Bioresource Technology 193 (2015) 213–218
Contents lists available at ScienceDirect
Bioresource Technology journal homepage: www.elsevier.com/locate/biortech
The effect of iron dosing on reducing waste activated sludge in the oxic-settling-anoxic process Nevin Yagci a,⇑, John T. Novak b, Clifford W. Randall b, Derin Orhon a,c a
Faculty of Civil Engineering, Environmental Engineering Department, Istanbul Technical University, 34469 Istanbul, Turkey Virginia Polytechnic Institute and State University, Department of Civil Engineering, Blacksburg, VA 24061, USA c ENVIS Energy and Environmental Systems R&D Ltd, ITU Arı Teknokent, Arı-1 Building, 34469 Istanbul, Turkey b
h i g h l i g h t s High iron concentrations in the wastewater enhances the sludge reduction. Reduced observed yield and suitable sludge volume index with fast feeding regime. Simultaneously occurring sludge reduction mechanisms.
a r t i c l e
i n f o
Article history: Received 21 April 2015 Received in revised form 21 June 2015 Accepted 22 June 2015 Available online 27 June 2015 Keywords: Oxic-settling-anoxic process Iron content Feeding regime Sludge reduction Observed sludge yield
a b s t r a c t This study evaluates the biological solid reduction in a conventional activated sludge system with an anoxic/anaerobic side stream reactor receiving 1/10 of return sludge mass. Influent iron concentrations and feeding modes were changed to explore the consistency between the influent iron concentration and yield values and to assess the impact of feeding pattern. The results indicated that sludge reduction occurs during alternately exposure of sludge to aerobic and anoxic/anaerobic conditions in a range of 38–87%. The sludge reduction values reached a maximum level with the higher iron concentrations. Thus, it is concluded that this configuration is more applicable for plants receiving high iron concentrations in the wastewaters. Ó 2015 Elsevier Ltd. All rights reserved.
1. Introduction Traditionally, excess sludge production has been regarded as one of the most vital issues of activated sludge systems. Recently promulgated stringent regulations significantly increased the cost of sludge handling and disposal as they practically prohibited conventional and economically feasible approaches such as disposal to landfill (Wei et al., 2003); in fact, recent studies provided conclusive evidence that, technically, the organic carbon content of sludge could not be reduced to meet the new regulatory requirements (Orhon, 2014). Consequently, research is now focused on exploring different strategies that would potentially reduce the magnitude of sludge generation in the face of more expensive technologies such as drying, incineration, etc.
⇑ Corresponding author at: Istanbul Technical University, Civil Engineering Faculty, Department of Environmental Engineering, 34469 Maslak, Istanbul, Turkey. Tel.: +90 (212) 2857302; fax: +90 (212) 2856545. E-mail address: [email protected] (N. Yagci). http://dx.doi.org/10.1016/j.biortech.2015.06.109 0960-8524/Ó 2015 Elsevier Ltd. All rights reserved.
Separate handling of activated sludge has been historically tested and implemented for more effective stabilization; it was initially promoted as different re-aeration systems that basically involved aerating settled activated sludge in side-stream stabilization reactors prior to mixing with wastewater (Orhon, 2014). Later, more effective sludge stabilization was also reported to occur under anoxic/anaerobic conditions (Wanner, 1994; Eckenfelder, 1998). The oxic-settling-anoxic (OSA) process – also commercialized as the Cannibal process – is one of the promising process configurations recommended for sludge minimization (Saby et al., 2003; Chon et al., 2011). Basically, it involves a side-stream anoxic/anaerobic reactor designed to stabilize and reduce the excess sludge generated in the main activated sludge unit. The principles and performance of this system are well covered in different review papers on sludge minimization (Guo et al., 2013; Wei et al., 2003; Liu and Tay, 2001; Semblante et al., 2014). Related research generally utilized the observed yield values for the assessment of sludge reduction achieved. Novak et al. (2006) found that the observed yield values were about three times lower
214
N. Yagci et al. / Bioresource Technology 193 (2015) 213–218
and argued that the Cannibal system was likely to generate up to 60% less solids as compared to the conventional activated sludge process. This result was in agreement with the finding of other studies reporting similar sludge reductions in the range of 37–58% (Saby et al., 2003; Wang et al., 2007). Low et al. (2000) further postulated that decreased biomass yield per unit mass of substrate removed indicates uncoupled metabolism in biological wastewater treatment processes. The function and the positive impact on the biochemical mechanisms associated with the Cannibal process were also explored: Iron is reduced from ferric to ferrous ions under anaerobic conditions and this reduction promotes the release of protein into solution (Novak et al., 2003, 2007); this mechanism was also supported by the floc model suggested by Park and Novak (2007), where the released iron-associated protein is recycled to the aerobic activated sludge reactor and becomes available for microbial metabolism (Park et al., 2006). Studies also suggested a similar effect of the substrate feeding regime, which was empirically called substrate pressure, on the rate of solids generation by the Cannibal system (Khanthongthip et al., 2015). Khanthongthip (2010) reported that a 64% reduction could be achieved with the fast feeding regime compared with the slow-feed system, presumably due to higher production of easily biodegradable floc structure with high substrate pressure. Different mechanisms such as lysis–cryptic growth, uncoupling metabolism, maintenance metabolism, and predation on bacteria were proposed based on experimental results obtained (Low and Chase, 1999; Liu and Tay, 2001; Chen et al., 2003; Guo et al., 2013). In this context, the objective of the study was to provide experimental support for the effect of iron and feeding regime on the sludge generation mechanism in the OSA process and, particularly, to assess the experimental correlation between the iron content and the observed sludge yield under fast and slow feeding regimes.
were initiated at the beginning of reaction phase of each sequence. Both systems were fed with synthetic wastewater with short and long feeding periods, to provide low and high substrate pressures. The feeding periods of each system were started simultaneously with the aeration phase at the beginning of each cycle and continued for 5 min and 120 min for short and long feeding, respectively. Once a day within the operation scheme of each SBR, a 5 min mixing period was devoted to provide complete mixing before feeding of settled sludge from the SBR to the anaerobic bioreactor following the withdrawal phase. Then, the same volume of sludge was interchanged between SBR and anaerobic bioreactor to provide 10 days of hydraulic retention time in the anaerobic bioreactor.
2.2. Synthetic wastewater Both systems were fed with the same synthetic wastewater containing a total COD of 400 mg/l. The synthetic wastewater was composed of peptone (300 mgCOD/L), acetate (65 mgCOD/L) and propionate (35 mgCOD/L) as organic carbon sources. The feed solution was prepared daily using tap water by completely mixing of carbon sources with macro- and micronutrients according to a previous study by Novak et al. (2007). The change in the influent characteristics was only limited to changes in the concentrations of ferric iron in the range of 2.7–16.0 mgFe/L by addition of FeCl3. The iron concentration range is justifiable based upon typical measured iron content of wastewaters of domestic and mixture of domestic and industrial sewage (higher for industrial contributions). Each phase of the experiments performed with a different concentration of ferric iron in the influent will be described herein as runs. Consequently, iron concentrations in the synthetic wastewater were adjusted to 10.7, 5.35, 2.68 and 16.05 mgFe/L in Run-1, Run-2, Run-3 and Run-4, respectively.
2. Methods 2.1. Experimental set-up
2.3. Batch anaerobic tests
The experimental study involved parallel operation of two OSA systems with identical design except for their feeding patterns. The survey period was continued for 115 days (excluding the sludge acclimation phase at the beginning of the total operation period). The sludge from Blacksburg Wastewater Treatment Plant, VA USA, was used to start up the systems. Each system consisted of an aerobic sequencing batch reactor (SBR) coupled with side-stream anaerobic bioreactor as described by Chudoba et al. (1992); the reactor system was sustained at room temperature (22 ± 1 °C). The sludge was not wasted intentionally from the SBRs during the entire study. The overall sludge loss from the SBR was the daily sludge interchange between SBR and anaerobic bioreactor and a low amount of sludge escaping the system with the treated effluent. The interchange rate was selected as 10% based on the previous study by Easwaran (2006). It was maintained by recycling 1/10th of the settled sludge from the SBR through the anaerobic reactor. Thus, the hydraulic retention time (HRT) and sludge age were maintained at 10 days in the anaerobic bioreactor. The SBR reactor was designed with a working volume of 7.5 L and an initial volume of 3.75 L before feeding. The total cycle time was 8 h. Thus, the SBR operated with a hydraulic retention time (HRT) of 16 h and a ratio of the initial volume to the fill volume (V0/VF) of 1.0. The operation cycle of each SBR consisted of 6 h of reaction (only aerobic) phase and 45 min of settling phase, 15 min of decanting phase, 15 min of interchange phase and 45 min of idle phase. Simultaneous feeding, mixing and aeration
Anaerobic digestion tests were performed using sludge from the SBRs fed with high iron (16.05 mg/L) concentration. Two digesters were initially fed with approximately 750 mL of settled sludge produced in fast feed and slow feed SBRs and tightly sealed to prevent gas leakage and monitored daily for 10 days. The reactors were purged with nitrogen gas to establish and maintain anaerobic conditions in the batch test vessels initially and after every sampling.
2.4. Observed yield Observed yield is accepted as a suitable marker for the sludge reduction by researchers (Saby et al., 2003; Wang et al., 2007; Low et al., 2000; Troiani et al., 2011; Khanthongthip et al., 2015). In this study, the observed biomass yield values were determined for slow and fast feed SBRs receiving different influent iron concentrations by using the mixed liquor solids data. The observed yield values were calculated based upon the cumulative increase in solids (in term of VSS) divided by the corresponding removal of COD as given by Zhou et al. (2014). The amount of cumulative solids was estimated by the sum of the solids increased in the SBRs, the solids lost in the effluent and the solids removed from the reactors for sampling. A reduction in the observed yield was demonstrated sludge reduction in the OSA system compared to reference system. And, reduction of sludge production was calculated by using calculated observed sludge yield data from reference and OSA systems.
N. Yagci et al. / Bioresource Technology 193 (2015) 213–218
2.5. Analytical procedures The total solids (TS), total suspended solids (TSS), total volatile solids (VS), volatile suspended solids (VSS) contents and chemical oxygen demand (COD) were measured according to the Standard Methods for the Examination of Water and Wastewater (APHA, 2005). The soluble cations and anions were determined by using a Dionex ion chromatograph (Sunnyvale, California).
3. Results and discussion 3.1. The fate of solids in fast and slow feed systems for different iron concentrations The behavior of the two experimental units operated with fast and slow feeding regimes in terms of solids mass balance was primarily evaluated based on solids concentrations in the mixed liquor, MLSS and MLVSS. Both parameters exhibited a gradual increase during the start-up phase and reached a constant level within the first month of operation (data not shown). The system is reached steady state condition as evidenced by low variability in effluent quality and especially suspended solid concentration in the reactor. Then, the system is characterized via monitoring related parameters in the mixed-liquor and effluent. Fig. 1 shows MLSS and MLVSS profiles for the SBR systems and the corresponding anaerobic reactors for the entire operation period. The
215
observed profiles indicate no appreciable effect of changes in the iron concentration on MLSS and MLVSS levels at steady state. Consequently, the average MLSS and MLVSS concentrations characterizing the unit sustained with fast substrate feeding were calculated as 4050 ± 300 and 2800 ± 230 mg/L respectively. The corresponding MLSS and MLVSS levels were 2815 ± 260 and 1920 ± 180 mg/L, significantly lower as compared with fast feeding, confirming the effect of feeding regime on the magnitude of biomass maintained in the aerobic reactor. While this particular issue is beyond the scope of this study, preferential storage of the volatile fatty acid fractions of the substrate mixture during the fast feeding regime under aerobic conditions, and subsequent utilization of intracellular storage compounds for secondary growth coupled with a very low yield in the anaerobic reactor may be underlined for future studies (Ciggin et al., 2012, 2013). The second parameter used for the assessment of the two experimental units was the observed yield (g VSS/g COD) associated with different iron concentrations in the influent. As previously indicated, cumulative solids generation for each run was evaluated for a period where the MLVSS remained practically unchanged based on fractions lost in the effluent and removed for sampling. The observed yield values were calculated by using variation of VSS in the reactor (loss of VSS with sampling and from effluent) and consumed COD data. Fig. 2 shows the variation of cumulative solids (in terms of VSS) (x axis) over cumulative COD consumption (y axis) for the experimental runs, with the corresponding observed yield values. For computing observed yield,
Fig. 1. MLSS and MLVSS profiles in the systems during the period of study, (a) fast feed system, and (b) slow feed system.
216
N. Yagci et al. / Bioresource Technology 193 (2015) 213–218
Fig. 2. The cumulative solids for operation of system in d slow feed, and s fast feed mode.
the corresponding COD removals were calculated considering effluent COD concentrations. Soluble COD concentrations in the effluent always remained in the range of approximately 40 ± 20 mg/L for all sets. Fig. 3 displays the effect of iron concentration in the influent on the magnitude of resulting observed yield values for both fast and slow feeding regimes. It basically shows the significant decrease in observed yield, i.e. sludge generation, at higher iron concentrations. In fact, the observed yield value of 0.23 g VSS/g COD calculated for the iron concentration of 2.67 mg/L exhibited an almost 4-fold reduction, down to 0.05 g VSS/g COD in the experimental run where the iron concentration was increased to 16.05 mg/L. Moreover, a similar reduction trend in sludge generation was also observed for the unit sustained with a fast feeding regime, while the corresponding observed yield values remained always lower when compared with the slow feeding regime. These results confirm the mechanisms proposed by Park et al. (2006) that the degradation of iron associated organic matter is degraded easily under anaerobic conditions. The deflocculated activated sludge under anaerobic condition is coupled with iron reduction which associate volatile solid reduction as postulated by Novak et al. (2003) and subsequent sludge reduction. Furthermore, the observed yield values obtained in this study were compared with the value achieved in a conventional activated sludge which was operated as a control system by
Khanthongthip (2010). The yield value of 0.45 gTSS/gCOD and MLVSS/MLSS ratio of 0.83 for the conventional activated sludge system were used for the comparison. Thus, the solid accumulation with varying influent iron concentrations and different feed patterns were estimated as given in Table 1. The significance of the net/observed yield values obtained in this study may be better visualized by comparing them with a similar yield value of 0.45 g SS/g COD and a MLVSS/MLSS ratio of 0.83 reported for a conventional activated sludge system operated with fast feed but no sidestream reactor (Khanthongthip, 2010). Table 1 outlines the relative sludge reduction potentially achievable in the OSA system when compared to the conventional activated sludge configuration tested by Khanthongthip (2010) assuming it as a suitable control unit. In this context, the results displayed in Table 1 provide clear indication that (i) the OSA process provided a minimum sludge reduction of 38–44% for the lowest influent iron concentration of 2.67 mg/L. (ii) Sludge reduction was gradually increased to 87% when the iron level was also increased to the highest tested level of 16.05 mg/L From the table, it is apparent that the solid reduction is lower for slow feed conditions compared to fast feed conditions at all influent iron concentrations. It is also observed that high solid reduction values as high as 87% may be obtained in OSA systems receiving high iron concentrations regardless of the feeding pattern. 3.2. Variation in the settling properties of sludge in SBR units The sludge settling ability was also monitored during the study by measuring sludge volume index (SVI) throughout the experiments. Fig. 4 shows the relation between SVI values and the influent iron concentration in both systems. Table 1 Relative sludge percent reduction of the OSA system at different iron concentrations and feeding conditions.
Fig. 3. Effect of influent iron concentration on observed yield values.
Influent iron conc.
Sludge reduction (%)
mgFe/L
Slow feed
Fast feed
2.675 5.35 10.7 16.05
38 46 65 87
44 65 79 87
N. Yagci et al. / Bioresource Technology 193 (2015) 213–218
217
The figure essentially reflects the impact of feeding regime and iron concentration on the settling properties of the biomass. While the SVI values were always higher than 200 ml/g for the SBR unit operated with slow feeding, the same parameter remained always lower than 200 ml/g and gradually decreased with increasing iron levels, reaching a minimum value of (SVI = 108 ± 12 mL/g) for the highest tested iron concentration of 16.05 mg/L under fast feeding conditions. It should be noted that the measured SVI values are only valid for the synthetic substrate utilized and the operating conditions of the SBR units. However the significant SVI difference between the two systems can better be interpreted as reflections of the variations in the composition of microbial communities sustained under different feeding regimes, as investigated in detail and reported by Ciggin et al. (2012). Fig. 4. The relation between iron concentration and feeding conditions on the SVI in d slow feed, and s fast feed mode.
3.3. Effect of anaerobic exposure on aerobic sludge In order to stimulate anaerobic digestion, 750 mL of sludge was withdrawn from each SBR and subsequently exposed to
Fig. 5. The variation of phosphate, ammonia-nitrogen, potassium, magnesium, sulfate and nitrate concentrations during the anaerobic batch test in d slow feed, and s fast feed mode.
218
N. Yagci et al. / Bioresource Technology 193 (2015) 213–218
completely anaerobic conditions for 10 days in batch reactors. As illustrated in Fig. 5, a drastic increase in orthophosphate, ammonia and magnesium were monitored in each vessel and a similar potassium decrease for both fast and slow feed sludge during the monitoring period. After 10 days, release of P, N and Mg, and decrease in K, was systematically higher for fast feeding sludge. In fact, uptake or dilution of K was 27.7 and 31.3 mg/L, release of Mg was 20.0 and 58.7 mg/L, release of NH3-N was 35.1 and 42.3 mg-N/L, and release of PO4-P was 10.0 and 16.5 mg-P/L for slow feeding and fast feeding sludge, respectively. The release of principal ions that are present within the cell structure, and especially that of ammonia nitrogen and phosphorus may be evaluated for supporting and quantifying the accelerated sludge decay presumably taking place in the anaerobic reactors, due to destruction and lysis of living cellular material. Historically, these ions were successfully traced and utilized for the understanding of microbial decay and substrate storage mechanisms (Gaudy and Engelbrecht, 1963; Orhon, 2014). Since the main reactor was operated only under aerobic conditions, the release of phosphate and potassium is not attributed to polyphosphate cleavage from polyphosphate accumulating organisms. Moreover, the batch tests showed significant sulfate reduction after depletion of oxidized nitrogen forms, as shown in Fig. 5, which therefore could indicate coexistence of sulfate reducing microorganisms as slow-growing bacteria in the sludge where sulfate is utilized by the metabolism of these bacteria after complete anaerobic conditions are sustained; this way, the dominant mechanism could diverge from decay to metabolism of sulfate-reducers. Thus, the presence of these slow-growing organisms might partially explain the reduction in the overall sludge yield in this system configuration. On the other hand, iron could be reduced and combined with sulfide as demonstrated by Novak et al. (2007). Therefore, the sulfate reduction rates obtained from batch tests could also be explained by chemical precipitation of iron sulfides. However, since the iron content of the sludge and bulk liquid is theoretically not enough for sufficient precipitation of iron sulfides to cause a decrease in sulfate concentration from more than 100 mg/L to zero in a few days, it is assumed that sulfate reduction could not solely be attributed to chemical precipitation. Conversely, partial biological sulfate reduction could be associated with predominance of sulfate reducers in the anaerobic batch reactor. This mechanism could also be supported by a higher sulfate reduction rate obtained for sludge yielding higher sludge reduction.
4. Conclusions Experimental results suggested that the OSA system would be applicable for plants treating sewage with a high-iron content. Iron dosing was always more effective under fast feeding conditions, both in terms of sludge reductions achieved and improved sludge-settleability. Experimental results also revealed dissolution and release of essential elements for growth, when the sludge was exposed to anaerobic conditions. The data obtained in this study verified and modified the proposed outcome of models previously suggested. This aspect particularly recommended for further support and verification by respirometric and molecular analyses. Acknowledgements This research is Granted by The Scientific and Technological _ Department of Science Research Council of Turkey (TÜBITAK), Fellowships and Grant Programs (BIDEB) through supporting the
first author’s visit to University post-doctoral research fellow.
of
Virginia
Tech
as
a
References APHA, 2005. Standard Methods for the Examination of Water and Wastewater, 21st ed. American Public Health Association, American Water Works Association, Water Environment Federation, Washington DC, USA. Chen, G., An, K., Saby, S., Brois, E., Djafer, M., 2003. Possible cause of excess sludge reduction in an oxic-settling-anaerobic activated sludge process (OSA process). Water Res. 37, 3855–3866. Chon, D., Rome, Kim, Y.M., Park, K.Y., Park, C., 2011. Investigation of the sludge reduction mechanism in the anaerobic side-stream reactor process using several control biological wastewater treatment processes. Water Res. 45, 6021–6029. Chudoba, P., Morel, A., Capedeville, B., 1992. The case of both energetic uncoupling and metabolic selection of microorganisms in the OSA activated sludge system. Environ. Technol. 13, 761–770. Ciggin, A.S., Rosetti, S., Majone, M., Orhon, D., 2012. Effect of feeding regime and the sludge age on the fate of acetate and the microbial composition in sequencing batch reactor. J. Environ. Sci. Heal. A 47, 192–203. Ciggin, A.S., Orhon, D., Capitani, D., Miccheli, A., Puccetti, C., Majone, M., 2013. Aerobic metabolism of mixed carbon sources in sequencing batch reactor under pulse and continuous feeding. Bioresour. Technol. 129, 118–126. Easwaran, S.P., 2006. Developing a mechanistic understanding and optimization of the cannibal process phase II (M.Sc. thesis). Virginia Polytechnic Institute and State University, Blacksburg, VA. Eckenfelder, W., 1998. Activated Sludge: Process Design and Control. CRC Press. Gaudy, A.F., Engelbrecht, R.S., 1963. Basic biochemical consideration during metabolism in growing vs respiring systems. In: Eckenfelder, W.W., Jr., McCabe, J. (Eds.), Advances in Biological Waste Treatment. Pergamon Press, New York, pp. 1–10. Guo, W.Q., Yang, S.S., Xiang, W.S., Wang, X.J., Ren, N.Q., 2013. Minimization of excess, sludge production by in-situ activated sludge treatment processes — a comprehensive review. Biotechnol. Adv. 31, 1386–1396. Khanthongthip, P., 2010. The biological sludge reduction by anaerobic/aerobic cycling (Ph.D. thesis). Virginia Polytechnic Institute and State University, Blacksburg, VA. Khanthongthip, P., Novak, J.T., Doyle, M.L., 2015. Impact of substrate feed patterns on solids reduction by the cannibal process. Water Environ. Res. 87 (3), 274– 280. Liu, Y., Tay, J.H., 2001. Strategy for minimization of excess sludge production from the activated sludge process. Biotechnol. Adv. 19, 97–107. Low, E.W., Chase, H.A., 1999. Reducing production of excess biomass during wastewater treatment. Water Res. 33, 1119–1132. Low, E.W., Chase, H.A., Milner, M.G., Curtis, T.P., 2000. Uncoupling of metabolism to reduce biomass production in the activated sludge process. Water Res. 34, 3204–3212. Novak, J.T., Sadler, M.E., Murthy, S.N., 2003. Mechanisms of floc destruction during anaerobic and aerobic digestion and the effect on conditioning and dewatering of biosolids. Water Res. 37, 3136–3144. Novak, J.T., Chon, D.H., Curtis, B.A., Doyle, M., 2006. Reduction of sludge generation using the cannibal process: mechanisms and performance. In: Residuals and Biosolids Management Conference. Novak, J.T., Verma, N., Muller, C.D., 2007. The role of iron and aluminum in digestion and odor formation. Water Sci. Technol. 56, 59–65. Orhon, D., 2014. Evolution of the activated sludge process: the first 50 years. J. Chem. Technol. Biotechnol. 90, 608–640. Park, C., Novak, J.T., 2007. Characterization of activated sludge exocellular polymers using several cation-associated extraction methods. Water Res. 41 (8), 1679– 1688. Park, C., Muller, C.D., Abu-Orf, M.M., Novak, J.T., 2006. The effect of wastewater cations on activated sludge characteristics: effects of aluminum and iron in floc. Water Environ. Res. 78, 31–40. Saby, S., Djafer, M., Chen, G.H., 2003. Effect of low ORP in anoxic sludge zone on excess sludge production in oxic-settling-anoxic activated sludge process. Water Res. 37 (1), 11–20. Semblante, G.U., Hai, F.I., Ngo, H.H., Guo, W., You, S.J., Price, W.E., Nghiem, L.D., 2014. Sludge cycling between aerobic, anoxic and anaerobic regimes to reduce sludge production during wastewater treatment: Performance, mechanisms, and implications. Bioresour. Technol. 155, 395–409. Troiani, C., Eusebi, A.L., Battistoni, P., 2011. Excess sludge reduction by biological way: from experimental experience to a real full scale application. Bioresour. Technol. 102 (22), 10352–10358. Wang, J.F., Jin, W.B., Zhao, Q.L., Liu, Z.G., Lin, J.K., 2007. Performance of treating wastewater and anti-impact in oxic-settling-anaerobic (OSA) process for minimization of excess sludge. Chin. J. Environ. Sci. 28, 2488–2493. Wanner, J., 1994. Activated Sludge: Bulking and Foaming Control. CRC Press. Wei, Y., Van Houten, R.T., Borger, A.R., Eikelboom, D.H., Fan, Y., 2003. Minimization of excess sludge production for biological wastewater treatment. Water Res. 37 (18), 4453–4467. Zhou, Z., Qiao, W., Xing, C., Wang, Y., Wang, C., Wang, Y., Wang, Y., Wang, L., 2014. Sludge reduction and performance analysis of a modified sludge reduction process. Water Sci. Technol. 69, 934–940.
Contents lists available at ScienceDirect
Bioresource Technology journal homepage: www.elsevier.com/locate/biortech
The effect of iron dosing on reducing waste activated sludge in the oxic-settling-anoxic process Nevin Yagci a,⇑, John T. Novak b, Clifford W. Randall b, Derin Orhon a,c a
Faculty of Civil Engineering, Environmental Engineering Department, Istanbul Technical University, 34469 Istanbul, Turkey Virginia Polytechnic Institute and State University, Department of Civil Engineering, Blacksburg, VA 24061, USA c ENVIS Energy and Environmental Systems R&D Ltd, ITU Arı Teknokent, Arı-1 Building, 34469 Istanbul, Turkey b
h i g h l i g h t s High iron concentrations in the wastewater enhances the sludge reduction. Reduced observed yield and suitable sludge volume index with fast feeding regime. Simultaneously occurring sludge reduction mechanisms.
a r t i c l e
i n f o
Article history: Received 21 April 2015 Received in revised form 21 June 2015 Accepted 22 June 2015 Available online 27 June 2015 Keywords: Oxic-settling-anoxic process Iron content Feeding regime Sludge reduction Observed sludge yield
a b s t r a c t This study evaluates the biological solid reduction in a conventional activated sludge system with an anoxic/anaerobic side stream reactor receiving 1/10 of return sludge mass. Influent iron concentrations and feeding modes were changed to explore the consistency between the influent iron concentration and yield values and to assess the impact of feeding pattern. The results indicated that sludge reduction occurs during alternately exposure of sludge to aerobic and anoxic/anaerobic conditions in a range of 38–87%. The sludge reduction values reached a maximum level with the higher iron concentrations. Thus, it is concluded that this configuration is more applicable for plants receiving high iron concentrations in the wastewaters. Ó 2015 Elsevier Ltd. All rights reserved.
1. Introduction Traditionally, excess sludge production has been regarded as one of the most vital issues of activated sludge systems. Recently promulgated stringent regulations significantly increased the cost of sludge handling and disposal as they practically prohibited conventional and economically feasible approaches such as disposal to landfill (Wei et al., 2003); in fact, recent studies provided conclusive evidence that, technically, the organic carbon content of sludge could not be reduced to meet the new regulatory requirements (Orhon, 2014). Consequently, research is now focused on exploring different strategies that would potentially reduce the magnitude of sludge generation in the face of more expensive technologies such as drying, incineration, etc.
⇑ Corresponding author at: Istanbul Technical University, Civil Engineering Faculty, Department of Environmental Engineering, 34469 Maslak, Istanbul, Turkey. Tel.: +90 (212) 2857302; fax: +90 (212) 2856545. E-mail address: [email protected] (N. Yagci). http://dx.doi.org/10.1016/j.biortech.2015.06.109 0960-8524/Ó 2015 Elsevier Ltd. All rights reserved.
Separate handling of activated sludge has been historically tested and implemented for more effective stabilization; it was initially promoted as different re-aeration systems that basically involved aerating settled activated sludge in side-stream stabilization reactors prior to mixing with wastewater (Orhon, 2014). Later, more effective sludge stabilization was also reported to occur under anoxic/anaerobic conditions (Wanner, 1994; Eckenfelder, 1998). The oxic-settling-anoxic (OSA) process – also commercialized as the Cannibal process – is one of the promising process configurations recommended for sludge minimization (Saby et al., 2003; Chon et al., 2011). Basically, it involves a side-stream anoxic/anaerobic reactor designed to stabilize and reduce the excess sludge generated in the main activated sludge unit. The principles and performance of this system are well covered in different review papers on sludge minimization (Guo et al., 2013; Wei et al., 2003; Liu and Tay, 2001; Semblante et al., 2014). Related research generally utilized the observed yield values for the assessment of sludge reduction achieved. Novak et al. (2006) found that the observed yield values were about three times lower
214
N. Yagci et al. / Bioresource Technology 193 (2015) 213–218
and argued that the Cannibal system was likely to generate up to 60% less solids as compared to the conventional activated sludge process. This result was in agreement with the finding of other studies reporting similar sludge reductions in the range of 37–58% (Saby et al., 2003; Wang et al., 2007). Low et al. (2000) further postulated that decreased biomass yield per unit mass of substrate removed indicates uncoupled metabolism in biological wastewater treatment processes. The function and the positive impact on the biochemical mechanisms associated with the Cannibal process were also explored: Iron is reduced from ferric to ferrous ions under anaerobic conditions and this reduction promotes the release of protein into solution (Novak et al., 2003, 2007); this mechanism was also supported by the floc model suggested by Park and Novak (2007), where the released iron-associated protein is recycled to the aerobic activated sludge reactor and becomes available for microbial metabolism (Park et al., 2006). Studies also suggested a similar effect of the substrate feeding regime, which was empirically called substrate pressure, on the rate of solids generation by the Cannibal system (Khanthongthip et al., 2015). Khanthongthip (2010) reported that a 64% reduction could be achieved with the fast feeding regime compared with the slow-feed system, presumably due to higher production of easily biodegradable floc structure with high substrate pressure. Different mechanisms such as lysis–cryptic growth, uncoupling metabolism, maintenance metabolism, and predation on bacteria were proposed based on experimental results obtained (Low and Chase, 1999; Liu and Tay, 2001; Chen et al., 2003; Guo et al., 2013). In this context, the objective of the study was to provide experimental support for the effect of iron and feeding regime on the sludge generation mechanism in the OSA process and, particularly, to assess the experimental correlation between the iron content and the observed sludge yield under fast and slow feeding regimes.
were initiated at the beginning of reaction phase of each sequence. Both systems were fed with synthetic wastewater with short and long feeding periods, to provide low and high substrate pressures. The feeding periods of each system were started simultaneously with the aeration phase at the beginning of each cycle and continued for 5 min and 120 min for short and long feeding, respectively. Once a day within the operation scheme of each SBR, a 5 min mixing period was devoted to provide complete mixing before feeding of settled sludge from the SBR to the anaerobic bioreactor following the withdrawal phase. Then, the same volume of sludge was interchanged between SBR and anaerobic bioreactor to provide 10 days of hydraulic retention time in the anaerobic bioreactor.
2.2. Synthetic wastewater Both systems were fed with the same synthetic wastewater containing a total COD of 400 mg/l. The synthetic wastewater was composed of peptone (300 mgCOD/L), acetate (65 mgCOD/L) and propionate (35 mgCOD/L) as organic carbon sources. The feed solution was prepared daily using tap water by completely mixing of carbon sources with macro- and micronutrients according to a previous study by Novak et al. (2007). The change in the influent characteristics was only limited to changes in the concentrations of ferric iron in the range of 2.7–16.0 mgFe/L by addition of FeCl3. The iron concentration range is justifiable based upon typical measured iron content of wastewaters of domestic and mixture of domestic and industrial sewage (higher for industrial contributions). Each phase of the experiments performed with a different concentration of ferric iron in the influent will be described herein as runs. Consequently, iron concentrations in the synthetic wastewater were adjusted to 10.7, 5.35, 2.68 and 16.05 mgFe/L in Run-1, Run-2, Run-3 and Run-4, respectively.
2. Methods 2.1. Experimental set-up
2.3. Batch anaerobic tests
The experimental study involved parallel operation of two OSA systems with identical design except for their feeding patterns. The survey period was continued for 115 days (excluding the sludge acclimation phase at the beginning of the total operation period). The sludge from Blacksburg Wastewater Treatment Plant, VA USA, was used to start up the systems. Each system consisted of an aerobic sequencing batch reactor (SBR) coupled with side-stream anaerobic bioreactor as described by Chudoba et al. (1992); the reactor system was sustained at room temperature (22 ± 1 °C). The sludge was not wasted intentionally from the SBRs during the entire study. The overall sludge loss from the SBR was the daily sludge interchange between SBR and anaerobic bioreactor and a low amount of sludge escaping the system with the treated effluent. The interchange rate was selected as 10% based on the previous study by Easwaran (2006). It was maintained by recycling 1/10th of the settled sludge from the SBR through the anaerobic reactor. Thus, the hydraulic retention time (HRT) and sludge age were maintained at 10 days in the anaerobic bioreactor. The SBR reactor was designed with a working volume of 7.5 L and an initial volume of 3.75 L before feeding. The total cycle time was 8 h. Thus, the SBR operated with a hydraulic retention time (HRT) of 16 h and a ratio of the initial volume to the fill volume (V0/VF) of 1.0. The operation cycle of each SBR consisted of 6 h of reaction (only aerobic) phase and 45 min of settling phase, 15 min of decanting phase, 15 min of interchange phase and 45 min of idle phase. Simultaneous feeding, mixing and aeration
Anaerobic digestion tests were performed using sludge from the SBRs fed with high iron (16.05 mg/L) concentration. Two digesters were initially fed with approximately 750 mL of settled sludge produced in fast feed and slow feed SBRs and tightly sealed to prevent gas leakage and monitored daily for 10 days. The reactors were purged with nitrogen gas to establish and maintain anaerobic conditions in the batch test vessels initially and after every sampling.
2.4. Observed yield Observed yield is accepted as a suitable marker for the sludge reduction by researchers (Saby et al., 2003; Wang et al., 2007; Low et al., 2000; Troiani et al., 2011; Khanthongthip et al., 2015). In this study, the observed biomass yield values were determined for slow and fast feed SBRs receiving different influent iron concentrations by using the mixed liquor solids data. The observed yield values were calculated based upon the cumulative increase in solids (in term of VSS) divided by the corresponding removal of COD as given by Zhou et al. (2014). The amount of cumulative solids was estimated by the sum of the solids increased in the SBRs, the solids lost in the effluent and the solids removed from the reactors for sampling. A reduction in the observed yield was demonstrated sludge reduction in the OSA system compared to reference system. And, reduction of sludge production was calculated by using calculated observed sludge yield data from reference and OSA systems.
N. Yagci et al. / Bioresource Technology 193 (2015) 213–218
2.5. Analytical procedures The total solids (TS), total suspended solids (TSS), total volatile solids (VS), volatile suspended solids (VSS) contents and chemical oxygen demand (COD) were measured according to the Standard Methods for the Examination of Water and Wastewater (APHA, 2005). The soluble cations and anions were determined by using a Dionex ion chromatograph (Sunnyvale, California).
3. Results and discussion 3.1. The fate of solids in fast and slow feed systems for different iron concentrations The behavior of the two experimental units operated with fast and slow feeding regimes in terms of solids mass balance was primarily evaluated based on solids concentrations in the mixed liquor, MLSS and MLVSS. Both parameters exhibited a gradual increase during the start-up phase and reached a constant level within the first month of operation (data not shown). The system is reached steady state condition as evidenced by low variability in effluent quality and especially suspended solid concentration in the reactor. Then, the system is characterized via monitoring related parameters in the mixed-liquor and effluent. Fig. 1 shows MLSS and MLVSS profiles for the SBR systems and the corresponding anaerobic reactors for the entire operation period. The
215
observed profiles indicate no appreciable effect of changes in the iron concentration on MLSS and MLVSS levels at steady state. Consequently, the average MLSS and MLVSS concentrations characterizing the unit sustained with fast substrate feeding were calculated as 4050 ± 300 and 2800 ± 230 mg/L respectively. The corresponding MLSS and MLVSS levels were 2815 ± 260 and 1920 ± 180 mg/L, significantly lower as compared with fast feeding, confirming the effect of feeding regime on the magnitude of biomass maintained in the aerobic reactor. While this particular issue is beyond the scope of this study, preferential storage of the volatile fatty acid fractions of the substrate mixture during the fast feeding regime under aerobic conditions, and subsequent utilization of intracellular storage compounds for secondary growth coupled with a very low yield in the anaerobic reactor may be underlined for future studies (Ciggin et al., 2012, 2013). The second parameter used for the assessment of the two experimental units was the observed yield (g VSS/g COD) associated with different iron concentrations in the influent. As previously indicated, cumulative solids generation for each run was evaluated for a period where the MLVSS remained practically unchanged based on fractions lost in the effluent and removed for sampling. The observed yield values were calculated by using variation of VSS in the reactor (loss of VSS with sampling and from effluent) and consumed COD data. Fig. 2 shows the variation of cumulative solids (in terms of VSS) (x axis) over cumulative COD consumption (y axis) for the experimental runs, with the corresponding observed yield values. For computing observed yield,
Fig. 1. MLSS and MLVSS profiles in the systems during the period of study, (a) fast feed system, and (b) slow feed system.
216
N. Yagci et al. / Bioresource Technology 193 (2015) 213–218
Fig. 2. The cumulative solids for operation of system in d slow feed, and s fast feed mode.
the corresponding COD removals were calculated considering effluent COD concentrations. Soluble COD concentrations in the effluent always remained in the range of approximately 40 ± 20 mg/L for all sets. Fig. 3 displays the effect of iron concentration in the influent on the magnitude of resulting observed yield values for both fast and slow feeding regimes. It basically shows the significant decrease in observed yield, i.e. sludge generation, at higher iron concentrations. In fact, the observed yield value of 0.23 g VSS/g COD calculated for the iron concentration of 2.67 mg/L exhibited an almost 4-fold reduction, down to 0.05 g VSS/g COD in the experimental run where the iron concentration was increased to 16.05 mg/L. Moreover, a similar reduction trend in sludge generation was also observed for the unit sustained with a fast feeding regime, while the corresponding observed yield values remained always lower when compared with the slow feeding regime. These results confirm the mechanisms proposed by Park et al. (2006) that the degradation of iron associated organic matter is degraded easily under anaerobic conditions. The deflocculated activated sludge under anaerobic condition is coupled with iron reduction which associate volatile solid reduction as postulated by Novak et al. (2003) and subsequent sludge reduction. Furthermore, the observed yield values obtained in this study were compared with the value achieved in a conventional activated sludge which was operated as a control system by
Khanthongthip (2010). The yield value of 0.45 gTSS/gCOD and MLVSS/MLSS ratio of 0.83 for the conventional activated sludge system were used for the comparison. Thus, the solid accumulation with varying influent iron concentrations and different feed patterns were estimated as given in Table 1. The significance of the net/observed yield values obtained in this study may be better visualized by comparing them with a similar yield value of 0.45 g SS/g COD and a MLVSS/MLSS ratio of 0.83 reported for a conventional activated sludge system operated with fast feed but no sidestream reactor (Khanthongthip, 2010). Table 1 outlines the relative sludge reduction potentially achievable in the OSA system when compared to the conventional activated sludge configuration tested by Khanthongthip (2010) assuming it as a suitable control unit. In this context, the results displayed in Table 1 provide clear indication that (i) the OSA process provided a minimum sludge reduction of 38–44% for the lowest influent iron concentration of 2.67 mg/L. (ii) Sludge reduction was gradually increased to 87% when the iron level was also increased to the highest tested level of 16.05 mg/L From the table, it is apparent that the solid reduction is lower for slow feed conditions compared to fast feed conditions at all influent iron concentrations. It is also observed that high solid reduction values as high as 87% may be obtained in OSA systems receiving high iron concentrations regardless of the feeding pattern. 3.2. Variation in the settling properties of sludge in SBR units The sludge settling ability was also monitored during the study by measuring sludge volume index (SVI) throughout the experiments. Fig. 4 shows the relation between SVI values and the influent iron concentration in both systems. Table 1 Relative sludge percent reduction of the OSA system at different iron concentrations and feeding conditions.
Fig. 3. Effect of influent iron concentration on observed yield values.
Influent iron conc.
Sludge reduction (%)
mgFe/L
Slow feed
Fast feed
2.675 5.35 10.7 16.05
38 46 65 87
44 65 79 87
N. Yagci et al. / Bioresource Technology 193 (2015) 213–218
217
The figure essentially reflects the impact of feeding regime and iron concentration on the settling properties of the biomass. While the SVI values were always higher than 200 ml/g for the SBR unit operated with slow feeding, the same parameter remained always lower than 200 ml/g and gradually decreased with increasing iron levels, reaching a minimum value of (SVI = 108 ± 12 mL/g) for the highest tested iron concentration of 16.05 mg/L under fast feeding conditions. It should be noted that the measured SVI values are only valid for the synthetic substrate utilized and the operating conditions of the SBR units. However the significant SVI difference between the two systems can better be interpreted as reflections of the variations in the composition of microbial communities sustained under different feeding regimes, as investigated in detail and reported by Ciggin et al. (2012). Fig. 4. The relation between iron concentration and feeding conditions on the SVI in d slow feed, and s fast feed mode.
3.3. Effect of anaerobic exposure on aerobic sludge In order to stimulate anaerobic digestion, 750 mL of sludge was withdrawn from each SBR and subsequently exposed to
Fig. 5. The variation of phosphate, ammonia-nitrogen, potassium, magnesium, sulfate and nitrate concentrations during the anaerobic batch test in d slow feed, and s fast feed mode.
218
N. Yagci et al. / Bioresource Technology 193 (2015) 213–218
completely anaerobic conditions for 10 days in batch reactors. As illustrated in Fig. 5, a drastic increase in orthophosphate, ammonia and magnesium were monitored in each vessel and a similar potassium decrease for both fast and slow feed sludge during the monitoring period. After 10 days, release of P, N and Mg, and decrease in K, was systematically higher for fast feeding sludge. In fact, uptake or dilution of K was 27.7 and 31.3 mg/L, release of Mg was 20.0 and 58.7 mg/L, release of NH3-N was 35.1 and 42.3 mg-N/L, and release of PO4-P was 10.0 and 16.5 mg-P/L for slow feeding and fast feeding sludge, respectively. The release of principal ions that are present within the cell structure, and especially that of ammonia nitrogen and phosphorus may be evaluated for supporting and quantifying the accelerated sludge decay presumably taking place in the anaerobic reactors, due to destruction and lysis of living cellular material. Historically, these ions were successfully traced and utilized for the understanding of microbial decay and substrate storage mechanisms (Gaudy and Engelbrecht, 1963; Orhon, 2014). Since the main reactor was operated only under aerobic conditions, the release of phosphate and potassium is not attributed to polyphosphate cleavage from polyphosphate accumulating organisms. Moreover, the batch tests showed significant sulfate reduction after depletion of oxidized nitrogen forms, as shown in Fig. 5, which therefore could indicate coexistence of sulfate reducing microorganisms as slow-growing bacteria in the sludge where sulfate is utilized by the metabolism of these bacteria after complete anaerobic conditions are sustained; this way, the dominant mechanism could diverge from decay to metabolism of sulfate-reducers. Thus, the presence of these slow-growing organisms might partially explain the reduction in the overall sludge yield in this system configuration. On the other hand, iron could be reduced and combined with sulfide as demonstrated by Novak et al. (2007). Therefore, the sulfate reduction rates obtained from batch tests could also be explained by chemical precipitation of iron sulfides. However, since the iron content of the sludge and bulk liquid is theoretically not enough for sufficient precipitation of iron sulfides to cause a decrease in sulfate concentration from more than 100 mg/L to zero in a few days, it is assumed that sulfate reduction could not solely be attributed to chemical precipitation. Conversely, partial biological sulfate reduction could be associated with predominance of sulfate reducers in the anaerobic batch reactor. This mechanism could also be supported by a higher sulfate reduction rate obtained for sludge yielding higher sludge reduction.
4. Conclusions Experimental results suggested that the OSA system would be applicable for plants treating sewage with a high-iron content. Iron dosing was always more effective under fast feeding conditions, both in terms of sludge reductions achieved and improved sludge-settleability. Experimental results also revealed dissolution and release of essential elements for growth, when the sludge was exposed to anaerobic conditions. The data obtained in this study verified and modified the proposed outcome of models previously suggested. This aspect particularly recommended for further support and verification by respirometric and molecular analyses. Acknowledgements This research is Granted by The Scientific and Technological _ Department of Science Research Council of Turkey (TÜBITAK), Fellowships and Grant Programs (BIDEB) through supporting the
first author’s visit to University post-doctoral research fellow.
of
Virginia
Tech
as
a
References APHA, 2005. Standard Methods for the Examination of Water and Wastewater, 21st ed. American Public Health Association, American Water Works Association, Water Environment Federation, Washington DC, USA. Chen, G., An, K., Saby, S., Brois, E., Djafer, M., 2003. Possible cause of excess sludge reduction in an oxic-settling-anaerobic activated sludge process (OSA process). Water Res. 37, 3855–3866. Chon, D., Rome, Kim, Y.M., Park, K.Y., Park, C., 2011. Investigation of the sludge reduction mechanism in the anaerobic side-stream reactor process using several control biological wastewater treatment processes. Water Res. 45, 6021–6029. Chudoba, P., Morel, A., Capedeville, B., 1992. The case of both energetic uncoupling and metabolic selection of microorganisms in the OSA activated sludge system. Environ. Technol. 13, 761–770. Ciggin, A.S., Rosetti, S., Majone, M., Orhon, D., 2012. Effect of feeding regime and the sludge age on the fate of acetate and the microbial composition in sequencing batch reactor. J. Environ. Sci. Heal. A 47, 192–203. Ciggin, A.S., Orhon, D., Capitani, D., Miccheli, A., Puccetti, C., Majone, M., 2013. Aerobic metabolism of mixed carbon sources in sequencing batch reactor under pulse and continuous feeding. Bioresour. Technol. 129, 118–126. Easwaran, S.P., 2006. Developing a mechanistic understanding and optimization of the cannibal process phase II (M.Sc. thesis). Virginia Polytechnic Institute and State University, Blacksburg, VA. Eckenfelder, W., 1998. Activated Sludge: Process Design and Control. CRC Press. Gaudy, A.F., Engelbrecht, R.S., 1963. Basic biochemical consideration during metabolism in growing vs respiring systems. In: Eckenfelder, W.W., Jr., McCabe, J. (Eds.), Advances in Biological Waste Treatment. Pergamon Press, New York, pp. 1–10. Guo, W.Q., Yang, S.S., Xiang, W.S., Wang, X.J., Ren, N.Q., 2013. Minimization of excess, sludge production by in-situ activated sludge treatment processes — a comprehensive review. Biotechnol. Adv. 31, 1386–1396. Khanthongthip, P., 2010. The biological sludge reduction by anaerobic/aerobic cycling (Ph.D. thesis). Virginia Polytechnic Institute and State University, Blacksburg, VA. Khanthongthip, P., Novak, J.T., Doyle, M.L., 2015. Impact of substrate feed patterns on solids reduction by the cannibal process. Water Environ. Res. 87 (3), 274– 280. Liu, Y., Tay, J.H., 2001. Strategy for minimization of excess sludge production from the activated sludge process. Biotechnol. Adv. 19, 97–107. Low, E.W., Chase, H.A., 1999. Reducing production of excess biomass during wastewater treatment. Water Res. 33, 1119–1132. Low, E.W., Chase, H.A., Milner, M.G., Curtis, T.P., 2000. Uncoupling of metabolism to reduce biomass production in the activated sludge process. Water Res. 34, 3204–3212. Novak, J.T., Sadler, M.E., Murthy, S.N., 2003. Mechanisms of floc destruction during anaerobic and aerobic digestion and the effect on conditioning and dewatering of biosolids. Water Res. 37, 3136–3144. Novak, J.T., Chon, D.H., Curtis, B.A., Doyle, M., 2006. Reduction of sludge generation using the cannibal process: mechanisms and performance. In: Residuals and Biosolids Management Conference. Novak, J.T., Verma, N., Muller, C.D., 2007. The role of iron and aluminum in digestion and odor formation. Water Sci. Technol. 56, 59–65. Orhon, D., 2014. Evolution of the activated sludge process: the first 50 years. J. Chem. Technol. Biotechnol. 90, 608–640. Park, C., Novak, J.T., 2007. Characterization of activated sludge exocellular polymers using several cation-associated extraction methods. Water Res. 41 (8), 1679– 1688. Park, C., Muller, C.D., Abu-Orf, M.M., Novak, J.T., 2006. The effect of wastewater cations on activated sludge characteristics: effects of aluminum and iron in floc. Water Environ. Res. 78, 31–40. Saby, S., Djafer, M., Chen, G.H., 2003. Effect of low ORP in anoxic sludge zone on excess sludge production in oxic-settling-anoxic activated sludge process. Water Res. 37 (1), 11–20. Semblante, G.U., Hai, F.I., Ngo, H.H., Guo, W., You, S.J., Price, W.E., Nghiem, L.D., 2014. Sludge cycling between aerobic, anoxic and anaerobic regimes to reduce sludge production during wastewater treatment: Performance, mechanisms, and implications. Bioresour. Technol. 155, 395–409. Troiani, C., Eusebi, A.L., Battistoni, P., 2011. Excess sludge reduction by biological way: from experimental experience to a real full scale application. Bioresour. Technol. 102 (22), 10352–10358. Wang, J.F., Jin, W.B., Zhao, Q.L., Liu, Z.G., Lin, J.K., 2007. Performance of treating wastewater and anti-impact in oxic-settling-anaerobic (OSA) process for minimization of excess sludge. Chin. J. Environ. Sci. 28, 2488–2493. Wanner, J., 1994. Activated Sludge: Bulking and Foaming Control. CRC Press. Wei, Y., Van Houten, R.T., Borger, A.R., Eikelboom, D.H., Fan, Y., 2003. Minimization of excess sludge production for biological wastewater treatment. Water Res. 37 (18), 4453–4467. Zhou, Z., Qiao, W., Xing, C., Wang, Y., Wang, C., Wang, Y., Wang, Y., Wang, L., 2014. Sludge reduction and performance analysis of a modified sludge reduction process. Water Sci. Technol. 69, 934–940.
