Bioresource Technology 171 (2014) 491–494
Contents lists available at ScienceDirect
Bioresource Technology journal homepage: www.elsevier.com/locate/biortech
Short Communication
Early warning indicators for monitoring the process failure of anaerobic digestion system of food waste Lei Li, Qingming He, Yunmei Wei, Qin He, Xuya Peng ⇑ Key Laboratory of Three Gorges Reservoir Region’s Eco-Environment, Ministry of Education, Chongqing University, Chongqing 400045, China
h i g h l i g h t s Operation conditions significantly affect the responses of state parameters. The recommended early warning indicators are different among digesters. None of the single indicators was universally valid for all the systems. A combination of total VFA, VFA/TA and BA/TA has a general adaptability.
a r t i c l e
i n f o
Article history: Received 16 June 2014 Received in revised form 18 August 2014 Accepted 21 August 2014 Available online 28 August 2014 Keywords: Anaerobic digestion Food waste Early warning indicators Organic overloading Process stability
a b s t r a c t To determine reliable state parameters which could be used as early warning indicators of process failure due to the acidification of anaerobic digestion of food waste, three mesophilic anaerobic digesters of food waste with different operation conditions were investigated. Such parameters as gas production, methane content, pH, concentrations of volatile fatty acid (VFA), alkalinity and their combined indicators were evaluated. Results revealed that operation conditions significantly affect the responses of parameters and thus the optimal early warning indicators of each reactor differ from each other. None of the single indicators was universally valid for all the systems. The universally valid indicators should combine several parameters to supply complementary information. A combination of total VFA, the ratio of VFA to total alkalinity (VFA/TA) and the ratio of bicarbonate alkalinity to total alkalinity (BA/TA) can reflect the metabolism of the digesting system and realize rapid and effective early warning. Ó 2014 Elsevier Ltd. All rights reserved.
1. Introduction Anaerobic digestion (AD) usually features process imbalance under high organic loading rate (OLR) (Méndez-Acosta et al., 2010), so the process monitoring and control are indispensable for achieving a high-efficiency and stabilized performance of the AD system (Bjornsson et al., 2001), while effective process parameters are the basis for ensuring the process monitoring and control. A multitude of studies have explored the process parameters of the AD system. To sum up, the most commonly used parameters include pH, volatile fatty acid (VFA), alkalinity, biogas production, methane content and their combination factors (Bjornsson et al., 2001). However, previous researches on early warning parameters always focused on a specific parameter, verified its effectiveness
⇑ Corresponding author. Address: No. 174, Shapingba Zhengjie Street, Chongqing 400045, China. Tel.: +86 13594614338; fax: +86 023 65121734. E-mail address: [email protected] (X. Peng). http://dx.doi.org/10.1016/j.biortech.2014.08.089 0960-8524/Ó 2014 Elsevier Ltd. All rights reserved.
and defined its threshold value (Martín-González et al., 2013; Wang et al., 2009); or studied the responses of multiple parameters under a specific condition, and suggested the optimal parameters for this operating conditions (Nielsen et al., 2007; Kleyböcker et al., 2012; Ahring et al., 1995). Thus, though lots of researches on early warning parameters have been conducted, the proposed early warning indicators are only effective when applied to these specific operating conditions, while less effective in systems with different operation conditions. For example, the ratio of intermediate alkalinity to partial alkalinity (IA/PA) of 0.9 was suggested in order to maintain stable operation in thermophilic reactors treating sewage sludge; however IA/PA of 0.4 was proposed to assure a stable reactor performance when a potato-starch wastewater was treated; and IA/PA below 0.3 was recommended to maintain a stable operation when municipal solid waste was treated (Martín-González et al., 2013). In contrast, Hill et al. (1987) suggested that the ratio of propionate to acetate (Pr /Ac ) higher than 1.4 indicated impending digester failure, but Weiland (2008) held that the Pr /Ac shall be lower than 1.0 in a stable system; and
492
L. Li et al. / Bioresource Technology 171 (2014) 491–494
Ahring et al. (1995) suggested that Pr /Ac was useless in determining process imbalance. Other parameters are also suffering the similar phenomenon. Therefore, the indicators suggested by each study are different, as for the same warning parameter there even existed multi-thresholds. So far there still lacks of an early warning index that is universally applicable. This study is aimed at comprehensive evaluation of the validity of parameters under different operation conditions, and find out the early warning index which is universal applicable for indicating process failure of food waste (FW) digesters. Three anaerobic digesters with different operation conditions were investigated; and such process parameters including gas production and composition, pH, VFA, alkalinity and their combined indicators were all evaluated.
parameters measured daily. VS removal rate (VSr) was calculated by the same equation as reported by Koch et al. (2009). 2.3. Analytical methods
2. Methods
TS and VS were measured according to standard methods (APHA et al., 1998). pH was measured using a pH meter (Horiba, B-212). Total VFA, total alkalinity (TA) bicarbonate alkalinity (BA), intermediate alkalinity (IA) and partial alkalinity (PA) were analyzed according to Anderson and Yang (1992). Biogas production was measured using wet gas meters. Biogas composition was determined by a Biogas 5000 portable instrument (Geotech, UK). Ammonia-nitrogen was analyzed using a DR-2800 spectrophotometer (HACH, USA). Individual VFA were analyzed on a gas chromatograph equipped with flame ionization detector.
2.1. Feedstock and seeding sludge
3. Results and discussion
FW was collected from a school canteen, and with the removals of such coarse impurities as bones and plastics, the collected FW was divided into to 2 portions, and 1/3 of the FW was placed in boiling water for the removal of the surface floating oil, then it was subject to solid–liquid separation through a sieve with 2 2 mm lattice, and the obtained solid part was shredded into particles with an average size of 5.0 mm by a Robot-Coupe Shredder. The remaining 2/3 were milled directly without degreasing. Then, the two kinds of materials were all packed into 4-L plastic storage bags, and cryopreserved at 18 °C. One week prior to use, the frozen feedstock was thawed, and stored at 4 °C. The seed sludge used in this study was respectively the digested sludge taken from Dadukou Municipal Wastewater Treatment Plant, Chongqing, China (MWS); and digested sludge from an anaerobic digester treating piggery waste (PWS) and digested sludge from a 50 L anaerobic digester of FW (FWS). The characteristics of the substrate and seed sludge were listed in the Supplemental materials (SM) Table S1.
3.1. Digester performance
2.2. Digesters operation Three complete-mix anaerobic digesters each with a working volume of 20 L were operated at 36 ± 1 °C. Motorized automatic stirring was provided at the top, with an agitating frequency of 20 min h 1 and a rotary speed of 40 rpm. Different substrate pretreatment modes, seeding sludge sources and retention time were respectively adopted for the three digesters, and the specific operating conditions of the various digesters are shown in Table 1. The digesters were operated at a semi-continuous mode; substrate was added once daily and before which manual sampling was conducted at the sampling opening. For an efficient injection, one token of the substrate was mixed with twice volume of retrieved sludge from the operating reactor. During the experiment, the total solids (TS) and volatile solids (VS) of digestate were measured every three days, and the other
OLR, methane yield and VSr are usually used for the evaluation of the anaerobic digester performance (Nagao et al., 2012). Perturbations of OLR were introduced, and steady state was determined by constant methane yield and VSr in this paper. The date when sharp declines appeared in methane yield or VSr was defined as the date of the process failure. Fig. 1 manifests the responses of methane yield and VSr along with the OLR during the experiment. It can be seen that the fluctuations of VSr were slight throughout the experiment, with the specific values being 87.5 ± 3.9%, 88.9 ± 2.8% and 86.6 ± 3.4% in Reactor A, B and C, respectively. Thus they are within the recommended range of 81–92% (Nagao et al., 2012) throughout the experiment, indicating that it is infeasible to use the VSr for performance description. However, the fluctuations in methane yield were much more obvious. Approximately 50% reductions appeared on day 66 and 72 in Reactor A and C, respectively, and the same reductions occurred on day 34 and 62 in Reactor B. Accordingly, the corresponding dates of process failure were defined, which was indicated by the vertical dotted line in Fig. 1 and SM Figs. S1–S3. 3.2. Response of process parameters As with methane yield, process parameters will respond to process perturbations, and once response from a process parameter is faster than that of methane yield, then the process parameter can be used as early warning indicator. The curve diagrams of the responses of various process parameters along with OLR in three reactors were shown in SM Figs. S1–S3; and Table 2 summarized the specific early warning information of various parameters. Among then the dates of the process imbalance for IA/PA, VFA/TA and BA/TA were, respectively defined as the date when the ratios exceeded their critical value of 0.3, 0.35 and 0.8 (Martín-González et al., 2013; Méndez-Acosta et al., 2010). The
Table 1 Specific operating conditions of three digesters.
a b
1
Inoculum
Substrate
Stage
OLR (g VS L
A B
MWS PWS
Pre-treated FW Raw FW
C
FWS
Raw FW
Gradual overload Gradual overload Recovery stage Sudden overload Recovery stage Gradual overload
4, 5, 6, 7, 7.5, 8, 8.5 4, 5, 6, 7, 7.5 1, 4 4–7.1b 1 2, 3, 4, 5, 6
Retention time of each OLR. Increased 10% daily.
d
1
Digester
)
Retention timea(d) 10 7 10 1 8 15
493
L. Li et al. / Bioresource Technology 171 (2014) 491–494 Table 2 Specific early warning information of three reactors. Reactor
Parameters
Date of process failure (d)
Date of process imbalance (d)
Warning time (d)
A
BA/TA Acetate VFA IA/PA VFA/TA Propionate CH4 Pr /Ac TA Gas production pH
66
53 54 54 54 56 56 62 63 64 66
13 12 12 12 10 10 4 3 2 0
B
VFA/TA IA/PA Acetate BA/TA pH Propionate VFA Pr /Ac Gas production TA CH4
34 (62)a
IA/PA Acetate VFA BA/TA VFA/TA Propionate Pr /Ac pH CH4 Gas production TA
72
C
Fig. 1. The evolution of OLR, methane yield and VSr in R(A), R(B) and R(C) during the experiment.
selected thresholds are valid in all three reactors, and seem more suitable than any other thresholds to indicate the process imbalance of FW digesters. While for other parameters no consistent thresholds can be used, so the dates of the process imbalance were those when sudden changes in a relatively constant value occur. The warning time is the difference between the date of process failure and imbalance. As shown in Table 2, though the evaluated parameters were consistent among digesters, the warning effects of parameters were different. Generally speaking, the most insensitive indicators were gasphase indexes, pH, TA and Pr /Ac , and their warning time were not more than 5 days in all digesters. The limitations in liquidto-gas mass transfer may be the main reason of the low sensitivity in gas-phase indexes (Bjornsson et al., 2001). The delay in pH may have been caused by a high buffering capacity in FW digesters. And TA showed almost no warning effect, since an increase in VFA will cause a decrease in BA, resulting in a basically constant TA-value (Bjornsson et al., 2001). While the delay of Pr /Ac was caused by the fact that acetate exhibited faster response to the disturbance than propionate. In contrast, total VFA, acetate and propionate had featured rapider responses, their warning times range from 2 to 12 days, and the warning orders were basically consistent. Total VFA and acetate responded simultaneously, followed by propionate response. Kleyböcker et al. (2012) also drew similar conclusions, and they found acetate accumulated first due to the inhibition of the acetoclastic methanogenesis, followed by the accumulation of propionate due to the inhibition of the acetogens. Combined indicators had similar or better early warning effect compared to VFA, while their warning effects were more unstable. For example, BA/TA was the best early warning indicator in Reactor A, but it lagged back in Reactor B; the VFA/TA recommended by Reactor B also was not so sensitive as the IA/PA in Reactor C. The differences in warning order of combined indicators may have
a
66 24 26 27 30 30 31 31 31 32
0 (56) (60) (58) (60) (61) (59) (60) (60) (62)
10 (6) 8 (2) 7 (4) 4 (2) 4 (1) 3 (3) 3 (2) 3 (2) 2 (0)
34 (62) 34 (62)
0 (0) 0 (0)
65 68 68 68 68 69 70 71 72 73
7 4 4 4 4 3 2 1 0 /
74
/
The values in brackets refer to the corresponding value of sudden overload.
been caused by the different contents of ammonia among digesters. As we know, the effective buffer capacity in AD systems includes the two parts of ammonia and BA. In high ammonia system, the ammonia would react with accumulated VFA sooner than BA due to the higher pKa (Boe, 2006), resulting in an decrease in TA, while BA had no change, thus the fluctuation orders of the three combined parameters would be VFA/TA earlier than BA/TA and VFA/BA. However, when the ammonia concentration was low, and even can be ignored, the BA would be consumed directly to counteract the accumulated VFA, thus causing the fluctuation order changes to VFA/BA, which was greater than the other two parameters. In our experiments, due to the effects of the substrate and inoculum, ammonia concentration in Reactor B was higher, so the warning effect of VFA/TA was the best; the ammonia concentration in Reactor A and C was lower, and the optimal indicator in Reactor C was IA/PA, while the earliest indicator in Reactor A was BA/TA. Background fluctuations may be one of the incentives. As can be seen from Fig. S1, the warning of BA/TA was not continuous, and the ratio is greater than its threshold at day 55, while IA/PA has begun to warn at day 54. Certainly there may be other unknown factors that improve the early warning performance of BA/TA in Reactor A. In addition, it is worth mentioning that IA/PA was used to replace VFA/BA in this paper, since these two were almost equivalent (Martín-González et al., 2013), and the threshold of IA/PA was lower than that of VFA/BA, so the warning effect of IA/PA is considered theoretically better than that of VFA/BA. In addition, as show in Table 2, the warning time of parameters were different among digesters, the optimal warning time were
494
L. Li et al. / Bioresource Technology 171 (2014) 491–494
13, 10 and 6, 7 days in Reactor A–C, respectively. The differences may mainly have been caused by the different retention time under each OLR. Ahring et al. (1995) reported that most of indicators were suitable for detecting gradual overloads, but were too slow for the response of sudden overloads. Accordingly, based on the different retention time of digesters, the warning time should be that Reactor C is longer than Reactor A and B. However, Reactor C actually held the shortest warning time in disturbance of gradual overloads. This abnormality may have been caused by seeding sludge. Elbeshbishy et al. (2012) found that the digesters’ performances were not as good as other kinds of seed sludge when FWS was used as the seed sludge. While in Reactor C what was used was not only digested sludge from FW digester, but the sludge was the restored sludge after acidification, thus its performance may have been subject to the dual impacts, which may have influenced the state parameters’ response to some extent. Considering the response of the state parameters was affected by many factors, it is difficult to obtain accurate information and explicit evidence about the inconsistencies of early warning indicators among digesters. However, the inference herein is still reasonable to some degree. 3.3. Suitability of process parameters Based on the above analysis, VFA and combined indicators are more suitable for applications as early warning indicators in AD systems of FW. However, each parameter showed different warning effects in different reactors, thus the optimal indicators were inconsistent among digesters, and none of these indicators was universally valid for all the systems. To describe the characteristics of the system effectively, it is necessary to combine several indicators to supply complementary information. The unbalance between acid production and consumption was the source of process failure, so it is necessary to choose an acid concentration as one of the indicators. Theoretically, the individual VFA are the best early warning indicators; however, the monitoring of individual VFA is difficult, while the total one, which possessed similar early warning performances, can be monitored simply, thus becoming an excellent substitute of the individual VFA. Among the three combined indicators, BA/TA reflects the buffer equilibrium of the system, while VFA/TA and IA/PA represent the acid–base balance of the system. To comprehensively reflect process statues, BA/TA should be chosen as one of the indicators and in view of the advantages of VFA/TA in the high ammonia nitrogen system, VFA/TA rather than IA/PA should be chosen to reflect acid–base balance of the system. The joint use of total VFA, VFA/TA and BA/TA can reflect the metabolism of the digesting system so as to realize rapid and effective early warning. 4. Conclusions VFA and combined indicators give fast and reliable information of process state compared to other parameters. While the operation conditions significantly affect the performance of digesters and
thus the most sensitive early warning indicators are still different among digesters. None of the single indicators was universally valid for all the systems. A combination of total VFA, VFA/TA and BA/TA to supply complementary information is suitable for monitoring this type of digesters effectively. Acknowledgements This work was granted by financial support of the Key Projects in the National Science & Technology Pillar Program during the Eleventh Five-year Plan Period (No. 2010BAC67B01). Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.biortech.2014.08. 089. References Ahring, B.K., Sandberg, M., Angelidaki, I., 1995. Volatile fatty acids as indicators of process imbalance in anaerobic digesters. Appl. Microbiol. Biotechnol. 43, 559– 565. Anderson, G.K., Yang, G., 1992. Determination of bicarbonate and total volatile acid concentration in anaerobic digesters using a simple titration. Water Environ. Res. 64, 53–59. APHA, AWWA, WEF, 1998. Standard Methods for the Examination of Water and Wastewater, Washington. Bjornsson, L., Murto, M., Jantsch, T.G., Mattiasson, B., 2001. Evaluation of new methods for the monitoring of alkalinity, dissolved hydrogen and the microbial community in anaerobic digestion. Water Res. 35, 2833–2840. Boe, K., 2006. Online Monitoring and Control of the Biogas Process (Ph.D. thesis). Institute of Environments & Resources, Technical University of Denmark. Elbeshbishy, E., Nakhla, G., Hafez, H., 2012. Biochemical methane potential (BMP) of food waste and primary sludge: influence of inoculum pre-incubation and inoculum source. Bioresour. Technol. 110, 18–25. Hill, D., Cobbs, S., Bolte, J., 1987. Using volatile fatty acid relationships to predict anaerobic digester failure. Trans. ASAE 30, 496–501. Kleyböcker, A., Liebrich, M., Verstraete, W., Kraume, M., Würdemann, H., 2012. Early warning indicators for process failure due to organic overloading by rapeseed oil in one-stage continuously stirred tank reactor, sewage sludge and waste digesters. Bioresour. Technol. 123, 534–541. Koch, K., Wichern, M., Lübken, M., Horn, H., 2009. Mono fermentation of grass silage by means of loop reactors. Bioresour. Technol. 100, 5934–5940. Martín-González, L., Font, X., Vicent, T., 2013. Alkalinity ratios to identify process imbalances in anaerobic digesters treating source-sorted organic fraction of municipal wastes. Biochem. Eng. J. 76, 1–5. Méndez-Acosta, H.O., Palacios-Ruiz, B., Alcaraz-González, V., González-Álvarez, V., García-Sandoval, J.P., 2010. A robust control scheme to improve the stability of anaerobic digestion processes. J. Process Control 20, 375–383. Nagao, N., Tajima, N., Kawai, M., Niwa, C., Kurosawa, N., Matsuyama, T., Yusoff, F.M., Toda, T., 2012. Maximum organic loading rate for the single-stage wet anaerobic digestion of food waste. Bioresour. Technol. 118, 210–218. Nielsen, H.B., Uellendahl, H., Ahring, B.K., 2007. Regulation and optimization of the biogas process: propionate as a key parameter. Biomass Bioenergy 31, 820–830. Wang, Y.Y., Zhang, Y.L., Wang, J.B., Meng, L., 2009. Effects of volatile fatty acid concentrations on methane yield and methanogenic bacteria. Biomass Bioenergy 33, 848–853. Weiland, P., 2008. Wichtige Messdaten für den Prozessablauf und Stand der Technik in der Praxis. Gülzower Fachgespräche-messen, steuern, regeln bei der Biogaserzeugung. Fachagentur für Nachwachsende Rohstoffe, 17–31, e.V. 27.
Contents lists available at ScienceDirect
Bioresource Technology journal homepage: www.elsevier.com/locate/biortech
Short Communication
Early warning indicators for monitoring the process failure of anaerobic digestion system of food waste Lei Li, Qingming He, Yunmei Wei, Qin He, Xuya Peng ⇑ Key Laboratory of Three Gorges Reservoir Region’s Eco-Environment, Ministry of Education, Chongqing University, Chongqing 400045, China
h i g h l i g h t s Operation conditions significantly affect the responses of state parameters. The recommended early warning indicators are different among digesters. None of the single indicators was universally valid for all the systems. A combination of total VFA, VFA/TA and BA/TA has a general adaptability.
a r t i c l e
i n f o
Article history: Received 16 June 2014 Received in revised form 18 August 2014 Accepted 21 August 2014 Available online 28 August 2014 Keywords: Anaerobic digestion Food waste Early warning indicators Organic overloading Process stability
a b s t r a c t To determine reliable state parameters which could be used as early warning indicators of process failure due to the acidification of anaerobic digestion of food waste, three mesophilic anaerobic digesters of food waste with different operation conditions were investigated. Such parameters as gas production, methane content, pH, concentrations of volatile fatty acid (VFA), alkalinity and their combined indicators were evaluated. Results revealed that operation conditions significantly affect the responses of parameters and thus the optimal early warning indicators of each reactor differ from each other. None of the single indicators was universally valid for all the systems. The universally valid indicators should combine several parameters to supply complementary information. A combination of total VFA, the ratio of VFA to total alkalinity (VFA/TA) and the ratio of bicarbonate alkalinity to total alkalinity (BA/TA) can reflect the metabolism of the digesting system and realize rapid and effective early warning. Ó 2014 Elsevier Ltd. All rights reserved.
1. Introduction Anaerobic digestion (AD) usually features process imbalance under high organic loading rate (OLR) (Méndez-Acosta et al., 2010), so the process monitoring and control are indispensable for achieving a high-efficiency and stabilized performance of the AD system (Bjornsson et al., 2001), while effective process parameters are the basis for ensuring the process monitoring and control. A multitude of studies have explored the process parameters of the AD system. To sum up, the most commonly used parameters include pH, volatile fatty acid (VFA), alkalinity, biogas production, methane content and their combination factors (Bjornsson et al., 2001). However, previous researches on early warning parameters always focused on a specific parameter, verified its effectiveness
⇑ Corresponding author. Address: No. 174, Shapingba Zhengjie Street, Chongqing 400045, China. Tel.: +86 13594614338; fax: +86 023 65121734. E-mail address: [email protected] (X. Peng). http://dx.doi.org/10.1016/j.biortech.2014.08.089 0960-8524/Ó 2014 Elsevier Ltd. All rights reserved.
and defined its threshold value (Martín-González et al., 2013; Wang et al., 2009); or studied the responses of multiple parameters under a specific condition, and suggested the optimal parameters for this operating conditions (Nielsen et al., 2007; Kleyböcker et al., 2012; Ahring et al., 1995). Thus, though lots of researches on early warning parameters have been conducted, the proposed early warning indicators are only effective when applied to these specific operating conditions, while less effective in systems with different operation conditions. For example, the ratio of intermediate alkalinity to partial alkalinity (IA/PA) of 0.9 was suggested in order to maintain stable operation in thermophilic reactors treating sewage sludge; however IA/PA of 0.4 was proposed to assure a stable reactor performance when a potato-starch wastewater was treated; and IA/PA below 0.3 was recommended to maintain a stable operation when municipal solid waste was treated (Martín-González et al., 2013). In contrast, Hill et al. (1987) suggested that the ratio of propionate to acetate (Pr /Ac ) higher than 1.4 indicated impending digester failure, but Weiland (2008) held that the Pr /Ac shall be lower than 1.0 in a stable system; and
492
L. Li et al. / Bioresource Technology 171 (2014) 491–494
Ahring et al. (1995) suggested that Pr /Ac was useless in determining process imbalance. Other parameters are also suffering the similar phenomenon. Therefore, the indicators suggested by each study are different, as for the same warning parameter there even existed multi-thresholds. So far there still lacks of an early warning index that is universally applicable. This study is aimed at comprehensive evaluation of the validity of parameters under different operation conditions, and find out the early warning index which is universal applicable for indicating process failure of food waste (FW) digesters. Three anaerobic digesters with different operation conditions were investigated; and such process parameters including gas production and composition, pH, VFA, alkalinity and their combined indicators were all evaluated.
parameters measured daily. VS removal rate (VSr) was calculated by the same equation as reported by Koch et al. (2009). 2.3. Analytical methods
2. Methods
TS and VS were measured according to standard methods (APHA et al., 1998). pH was measured using a pH meter (Horiba, B-212). Total VFA, total alkalinity (TA) bicarbonate alkalinity (BA), intermediate alkalinity (IA) and partial alkalinity (PA) were analyzed according to Anderson and Yang (1992). Biogas production was measured using wet gas meters. Biogas composition was determined by a Biogas 5000 portable instrument (Geotech, UK). Ammonia-nitrogen was analyzed using a DR-2800 spectrophotometer (HACH, USA). Individual VFA were analyzed on a gas chromatograph equipped with flame ionization detector.
2.1. Feedstock and seeding sludge
3. Results and discussion
FW was collected from a school canteen, and with the removals of such coarse impurities as bones and plastics, the collected FW was divided into to 2 portions, and 1/3 of the FW was placed in boiling water for the removal of the surface floating oil, then it was subject to solid–liquid separation through a sieve with 2 2 mm lattice, and the obtained solid part was shredded into particles with an average size of 5.0 mm by a Robot-Coupe Shredder. The remaining 2/3 were milled directly without degreasing. Then, the two kinds of materials were all packed into 4-L plastic storage bags, and cryopreserved at 18 °C. One week prior to use, the frozen feedstock was thawed, and stored at 4 °C. The seed sludge used in this study was respectively the digested sludge taken from Dadukou Municipal Wastewater Treatment Plant, Chongqing, China (MWS); and digested sludge from an anaerobic digester treating piggery waste (PWS) and digested sludge from a 50 L anaerobic digester of FW (FWS). The characteristics of the substrate and seed sludge were listed in the Supplemental materials (SM) Table S1.
3.1. Digester performance
2.2. Digesters operation Three complete-mix anaerobic digesters each with a working volume of 20 L were operated at 36 ± 1 °C. Motorized automatic stirring was provided at the top, with an agitating frequency of 20 min h 1 and a rotary speed of 40 rpm. Different substrate pretreatment modes, seeding sludge sources and retention time were respectively adopted for the three digesters, and the specific operating conditions of the various digesters are shown in Table 1. The digesters were operated at a semi-continuous mode; substrate was added once daily and before which manual sampling was conducted at the sampling opening. For an efficient injection, one token of the substrate was mixed with twice volume of retrieved sludge from the operating reactor. During the experiment, the total solids (TS) and volatile solids (VS) of digestate were measured every three days, and the other
OLR, methane yield and VSr are usually used for the evaluation of the anaerobic digester performance (Nagao et al., 2012). Perturbations of OLR were introduced, and steady state was determined by constant methane yield and VSr in this paper. The date when sharp declines appeared in methane yield or VSr was defined as the date of the process failure. Fig. 1 manifests the responses of methane yield and VSr along with the OLR during the experiment. It can be seen that the fluctuations of VSr were slight throughout the experiment, with the specific values being 87.5 ± 3.9%, 88.9 ± 2.8% and 86.6 ± 3.4% in Reactor A, B and C, respectively. Thus they are within the recommended range of 81–92% (Nagao et al., 2012) throughout the experiment, indicating that it is infeasible to use the VSr for performance description. However, the fluctuations in methane yield were much more obvious. Approximately 50% reductions appeared on day 66 and 72 in Reactor A and C, respectively, and the same reductions occurred on day 34 and 62 in Reactor B. Accordingly, the corresponding dates of process failure were defined, which was indicated by the vertical dotted line in Fig. 1 and SM Figs. S1–S3. 3.2. Response of process parameters As with methane yield, process parameters will respond to process perturbations, and once response from a process parameter is faster than that of methane yield, then the process parameter can be used as early warning indicator. The curve diagrams of the responses of various process parameters along with OLR in three reactors were shown in SM Figs. S1–S3; and Table 2 summarized the specific early warning information of various parameters. Among then the dates of the process imbalance for IA/PA, VFA/TA and BA/TA were, respectively defined as the date when the ratios exceeded their critical value of 0.3, 0.35 and 0.8 (Martín-González et al., 2013; Méndez-Acosta et al., 2010). The
Table 1 Specific operating conditions of three digesters.
a b
1
Inoculum
Substrate
Stage
OLR (g VS L
A B
MWS PWS
Pre-treated FW Raw FW
C
FWS
Raw FW
Gradual overload Gradual overload Recovery stage Sudden overload Recovery stage Gradual overload
4, 5, 6, 7, 7.5, 8, 8.5 4, 5, 6, 7, 7.5 1, 4 4–7.1b 1 2, 3, 4, 5, 6
Retention time of each OLR. Increased 10% daily.
d
1
Digester
)
Retention timea(d) 10 7 10 1 8 15
493
L. Li et al. / Bioresource Technology 171 (2014) 491–494 Table 2 Specific early warning information of three reactors. Reactor
Parameters
Date of process failure (d)
Date of process imbalance (d)
Warning time (d)
A
BA/TA Acetate VFA IA/PA VFA/TA Propionate CH4 Pr /Ac TA Gas production pH
66
53 54 54 54 56 56 62 63 64 66
13 12 12 12 10 10 4 3 2 0
B
VFA/TA IA/PA Acetate BA/TA pH Propionate VFA Pr /Ac Gas production TA CH4
34 (62)a
IA/PA Acetate VFA BA/TA VFA/TA Propionate Pr /Ac pH CH4 Gas production TA
72
C
Fig. 1. The evolution of OLR, methane yield and VSr in R(A), R(B) and R(C) during the experiment.
selected thresholds are valid in all three reactors, and seem more suitable than any other thresholds to indicate the process imbalance of FW digesters. While for other parameters no consistent thresholds can be used, so the dates of the process imbalance were those when sudden changes in a relatively constant value occur. The warning time is the difference between the date of process failure and imbalance. As shown in Table 2, though the evaluated parameters were consistent among digesters, the warning effects of parameters were different. Generally speaking, the most insensitive indicators were gasphase indexes, pH, TA and Pr /Ac , and their warning time were not more than 5 days in all digesters. The limitations in liquidto-gas mass transfer may be the main reason of the low sensitivity in gas-phase indexes (Bjornsson et al., 2001). The delay in pH may have been caused by a high buffering capacity in FW digesters. And TA showed almost no warning effect, since an increase in VFA will cause a decrease in BA, resulting in a basically constant TA-value (Bjornsson et al., 2001). While the delay of Pr /Ac was caused by the fact that acetate exhibited faster response to the disturbance than propionate. In contrast, total VFA, acetate and propionate had featured rapider responses, their warning times range from 2 to 12 days, and the warning orders were basically consistent. Total VFA and acetate responded simultaneously, followed by propionate response. Kleyböcker et al. (2012) also drew similar conclusions, and they found acetate accumulated first due to the inhibition of the acetoclastic methanogenesis, followed by the accumulation of propionate due to the inhibition of the acetogens. Combined indicators had similar or better early warning effect compared to VFA, while their warning effects were more unstable. For example, BA/TA was the best early warning indicator in Reactor A, but it lagged back in Reactor B; the VFA/TA recommended by Reactor B also was not so sensitive as the IA/PA in Reactor C. The differences in warning order of combined indicators may have
a
66 24 26 27 30 30 31 31 31 32
0 (56) (60) (58) (60) (61) (59) (60) (60) (62)
10 (6) 8 (2) 7 (4) 4 (2) 4 (1) 3 (3) 3 (2) 3 (2) 2 (0)
34 (62) 34 (62)
0 (0) 0 (0)
65 68 68 68 68 69 70 71 72 73
7 4 4 4 4 3 2 1 0 /
74
/
The values in brackets refer to the corresponding value of sudden overload.
been caused by the different contents of ammonia among digesters. As we know, the effective buffer capacity in AD systems includes the two parts of ammonia and BA. In high ammonia system, the ammonia would react with accumulated VFA sooner than BA due to the higher pKa (Boe, 2006), resulting in an decrease in TA, while BA had no change, thus the fluctuation orders of the three combined parameters would be VFA/TA earlier than BA/TA and VFA/BA. However, when the ammonia concentration was low, and even can be ignored, the BA would be consumed directly to counteract the accumulated VFA, thus causing the fluctuation order changes to VFA/BA, which was greater than the other two parameters. In our experiments, due to the effects of the substrate and inoculum, ammonia concentration in Reactor B was higher, so the warning effect of VFA/TA was the best; the ammonia concentration in Reactor A and C was lower, and the optimal indicator in Reactor C was IA/PA, while the earliest indicator in Reactor A was BA/TA. Background fluctuations may be one of the incentives. As can be seen from Fig. S1, the warning of BA/TA was not continuous, and the ratio is greater than its threshold at day 55, while IA/PA has begun to warn at day 54. Certainly there may be other unknown factors that improve the early warning performance of BA/TA in Reactor A. In addition, it is worth mentioning that IA/PA was used to replace VFA/BA in this paper, since these two were almost equivalent (Martín-González et al., 2013), and the threshold of IA/PA was lower than that of VFA/BA, so the warning effect of IA/PA is considered theoretically better than that of VFA/BA. In addition, as show in Table 2, the warning time of parameters were different among digesters, the optimal warning time were
494
L. Li et al. / Bioresource Technology 171 (2014) 491–494
13, 10 and 6, 7 days in Reactor A–C, respectively. The differences may mainly have been caused by the different retention time under each OLR. Ahring et al. (1995) reported that most of indicators were suitable for detecting gradual overloads, but were too slow for the response of sudden overloads. Accordingly, based on the different retention time of digesters, the warning time should be that Reactor C is longer than Reactor A and B. However, Reactor C actually held the shortest warning time in disturbance of gradual overloads. This abnormality may have been caused by seeding sludge. Elbeshbishy et al. (2012) found that the digesters’ performances were not as good as other kinds of seed sludge when FWS was used as the seed sludge. While in Reactor C what was used was not only digested sludge from FW digester, but the sludge was the restored sludge after acidification, thus its performance may have been subject to the dual impacts, which may have influenced the state parameters’ response to some extent. Considering the response of the state parameters was affected by many factors, it is difficult to obtain accurate information and explicit evidence about the inconsistencies of early warning indicators among digesters. However, the inference herein is still reasonable to some degree. 3.3. Suitability of process parameters Based on the above analysis, VFA and combined indicators are more suitable for applications as early warning indicators in AD systems of FW. However, each parameter showed different warning effects in different reactors, thus the optimal indicators were inconsistent among digesters, and none of these indicators was universally valid for all the systems. To describe the characteristics of the system effectively, it is necessary to combine several indicators to supply complementary information. The unbalance between acid production and consumption was the source of process failure, so it is necessary to choose an acid concentration as one of the indicators. Theoretically, the individual VFA are the best early warning indicators; however, the monitoring of individual VFA is difficult, while the total one, which possessed similar early warning performances, can be monitored simply, thus becoming an excellent substitute of the individual VFA. Among the three combined indicators, BA/TA reflects the buffer equilibrium of the system, while VFA/TA and IA/PA represent the acid–base balance of the system. To comprehensively reflect process statues, BA/TA should be chosen as one of the indicators and in view of the advantages of VFA/TA in the high ammonia nitrogen system, VFA/TA rather than IA/PA should be chosen to reflect acid–base balance of the system. The joint use of total VFA, VFA/TA and BA/TA can reflect the metabolism of the digesting system so as to realize rapid and effective early warning. 4. Conclusions VFA and combined indicators give fast and reliable information of process state compared to other parameters. While the operation conditions significantly affect the performance of digesters and
thus the most sensitive early warning indicators are still different among digesters. None of the single indicators was universally valid for all the systems. A combination of total VFA, VFA/TA and BA/TA to supply complementary information is suitable for monitoring this type of digesters effectively. Acknowledgements This work was granted by financial support of the Key Projects in the National Science & Technology Pillar Program during the Eleventh Five-year Plan Period (No. 2010BAC67B01). Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.biortech.2014.08. 089. References Ahring, B.K., Sandberg, M., Angelidaki, I., 1995. Volatile fatty acids as indicators of process imbalance in anaerobic digesters. Appl. Microbiol. Biotechnol. 43, 559– 565. Anderson, G.K., Yang, G., 1992. Determination of bicarbonate and total volatile acid concentration in anaerobic digesters using a simple titration. Water Environ. Res. 64, 53–59. APHA, AWWA, WEF, 1998. Standard Methods for the Examination of Water and Wastewater, Washington. Bjornsson, L., Murto, M., Jantsch, T.G., Mattiasson, B., 2001. Evaluation of new methods for the monitoring of alkalinity, dissolved hydrogen and the microbial community in anaerobic digestion. Water Res. 35, 2833–2840. Boe, K., 2006. Online Monitoring and Control of the Biogas Process (Ph.D. thesis). Institute of Environments & Resources, Technical University of Denmark. Elbeshbishy, E., Nakhla, G., Hafez, H., 2012. Biochemical methane potential (BMP) of food waste and primary sludge: influence of inoculum pre-incubation and inoculum source. Bioresour. Technol. 110, 18–25. Hill, D., Cobbs, S., Bolte, J., 1987. Using volatile fatty acid relationships to predict anaerobic digester failure. Trans. ASAE 30, 496–501. Kleyböcker, A., Liebrich, M., Verstraete, W., Kraume, M., Würdemann, H., 2012. Early warning indicators for process failure due to organic overloading by rapeseed oil in one-stage continuously stirred tank reactor, sewage sludge and waste digesters. Bioresour. Technol. 123, 534–541. Koch, K., Wichern, M., Lübken, M., Horn, H., 2009. Mono fermentation of grass silage by means of loop reactors. Bioresour. Technol. 100, 5934–5940. Martín-González, L., Font, X., Vicent, T., 2013. Alkalinity ratios to identify process imbalances in anaerobic digesters treating source-sorted organic fraction of municipal wastes. Biochem. Eng. J. 76, 1–5. Méndez-Acosta, H.O., Palacios-Ruiz, B., Alcaraz-González, V., González-Álvarez, V., García-Sandoval, J.P., 2010. A robust control scheme to improve the stability of anaerobic digestion processes. J. Process Control 20, 375–383. Nagao, N., Tajima, N., Kawai, M., Niwa, C., Kurosawa, N., Matsuyama, T., Yusoff, F.M., Toda, T., 2012. Maximum organic loading rate for the single-stage wet anaerobic digestion of food waste. Bioresour. Technol. 118, 210–218. Nielsen, H.B., Uellendahl, H., Ahring, B.K., 2007. Regulation and optimization of the biogas process: propionate as a key parameter. Biomass Bioenergy 31, 820–830. Wang, Y.Y., Zhang, Y.L., Wang, J.B., Meng, L., 2009. Effects of volatile fatty acid concentrations on methane yield and methanogenic bacteria. Biomass Bioenergy 33, 848–853. Weiland, P., 2008. Wichtige Messdaten für den Prozessablauf und Stand der Technik in der Praxis. Gülzower Fachgespräche-messen, steuern, regeln bei der Biogaserzeugung. Fachagentur für Nachwachsende Rohstoffe, 17–31, e.V. 27.
