016/j.ibiod.2014.05.005
J.-W. Shin et al. / International Biodeterioration & Biodegradation xxx (2014) 1e4 Table 2 Organic loading rate of each step. Stage
OLR step
OLR (kg m3 day1)
1
1 2 3 4 5 6 7 8 9
2.3 7.0 4.0 0.3 0.6 2.1 2.5 6.3 5.4
2
using Standard Methods, 20th Edition (APHA et al., 1998). Biogas generation was measured using a gas counter in the field pilot. 3. Results and discussion 3.1. Initial start-up Fig. 3 shows the profiles of (a) COD, (b) pH, (c) TSS and (d) VSS during the entire operation period. OLRs were incrementally increased as shown in Table 2. There were nine different OLR steps. OLR steps 1, 2 and 3 belong to the initial stage of start-up. The second stage of start-up includes the remaining steps. The initial start-up progressed well. As shown in Fig. 3, COD was removed considerably and pH was maintained at 7.3e8.5 until step 2.
3
However, at the beginning of OLR step 3, POME of very high COD concentration was introduced into the reactor. This caused an organic overload in the reactor, thereby causing a significant pH drop to 4.4 due to excessive accumulation of volatile fatty acids (VFAs). After that overload, effluent COD increased rapidly although the pH increased with the addition of soda ash. At the end of step 3, COD was virtually not removed. Initial start-up ended up with a failure not only due to the high organic load but to other reasons. SS concentrations were higher than expected and designed for (2000 mg L1). This made the reactors fill with SS and scum, preventing soluble organics from accessing granules. High SS concentrations in influent can cause adverse effects on the performance of a reactor at a high OLR by reducing microbial activities (Poh and Chong, 2009). At the beginning of OLR step 1, TSS concentrations in the reactors were very low (below 1000 mg L1) because of freshwater remaining in the reactors. As incoming TSS to the primary reactor increased rapidly at OLR step 2, TSS in the effluent of the AHR increased rapidly as well, indicating a shock load of TSS. The other reason was related with the recycle ratio of the primary reactor, which is the ratio of the recycle to influent flow rate. The recycle ratio was 30 and the recycle flow rate reached at 100 m3 day1 at the initial stage, which was later found to be too high. Recycling was intended to reduce shock loads of organics, distribute accumulated VFAs and provide alkalinity for freshly introduced influent (Zinatizadeh et al., 2010). However, excessive recycling is known to fracture granules due to excessive collisions
Fig. 3. Profiles of (a) COD, (b) pH, (c) TSS and (d) VSS. “A” indicates the period that reactors were cleaned after hydraulic shock and “B” indicates the period that reactors were just recycled without feed after organic shock load.
Please cite this article in press as: Shin, J.-W., et al., Performance evaluation of a field-scale pilot bioreactor for anaerobic treatment of palm oil mill effluent, International Biodeterioration & Biodegradation (2014), http://dx.doi.org/10.1016/j.ibiod.2014.05.005
4
J.-W. Shin et al. / International Biodeterioration & Biodegradation xxx (2014) 1e4
among them (Zhang et al., 2009). In our case, this excessive recycling resulted in wash-out of the granules from the primary reactor (AHR). This was identified after shut down. An excessive amount of granules was found in the secondary reactor (ABF) and later recovered for seed in the secondary start-up. 3.2. Secondary start-up and operation After the shutdown due to shock loads, the reactors were cleaned with fresh water to remove scum and interstitial SS between media. Then the second start-up was started with careful measures. TSS concentrations in influent POME were kept as low as 2000 mg L1. Fluctuations in COD and SS concentrations were greatly reduced with an additional holding capacity of raw POME. Recycle flow rate was adjusted to 3 m3 h1 to obtain a recycle ratio of 7e8. OLR was increased carefully from 2 to 7 kg m3 day1 (step 4e8). More than 80% of COD removal was obtained from step 4e7. The pH in the AHR effluent was maintained at 6.5e7.0 although no soda ash was added. However, at the end of step 8, an additional shock load of organics was experienced with a pH drop and lower COD removal. Therefore, we decided to stop OLR step 8 and to reduce the OLR to 6 kg m3 day1 after a period of recovery. Bajaj et al. (2009) reported that recovery of a reactor was possible only when no feed was provided to the reactor. Accordingly, in this stage, feeding was also stopped for the next two weeks. OLR step 9 was then started at a reduced load of 6 kg m3 day1. The COD concentrations were reduced to 17000 mg L1 in AHR and to 9800 mg L1 in ABF effluents. The recovery run was successful. It showed relatively stable effluent water qualities as shown in Fig. 3. Around 50% of COD was removed in AHR and additional COD removal was obtained in ABF. In total, 70% of COD was removed by the combined system. During the secondary start-up, pH in the anaerobic reactors was maintained at a desirable level even though no soda ash was added. This is mainly due to the rapid degradation of amino and fatty acids dissolved in POME and also may be due to recycling of effluent. Recycling is helpful to maintain the pH of anaerobic reactors (Sam-Soon et al., 1991; Najafpour et al., 2006). Faisal and Unno (2001) reported that recycling was required to maintain a pH higher than 6.8 without alkalinity supplementation. As shown in Fig. 3, although the pH of AHR effluent decreased rapidly to 5.8, the pH of ABF effluent was maintained in an optimum region because organic shock load only occurred in the primary reactor. Although organic shock load happened at step 8, TSS did not change significantly. The SS removal seems to be more stable in ABF than in AHR. The ABF, with a higher packing ratio, was able to perform a good role as a filter to prevent the washout of sludge from the primary reactor. Thus, it is considered that the high packing ratio of the secondary reactor is effective.
4. Conclusions Start-up of a combined system of anaerobic reactors was investigated. Initial start-up failed due to high organic and hydraulic shock loads. The second start-up was successful with careful controls of incoming COD and SS fluctuations. So far, the field pilot can remove 70% of COD at an OLR of 6 kg m3 day1. Additional operation will provide the efficiency at an optimum OLR. The addition of nutrients and alkalinity was not required in this anaerobic treatment of POME. Acknowledgements This study was financially supported by the “Center for Waste Eco-Energy and Non-CO2 Greenhouse Gases (CWEG)” as an EcoSTAR Project by the Ministry of Environment, Korea (No. 10-01021). References APHA, AWWA, WPCF, 1998. Standard Methods for Examination of Water and Wastewater. American Public Health Association, Washington, DC. Bajaj, M., Gallert, C., Winter, J., 2009. Treatment of phenolic wastewater in an anaerobic fixed bed reactor (AFBR)-Recovery after shock loading. J. Hazard. Mater. 162(2), 1330e1339. Bodkhe, S., 2008. Development of an improved anaerobic filter for municipal wastewater treatment. Bioresour. Technol. 99(1), 222e226. Choi, W.H., Shin, C.H., Son, S.M., Ghorpade, P.A., Kim, J.J., Park, J.Y., 2013. Anaerobic treatment of palm oil mill effluent using combined high rate anaerobic reactors. Bioresour. Technol. 141, 138e144. Faisal, M., Unno, H., 2001. Kinetic analysis of palm oil mill wastewater treatment by a modified anaerobic baffled reactor. Biochem. Eng. J. 9(1), 25e31. Latif, M.A., Ghufran, R., Wahid, Z.A., Ahmad, A., 2011. Integrated application of upflow anaerobic sludge blanket reactor for the treatment of wastewaters. Water Res. 45(16), 4683e4699. Najafpour, G.D., Zinatizadeh, A.A.L., Mohamed, A.R., Hasnain Isa, M., Nasrollahzadeh, H., 2006. High-rate anaerobic digestion of palm oil mill effluent in an upflow anaerobic sludge-fixed film bioreactor. Process Biochem. 41(2), 370e379. Poh, P.E., Chong, M.F., 2009. Development of anaerobic digestion methods for palm oil mill effluent (POME) treatment. Bioresour. Technol. 100(1), 1e9. Sam-Soon, P.A., Wentzel, M.C., Moosbrugger, R.E., Marais, G., Loewenthal, R.E., 1991. Effects of a recycle in upflow anaerobic sludge bed (UASB) systems. Water SA 17(1), 37e46. Wu, M., Wilson, F., Tay, J.H., 2000. Influence of media-packing ratio on performance of anaerobic hybrid reactors. Bioresour. Technol. 71(2), 151e157. Wu, T.Y., Mohammad, A.W., Jahim, J.M., Anuar, N., 2010. Pollution control technologies for the treatment of palm oil mill effluent (POME) through end-of-pipe processes. J. Environ. Manag. 91(7), 1467e1490. Yacob, S., Ali Hassan, M., Shirai, Y., Wakisaka, M., Subash, S., 2005. Baseline study of methane emission from open digesting tanks of palm oil mill effluent treatment. Chemosphere 59(11), 1575e1581. Yacob, S., Ali Hassan, M., Shirai, Y., Wakisaka, M., Subash, S., 2006. Baseline study of methane emission from anaerobic ponds of palm oil mill effluent treatment. Sci. Total Environ. 366(1), 187e196. Zhang, Y., Ma, Y., Quan, X., Jing, Y., Dai, S., 2009. Rapid startup of a hybrid UASB-AFF reactor using bi-circulation. Chem. Eng. J. 155(1e2), 266e271. Zinatizadeh, A.A.L., Pirsaheb, M., Bonakdari, H., Younesi, H., 2010. Response surface analysis of effects of hydraulic retention time and influent feed concentration on performance of an UASFF bioreactor. Waste Manag. 30(10), 1798e1807.
Please cite this article in press as: Shin, J.-W., et al., Performance evaluation of a field-scale pilot bioreactor for anaerobic treatment of palm oil mill effluent, International Biodeterioration & Biodegradation (2014), http://dx.doi.org/10.1016/j.ibiod.2014.05.005
Contents lists available at ScienceDirect
International Biodeterioration & Biodegradation journal homepage: www.elsevier.com/locate/ibiod
Performance evaluation of a field-scale pilot bioreactor for anaerobic treatment of palm oil mill effluent Ja-Won Shin, Jun-Hyeon Pyeon, Sung-Min Son, Joo-Young Jeong, Joo-Yang Park* Department of Civil and Environmental Engineering, Hanyang University, 17 Haengdang-dong, Seongdong-gu, Seoul 133-791, Republic of Korea
a r t i c l e i n f o
a b s t r a c t
Article history: Received 15 January 2014 Received in revised form 1 May 2014 Accepted 3 May 2014 Available online xxx
A combined system of anaerobic reactors was applied in a field-scale pilot bioreactor to treat palm oil mill effluent (POME) from the Amagra palm plantation in Sumatra Island, Indonesia. The initial start-up of this system failed due to shock loads of suspended solids, organics, and flow rates. After adjusting for influent conditions and operation variables, the second start-up successfully proceeded, where 70% of chemical oxygen demand (COD) was removed at 6 kg m3 day1 of the organic loading rate. Nutrients such as N or P were not added. An addition of alkalinity was made at the initial start-up and later found to be not necessary because amino and fatty acids were rapidly removed in this high rate system. Ó 2014 Elsevier Ltd. All rights reserved.
Keywords: Anaerobic treatment Hybrid reactor Combined system Start-up Palm oil mill effluent Field-scale
1. Introduction Palm oil is a valuable resource in tropical areas, particularly in Indonesia and Malaysia. Approximately 80% of palm oil is used for food purposes. Over the past three decades, there has been a significant growth in the global consumption of vegetable oil. Palm oil production involves the generation of a huge amount of highly concentrated organic wastewater called palm oil mill effluent (POME). Treatment of POME still depends greatly upon a conventional pond system that consists of anaerobic and facultative ponds followed by polishing ponds. Open digesting tanks are also frequently adopted due to their convenience in removing scum and bottom sludge (Najafpour et al., 2006). In both methods, methane gas is released directly to the atmosphere, which produces serious concerns for global warming due to POME (Yacob et al., 2005). In addition, the infiltration of untreated POME of high organic content into surrounding water has caused serious water pollution problems (Wu et al., 2010). Closed anaerobic digesters are very effective in preventing greenhouse gas emissions. There has been great improvement in developing high rate anaerobic reactors. The most successful processes are upflow anaerobic sludge blanket (UASB) reactors and anaerobic filters (AF). The disadvantages of these reactors are washout of sludge in UASB and clogging in AF (Wu et al., 2000; Bodkhe, * Corresponding author. Tel.: þ82 2 2220 0411; fax: þ82 2 2220 1945. E-mail address: [email protected] (J.-Y. Park).
2008). Hybrid reactors have also been developed, combining the advantages of both reactors. Recently, a combined system of an anaerobic hybrid reactor (AHR) with an anaerobic baffled filter (ABF) was developed. A laboratory pilot of this system was used to investigate its applicability for POME treatment and proved its effectiveness and stability in treating highly concentrated organic waste (Choi et al., 2013). One of the drawbacks in anaerobic treatment is the difficulty of start-up (Yacob et al., 2006). Start-up of an anaerobic reactor takes a longer time due to the low growth yield of microorganisms. Recently, research has been done on the anaerobic digestion of POME. However, only a few studies are available for the field application of high rate anaerobic reactors to POME. In this study, a combined system of AHR and ABF was scaled up to a field-scale pilot bioreactor to treat 16 m3 day1 of 55000 mg L1 of POME. The field pilot was constructed in a palm oil mill plant in Sumatra Island, Indonesia. The objective of this study was to investigate the performance of the combined system at a field scale for anaerobic high rate treatment of POME, especially at the start-up period.
2. Materials and methods 2.1. Palm oil mill effluent Raw POME from the Amagra palm plantation in Sumatra Island, Indonesia was examined to contain 4e5% of total solids (TS), 2e4%
http://dx.doi.org/10.1016/j.ibiod.2014.05.005 0964-8305/Ó 2014 Elsevier Ltd. All rights reserved.
Please cite this article in press as: Shin, J.-W., et al., Performance evaluation of a field-scale pilot bioreactor for anaerobic treatment of palm oil mill effluent, International Biodeterioration & Biodegradation (2014), http://dx.doi.org/10.1016/j.ibiod.2014.05.005
2
J.-W. Shin et al. / International Biodeterioration & Biodegradation xxx (2014) 1e4 Table 1 Characteristics of the POME used in this study. Parameters
Value
pH COD (mg L1) TSS (mg L1) VSS (mg L1) T-N (mg L1) T-P (mg L1)
4.0e5.6 32,300e63,200 2,060e24,720 2,040e20,360 380e880 182e1,500
suspended solids (SS), 0.6e0.7% of residual oil and 95e96% of water. High concentrations of SS and oil can cause serious operational problems such as excessive scum build-up and clogging in high rate anaerobic treatment processes (Latif et al., 2011). The raw POME was pretreated using a three-phase screw decanter. Table 1 shows the characteristics of the pretreated POME. Further treatment using air flotation was also performed to prevent scum build-up. When the pH in AHR decreased below 5.5, NaHCO3 was added at the rate of 250 mg L1 at the beginning of start-up. No nutrients such as N or P were added. 2.2. Reactor design and configuration A schematic diagram of the field-scale reactor used in this research is shown in Fig. 1. Anaerobic processes consist of a primary reactor (AHR) and secondary reactor (ABF). The AHR was made of reinforced concrete with internal epoxy coating to have an octagonal pillar shape, of which the volume was 60 m3 (height, 8.6 m; cross sectional area, 7 m2). The upper zone of the AHR was packed with Tri-Pack media up to a depth of 2.6 m from the top (polyethylene, Solmaro Trading & Engineering Company, South Korea; diameter, 3.5 in; specific gravity, 0.95; specific area, 38 ft2 ft3; porosity, 95%). The surge tank was placed just after the primary reactor to recycle the effluent. Recycling for AHR was intended to improve buffering capacities for pH and shock loads. The second reactor was made in the same manner to have half of the volume of the primary reactor (volume, 30 m3; height, 7.64 m; cross sectional area, 4 m2). The ABF was designed to have a longer
Fig. 2. A picture of the field pilot located in Riau province of Sumatra Island, Indonesia.
flow path using a vertical baffle. The ABF was packed with the same Tri-Pack from the top to a depth of 5.32 m (70%, H H1). Generated biogas flowed through water traps to prevent backflow of gas. The height of the trap was about 50 cm, and then biogas flowed to a dry desulfurizer. The desulfurizer was filled with a powdered activated carbon to adsorb H2S gas generated from the reactors. A front view of the high rate anaerobic reactors (HRARs) system in Indonesia is shown in Fig. 2. 2.3. Start-up and experimental methods The reactors were operated at mesophilic temperatures between 32 and 37 C. Organic loading rates (OLR) to the reactors were altered by changing the flow rate (Table 2). A step at the same OLR lasted from 6 to 55 days. The intended OLR at the same step was not maintained due to fluctuations in water quality of the POME from the field. A slurry of mesophilic anaerobic granules from a UASB reactor in a brewery plant in South Korea was seeded as 5 m3 for AHR and 1 m3 for ABF. Chemical oxygen demand (COD) concentrations were measured using the reactor digestion method (HACH, method 8000, range: 20e1500 mg L1). Total suspended solids (TSS) and volatile suspended solids (VSS) were measured
Fig. 1. Schematic of the field pilot, showing the anaerobic reactors combined with AHR and ABF to treat POME.
Please cite this article in press as: Shin, J.-W., et al., Performance evaluation of a field-scale pilot bioreactor for anaerobic treatment of palm oil mill effluent, International Biodeterioration & Biodegradation (2014), http://dx.doi.org/10.1016/j.ibiod.2014.05.005
J.-W. Shin et al. / International Biodeterioration & Biodegradation xxx (2014) 1e4 Table 2 Organic loading rate of each step. Stage
OLR step
OLR (kg m3 day1)
1
1 2 3 4 5 6 7 8 9
2.3 7.0 4.0 0.3 0.6 2.1 2.5 6.3 5.4
2
using Standard Methods, 20th Edition (APHA et al., 1998). Biogas generation was measured using a gas counter in the field pilot. 3. Results and discussion 3.1. Initial start-up Fig. 3 shows the profiles of (a) COD, (b) pH, (c) TSS and (d) VSS during the entire operation period. OLRs were incrementally increased as shown in Table 2. There were nine different OLR steps. OLR steps 1, 2 and 3 belong to the initial stage of start-up. The second stage of start-up includes the remaining steps. The initial start-up progressed well. As shown in Fig. 3, COD was removed considerably and pH was maintained at 7.3e8.5 until step 2.
3
However, at the beginning of OLR step 3, POME of very high COD concentration was introduced into the reactor. This caused an organic overload in the reactor, thereby causing a significant pH drop to 4.4 due to excessive accumulation of volatile fatty acids (VFAs). After that overload, effluent COD increased rapidly although the pH increased with the addition of soda ash. At the end of step 3, COD was virtually not removed. Initial start-up ended up with a failure not only due to the high organic load but to other reasons. SS concentrations were higher than expected and designed for (2000 mg L1). This made the reactors fill with SS and scum, preventing soluble organics from accessing granules. High SS concentrations in influent can cause adverse effects on the performance of a reactor at a high OLR by reducing microbial activities (Poh and Chong, 2009). At the beginning of OLR step 1, TSS concentrations in the reactors were very low (below 1000 mg L1) because of freshwater remaining in the reactors. As incoming TSS to the primary reactor increased rapidly at OLR step 2, TSS in the effluent of the AHR increased rapidly as well, indicating a shock load of TSS. The other reason was related with the recycle ratio of the primary reactor, which is the ratio of the recycle to influent flow rate. The recycle ratio was 30 and the recycle flow rate reached at 100 m3 day1 at the initial stage, which was later found to be too high. Recycling was intended to reduce shock loads of organics, distribute accumulated VFAs and provide alkalinity for freshly introduced influent (Zinatizadeh et al., 2010). However, excessive recycling is known to fracture granules due to excessive collisions
Fig. 3. Profiles of (a) COD, (b) pH, (c) TSS and (d) VSS. “A” indicates the period that reactors were cleaned after hydraulic shock and “B” indicates the period that reactors were just recycled without feed after organic shock load.
Please cite this article in press as: Shin, J.-W., et al., Performance evaluation of a field-scale pilot bioreactor for anaerobic treatment of palm oil mill effluent, International Biodeterioration & Biodegradation (2014), http://dx.doi.org/10.1016/j.ibiod.2014.05.005
4
J.-W. Shin et al. / International Biodeterioration & Biodegradation xxx (2014) 1e4
among them (Zhang et al., 2009). In our case, this excessive recycling resulted in wash-out of the granules from the primary reactor (AHR). This was identified after shut down. An excessive amount of granules was found in the secondary reactor (ABF) and later recovered for seed in the secondary start-up. 3.2. Secondary start-up and operation After the shutdown due to shock loads, the reactors were cleaned with fresh water to remove scum and interstitial SS between media. Then the second start-up was started with careful measures. TSS concentrations in influent POME were kept as low as 2000 mg L1. Fluctuations in COD and SS concentrations were greatly reduced with an additional holding capacity of raw POME. Recycle flow rate was adjusted to 3 m3 h1 to obtain a recycle ratio of 7e8. OLR was increased carefully from 2 to 7 kg m3 day1 (step 4e8). More than 80% of COD removal was obtained from step 4e7. The pH in the AHR effluent was maintained at 6.5e7.0 although no soda ash was added. However, at the end of step 8, an additional shock load of organics was experienced with a pH drop and lower COD removal. Therefore, we decided to stop OLR step 8 and to reduce the OLR to 6 kg m3 day1 after a period of recovery. Bajaj et al. (2009) reported that recovery of a reactor was possible only when no feed was provided to the reactor. Accordingly, in this stage, feeding was also stopped for the next two weeks. OLR step 9 was then started at a reduced load of 6 kg m3 day1. The COD concentrations were reduced to 17000 mg L1 in AHR and to 9800 mg L1 in ABF effluents. The recovery run was successful. It showed relatively stable effluent water qualities as shown in Fig. 3. Around 50% of COD was removed in AHR and additional COD removal was obtained in ABF. In total, 70% of COD was removed by the combined system. During the secondary start-up, pH in the anaerobic reactors was maintained at a desirable level even though no soda ash was added. This is mainly due to the rapid degradation of amino and fatty acids dissolved in POME and also may be due to recycling of effluent. Recycling is helpful to maintain the pH of anaerobic reactors (Sam-Soon et al., 1991; Najafpour et al., 2006). Faisal and Unno (2001) reported that recycling was required to maintain a pH higher than 6.8 without alkalinity supplementation. As shown in Fig. 3, although the pH of AHR effluent decreased rapidly to 5.8, the pH of ABF effluent was maintained in an optimum region because organic shock load only occurred in the primary reactor. Although organic shock load happened at step 8, TSS did not change significantly. The SS removal seems to be more stable in ABF than in AHR. The ABF, with a higher packing ratio, was able to perform a good role as a filter to prevent the washout of sludge from the primary reactor. Thus, it is considered that the high packing ratio of the secondary reactor is effective.
4. Conclusions Start-up of a combined system of anaerobic reactors was investigated. Initial start-up failed due to high organic and hydraulic shock loads. The second start-up was successful with careful controls of incoming COD and SS fluctuations. So far, the field pilot can remove 70% of COD at an OLR of 6 kg m3 day1. Additional operation will provide the efficiency at an optimum OLR. The addition of nutrients and alkalinity was not required in this anaerobic treatment of POME. Acknowledgements This study was financially supported by the “Center for Waste Eco-Energy and Non-CO2 Greenhouse Gases (CWEG)” as an EcoSTAR Project by the Ministry of Environment, Korea (No. 10-01021). References APHA, AWWA, WPCF, 1998. Standard Methods for Examination of Water and Wastewater. American Public Health Association, Washington, DC. Bajaj, M., Gallert, C., Winter, J., 2009. Treatment of phenolic wastewater in an anaerobic fixed bed reactor (AFBR)-Recovery after shock loading. J. Hazard. Mater. 162(2), 1330e1339. Bodkhe, S., 2008. Development of an improved anaerobic filter for municipal wastewater treatment. Bioresour. Technol. 99(1), 222e226. Choi, W.H., Shin, C.H., Son, S.M., Ghorpade, P.A., Kim, J.J., Park, J.Y., 2013. Anaerobic treatment of palm oil mill effluent using combined high rate anaerobic reactors. Bioresour. Technol. 141, 138e144. Faisal, M., Unno, H., 2001. Kinetic analysis of palm oil mill wastewater treatment by a modified anaerobic baffled reactor. Biochem. Eng. J. 9(1), 25e31. Latif, M.A., Ghufran, R., Wahid, Z.A., Ahmad, A., 2011. Integrated application of upflow anaerobic sludge blanket reactor for the treatment of wastewaters. Water Res. 45(16), 4683e4699. Najafpour, G.D., Zinatizadeh, A.A.L., Mohamed, A.R., Hasnain Isa, M., Nasrollahzadeh, H., 2006. High-rate anaerobic digestion of palm oil mill effluent in an upflow anaerobic sludge-fixed film bioreactor. Process Biochem. 41(2), 370e379. Poh, P.E., Chong, M.F., 2009. Development of anaerobic digestion methods for palm oil mill effluent (POME) treatment. Bioresour. Technol. 100(1), 1e9. Sam-Soon, P.A., Wentzel, M.C., Moosbrugger, R.E., Marais, G., Loewenthal, R.E., 1991. Effects of a recycle in upflow anaerobic sludge bed (UASB) systems. Water SA 17(1), 37e46. Wu, M., Wilson, F., Tay, J.H., 2000. Influence of media-packing ratio on performance of anaerobic hybrid reactors. Bioresour. Technol. 71(2), 151e157. Wu, T.Y., Mohammad, A.W., Jahim, J.M., Anuar, N., 2010. Pollution control technologies for the treatment of palm oil mill effluent (POME) through end-of-pipe processes. J. Environ. Manag. 91(7), 1467e1490. Yacob, S., Ali Hassan, M., Shirai, Y., Wakisaka, M., Subash, S., 2005. Baseline study of methane emission from open digesting tanks of palm oil mill effluent treatment. Chemosphere 59(11), 1575e1581. Yacob, S., Ali Hassan, M., Shirai, Y., Wakisaka, M., Subash, S., 2006. Baseline study of methane emission from anaerobic ponds of palm oil mill effluent treatment. Sci. Total Environ. 366(1), 187e196. Zhang, Y., Ma, Y., Quan, X., Jing, Y., Dai, S., 2009. Rapid startup of a hybrid UASB-AFF reactor using bi-circulation. Chem. Eng. J. 155(1e2), 266e271. Zinatizadeh, A.A.L., Pirsaheb, M., Bonakdari, H., Younesi, H., 2010. Response surface analysis of effects of hydraulic retention time and influent feed concentration on performance of an UASFF bioreactor. Waste Manag. 30(10), 1798e1807.
Please cite this article in press as: Shin, J.-W., et al., Performance evaluation of a field-scale pilot bioreactor for anaerobic treatment of palm oil mill effluent, International Biodeterioration & Biodegradation (2014), http://dx.doi.org/10.1016/j.ibiod.2014.05.005