- Email: [email protected]
Simulation of the Anaerobic Digestion of Microwave Pre-Treated Waste Activated Sludge with ADM1
Simulation of the Anaerobic Digestion of Microwave Pre-Treated Waste Activated Sludge with ADM1 Joost Lauwers*, Lise Appels*, Jan Van Impe*, Raf Dewil*
*Chemical and Biochemical Process Technology and Control section, Department of Chemical Engineering, KU Leuven (University of Leuven), Heverlee, Belgium (Tel:32-16-322675; e-mail: [email protected]; [email protected]; [email protected]; [email protected]) Abstract: The disposal of sludge originating from municipal waste water treatment is a major issue and represents up to 50% of the operating costs of wastewater treatment plants (WWTP). (Appels et al., 2008) This sludge, a by-product of the treatment processes, however, has the potential to be converted into an energy rich biogas, i.e. a mixture of ca. 65% CH4 and 35% CO2, which can be utilized for the sustainable production of heat and/or electricity. Various pre-treatment methods have been suggested in literature for improving the solids reduction and biogas production rate by enhancing the digestion’s rate limiting step, i.e. organic matter hydrolysis. They all induce the solubilization of complex particulate matter so this is more rapidly and completely consumed during the anaerobic digestion process. Methods that have been shown to have a positive effect on anaerobic digestion include chemical, mechanical, biological and thermal processes. Microwave disintegration is one of the more recently applied pretreatment methods. The disintegration is caused by the combination of thermal and athermal effects and may hence provide superior results compared to a heat treatment (Eskicioglu et al., 2007). The application of mathematical models for the optimization of the digestion process is widely acknowledged. Due to the complexity of the microbial process, accurate modelling of anaerobic digestion is, however, a daunting task. A major stride countering this problem was achieved by the development of the Anaerobic Digestion Model no 1 (ADM1) by the corresponding IWA task group (Batstone et al. 2002). This model is ever since, considered to be the state of the art in modelling of anaerobic digestion and has been the platform for further refinements and numerous applications (Batstone et al. 2006). The main aim of this work is to investigate the ability of ADM1 to describe anaerobic digestion of microwave pre-treated sludge. Compared to untreated sludge, the composition and structure of the sludge is changed dramatically due to the treatment. The recorded data and digester performances were compared with the results of a modelling in ADM1 (Batstone et al., 2002). Practically, the ADM1 implementation as described by Rosen & Jeppsson (Rosen and Jeppsson (2005)) is chosen as it effectively completes the mass balance for COD, carbon and nitrogen. Keywords: anaerobic digestion, microwave, pre-treatment, ADM1
1. INTRODUCTION The use of renewable resources is of crucial importance in the current CO2-mitigation policy. Energy from biomass and waste is seen as one of the most promising future renewable energy sources, especially because a continuous power generation can be guaranteed, unlike other types such as solar and wind energy. Sewage sludge is a type of waste that can be used as a renewable energy source. Large amounts are produced during the treatment of wastewater. These amounts continue to increase due to more stringent legislation and a more efficient/thorough wastewater purification. The handling and disposal of the vast amounts of waste sludge accounts up to 50 % of the total wastewater treatment costs (Neyens et al. (2004)). Anaerobic digestion (AD) is a key factor in converting this waste into a renewable energy resource. It is a microbiological process that converts the organic fraction into an energy-rich biogas (55-70% CH4),
which can be valorized energetically (Appels et al. (2008)), e.g., in a CHP application. A major benefit is the large volume reduction of the sludge. Other beneficial features include stabilization of the sludge, inactivation and reduction of pathogens, and improvement of the sludge dewaterability (Appels et al. (2010)), which is very important for further handling after AD. As depicted in Figure 1, the process consists of four (or five if disintegration is included) subsequent stages. Each stage is governed by a different group of micro-organisms, each with their own specific operating conditions. The activity, or conversion efficiency, of the respective bacterial groups is strongly dependent on the activity of the other groups, which renders this process complex and delicate. Although this technology has been applied for several decades, there is still a lack of fundamental knowledge on the mechanism of anaerobic digestion, which is mainly due to the very high complexity of the process. As a result, the design of digester
1
systems is still generally performed by rule-of thumb (De Baere (2006)). One of the main backdraws of anaerobic digestion is the rather slow degradation of the particulate organic matter, which leads to long residence times and/or low degradation efficiencies (generally 40 % of the dry solids (DS) are degraded). This is caused by the hydrolysis step of the process, generally considered to be rate-limiting (Appels et al. (2008), Ghyoot and Verstraete (1997), Tiehm et al. (2001)). To overcome this problem, a lot of research has been conducted towards pre-treatment of the sludge, including chemical (Dewil et al. (2007), Lin et al. (2009)), mechanical (Climent et al. (2007)), enzymatic (Barjenbruch and Kopplow (2003)) and thermal (Climent et al. (2007), Appels et al. (2010)) treatments. Application of these techniques leads to a (partial) disintegration of the sludge, hereby facilitating and partly bypassing the hydrolysis step. In this way, soluble organic matter is readily available for the anaerobic micro-organisms, hence leading to a higher biogas production. In this paper, a microwave pre-treatment was applied for the enhancement of the anaerobic digestion. Microwaves are electromagnetic radiation with an oscillating frequency between 0.3 and 300 GHz and induce structural changes in the sludge flocs, micro-organisms and organic components present in the sludge, and, furthermore, generate heat (Climent et al. (2007)). According to some authors, microwaves can also induce an athermal effect in addition to their thermal effect due to dipole orientation, which results in possible breakage of hydrogen bonds and subsequently leads to the disintegration of the floc matrix (Eskicioglu et al. (2007)). In order to improve digester performance and knowledge on the process, proper dynamic modeling of anaerobic digestion is required. Since the first dynamic models of Andrews (1969) and Andrews & Graef (1971) a lot of models have been conceived. The development of the Anaerobic Digestion Model No.1 (ADM1), however, was a considerable milestone in model development for anaerobic digestion. Its structure is not revolutionary different from its predecessors, but is rather a consensus of all modeling knowledge and experience condensed in a state of the art model. The biochemical structure as it is included in ADM1, is illustrated in Figure 1.
Fig. 1. The reaction paths described in ADM1 (Batstone et al. (2002)) Since its establishment, a lot of updates and extensions have been suggested for the model. A few of them, as well as some criticisms have been noted by Batstone et al. (2006). Rosen & Jeppson (2006) discuss some issues concerning the materials balance of C and N in ADM1. They also supply a full description of implementation of ADM1 in Matlab/Simulink®. 2. MATERIALS AND METHODS 2.1 Reactor set-up and sampling In this research, 2 pilot-scale CSTR digesters, each with a 50 L working volume, were run in parallel for 50 days at 37 °C. One digester was fed with blank sludge, while the other received microwave pre-treated sludge. A photograph of the digesters is presented in Figure 2. The digesters are continuously mixed with a blade stirrer at 2 rpm. 1.25 L sludge is fed to the digesters every 12 hours, leading to a HRT of 20 days. The waste activated sludge originates from the buffer tank of a municipal wastewater treatment plant (WWTP) located in Mechelen, Belgium. After sampling, the sludge was immediately stored at 4 °C prior to further treatment and analysis. The microwave treatment was conducted in a household type microwave (Sharp). 500 mL of sludge was microwave treated at 800 W for 2.5 minutes. After the treatment, the sludge was immediately cooled in an ice bath to room temperature prior to analysis and digestion. Prior to the digestion experiment, the reactors were filled with digested sludge from the industrial scale digester of the WWTP of Antwerpen-Zuid (Belgium) and were fed during 3 HRT’s with untreated sludge to stabilize. After this period the long term experiment started.
2
detector. The operating temperatures for the injection port and the detector were 240 and 250 °C, respectively. The column temperature was: 45 °C (1 min hold), at 10 °C/min to 200 °C (3 min hold). Helium was used as the carrier gas and the inlet pressure was 10 psi. The following VFAs were analyzed: acetic acid, propionic acid, iso-butyric acid, butyric acid, iso-valeric acid, valeric acid, caproic acid. The bicarbonate content is estimated from the partial alkalinity (Ripley et al., 1986). Cations are estimated from the total alkalinity (Zaher et al., 2003). All analyses were carried out in triplicate. 2.3 ADM1 Implementation
Fig. 2. Photograph of the pilot-scale digesters 2.2 Measurements and analysis During the digestion experiment, the produced biogas was measured using drum-type gas meters (Ritter). The methane content was relatively constant at 61 vol% methane and the total methane production was calculated accordingly. Regularly, i.e. every 3 to 4 days, samples of both the influent and digester content, i.e. the effluent were taken. The dry matter content (g DM/kg sludge) of the sludge was determined as the residue after drying a sludge sample at 105 °C to constant weight. Further heating at 605 °C (to constant weight) drives off the organic dry matter (ODM), usually called ignitable matter, leaving the mineral dry matter content (MDM) as residual ash (APHA (2006)). The COD was determined using standard Nanocolor® COD 1500 test tubes (Macherey-Nagel) in a digital photometer Nanocolor® 500D (Macherey-Nagel). Ammonia/ammonium nitrogen was determined using Nanocolor® 3,50 and 200 test tubes (Macherey – Nagel). The pH of the sludge was measured using a pH-electrode (Mettler Toledo). The analysis of carbohydrates was based on the Anthrone method, described by Gerhardt et al. (1994). The amount of proteins in the sludge and the supernatant was determined using the Bicinchoninic Acid method (Smith et al., 1985). The analyses were performed on both the sludge and the supernatant (obtained by centrifugation at 12000 g and subsequent filtration through 1.3 µm glass microfibre filter paper (Merck Eurolab) to identify total and soluble fractions of each specific composition variable, respectively. The lipid content of the sludge was measured through a soxhlet extraction with petroleum ether. Prior to volatile fatty acid (VFA) analysis, the sludge samples were acidified using H2SO4. A fixed amount of an internal standard (i.e. 2methylvaleric acid) was added to the acidified sample and the VFAs were extracted with diethyl ether by shaking for two minutes. The supernatant ether-phase was transferred to a volumetric flask using a Pasteur pipette. 1 µL of the extract was subsequently injected into the GC instrument. The concentrations of each individual VFA were analyzed using a Varian 3800 gas chromatograph, equipped with a DB-FFAP (J&W Scientific) column (length 30 m, internal diameter 0.25 mm, and film thickness 0.25 µm) and a flame ionization
The equations are implemented in Matlab version 7.7 according to the approach described in Rosen & Jeppsson (2006). All reactions, apart from the calculation of pH, are implemented as ordinary differential equations (ODE). As suggested by the same authors, the acid-base equilibrium is calculated using a nested routine in which the concentrations of acetate, butyrate, valerate, propionate, ammonium and hydronium are calculated. All kinetic and stoichiometric parameters used in the model, except the ones later mentioned in the parameter estimation, are listed by Rosen & Jeppsson (2006). 2.4 Parameter estimation A basic parameter search routine was implemented in which different parameter combinations were tested. The first-order dissociation kinetic and hydrolysis constants were optimized, both for the untreated and microwave pre-treated sludge. As criterion, the sum of squared errors was used between the predicted and measured values of the biogas production.
3. RESULTS & DISCUSSION 3.1. Pre-treatment and digester performance From Table 1 it is clear that the microwave pre-treatment has a considerable effect on the solubilization of the sludge. The concentration of the soluble COD and carbohydrate concentrations increases with a factor four and 3.5 respectively. The soluble protein concentration experiences an enormous increase: up to a factor 33.5. The amount of the total organic components remains the same. This means that during pre-treatment, no components are degraded (cfr. the Maillard reaction), only transferred to the liquid phase of the sludge. Figures 3 and 4 depict the cumulative biogas production of the untreated and micro-wave treated sludge. It is clear that the treated sludge was able to produce more biogas, with an increase of 50 %.
3.2. Modeling with ADM1 3.2.1 Substrate characterization and implementation issues
3
Despite the inclusion of specific components in the model, application of ADM1 exhibits some uncertainty. This is, however, due to the incomplete characterization of the substrate or digestion mixture. To the best of the authors’ knowledge, no reports have been made of a complete characterization of all specific components included in the model. Because of the inclusion of 7 microbial groups, which can be considered as undefined selections of microorganisms, a complete characterization is in effect almost unfeasible. As a result, some assumptions are required to cope with these and other issues. Firstly, the initial biomass concentrations are estimated by simulating ADM1 through several HRTs with an average feed composition and using the default parameters given by Rosen & Jeppsson (2007). The final biomass concentrations are taken as starting points for the simulation of the runs For implementation in ADM1, all components, except for the inorganic carbon and nitrogen, are expressed in terms of COD. Conversion of the measurements is done according to Table 2, which lists the amount of COD per mg of each unit. The average deviation between the actual and the calculated sum of COD was about 10 %. The fraction ODM/DM in Table 1 is used to determine the fraction of inerts in the mixture.
concentrations. In the implementation, fractionating parameters fsI,xc, fxI,xc, fch,xc, fpr,xc, fli,xc that determine the distribution of COD and moles of C and N are made dependent on the (varying) feed.
Unit
mg COD
1 mg glucose eq
1.066
1 mg acetatic acid
1.067
1 mg butyric acid
1.820
1 mg valeric acid
2.037
1 mg propionic acid
1.512
1 mg BSAeq
1.443
1mg fats
2.90
Table 2. Conversion table
3.2.2 Results
CODtotaL (gO2/L)
45331 ± 12064
CODsoLubLe (gO2/L)
1120 ± 744
CODsoLubLe/ CODtotaL
2.9% ± 2.5%
A first observation is that the cumulative biogas production curves depicted in Figures 3 and 4 have a jagged pattern. This is attributed to the heterogeneous composition of the sludge, Microwave pre- i.e. the feed to the digesters varies in time. A few large peaks treated sludge are observed. On day 3, a large peak both for untreated and pre-treated sludge is observed. This is attributed to a 46008 ± 11853 measurement error and was omitted for the parameter estimation. After a few days, the production increased for 4401 ± 1152 both digesters. The peak at day 30 is attributed to the high 10.0% ± 3.4% COD (70 g O2/L) in the sludge feed.
DM (g/kg)
40.8 ± 7.6
40.5 ± 8.2
ODM (g/kg)
28.5 ± 5.4
28.1 ± 5.7
sugarstotaL (mg gLueq/L)
5447 ± 2835
5429 ± 2516
sugarssoluble (mg gLueq/L)
77 ± 43
285 ± 145
proteinstotaL (mg BSAeq/L)
36798 ± 12428
38986 ± 10911
proteinssoLubLe (mg BSAeq/L)
390 ± 166
13167 ± 310
Untreated sludge
Table 1. Average values of several composition variables for untreated and microwave pre-treated sludge. Another concern is the determination of the particulate matter concentration. In ADM1, the hydrolysis of sugars, proteins and lipids is preceded by a disintegration process in which particulate matter is degraded into the abovementioned particulate components. However, with the used measurements it is not possible to differentiate between both. As a consequence, the influent is considered to be completely composed of particulate matter and soluble components. The latter is of course of greater importance for the microwave solubilized sludge in which these appear in higher
After recalculating different combinations, the parameter combinations presented in Table 3 are considered to be the most optimal. They result in a RMSE of 6.0 L/d for the untreated and 9.0 L/d for the pre-treated sludge. The predictions for the untreated sludge give no exact results, but do predict the major trends in the biogas production such as the increases and declines on day 15,17 28 and 32. Additionally, the model predicts a smoother and equally distributed digestion than the observed measurements. This can be attributed to the discontinuous feeding in the set-up, which gives small but sudden surges of feed. The same phenomenon occurs for the microwave pre-treated sludge although the effect is less due to the higher overall biogas production. Visually, the simulations for the pre-treated sludge provide a better estimation of the experimental results. After the first 3 days, it follows neatly the measurements for about 35 days. After day 35 the simulation, however, overestimates the declining biogas production. No obvious explanation can be given for the discrepancy between the two (this will be the subject of further research). After recalculating different combinations, the parameter combinations presented in Table 3 are considered to be the
4
most optimal. They result in a RMSE of 6.0 l/d for the untreated and 9.0 l/d for the pre-treated sludge.
Fig. 4. Measured (-) and predicted (- -) biogas production for the microwave pre-treated sludge.
The predictions for the untreated sludge give no exact results, but do predict the major trends in the biogas production such as the increases and declines on day 15,17 28 and 32. Additionally, the model predicts a smoother and equally distributed digestion than the observed measurements. This can be attributed to the discontinuous feeding in the set-up, which gives small but sudden surges of feed. The same phenomena occurs for the microwave pre-treated sludge although the effect is lessened by the higher overall production of biogas.
US
MW
Batstone et Rosen & al., 2002 Jeppsson, 2006
kdis (d-1)
1.5
0.05
0.4
0.5
khyd,ch (d-1)
1.5
0.5
0.25
10
khyd,li (d-1)
1
0.5
0.2
10
1
0.5
0.1
10
-1
khyd,pr (d )
Table 3. Estimated parameters values for the untreated sludge (US) and the microwave pre-treated sludge (MW). As a reference, the default parameter values reported by Batstone et al., (2002) for mesophilic digestion and Rosen & Jeppsson (2006).
Visually, the simulations for the pre-treated sludge corresponds more to the measurements. After the first 3 days, it follows neatly the measurements for about 35 days. After day 35 the simulation, however, overestimates the declining biogas production. Unfortunately, no easy explanation can be given for the discrepancy between the two. The found parameter values are a bit counter-intuitive because the microwave pre-treated sludge is described by slower kinetics than the untreated sludge. However, this decrease in digestion potential is onset by the large portion of the substrate that is solubilized and as such skips this step resulting in an overall increase of the biogas production.
4. CONCLUSIONS
Fig. 3. Measured (-) and predicted (- -) biogas production for untreated sludge.
In this paper, a microwave pre-treatment was applied to waste activated sludge in order to enhance the digestion efficiency through solubilization of the organic matter present in the sludge. Various organic components were monitored: (O)DM, total and soluble COD, total and soluble protein and carbohydrate concentration, seven individual volatile fatty acids and biogas production. Subsequently, the measured data for untreated sludge and microwave-treated sludge is implemented in the ADM1. The most important conclusions that can be drawn are: x
The microwave pre-treatment was able to effectively solubilize the organic components present in the sludge. The biogas production could be increased with 50 %.
x
The substrate characterization necessary for the implementation of the ADM1 was proven to be difficult. Nevertheless, the ADM1 is a valuable tool for simulation and predicting trends, rather than exact predictions of the results.
x
The data of the microwave-treated sludge seemed to be more accurately predictable than the data of the untreated sludge.
x
The results of the parameter estimation for microwave-treated sludge indicate slower kinetics, although this effect is clearly surpassed by the large
5
solubilization of the feed. By this solubilization, the relatively slow step of disintegration and hydrolysis is skipped and should result in faster kinetics of the digestion process. The results presented in this paper are very interesting, but many questions are still unfulfilled. Future research will include the analysis of more parameters to improve the prediction results, although it should be taken into account that the identifiability of the parameters based on the measurements is doubtful. Also, the dataset of the microwave treatment results will be expanded to investigate the reproducibility of the obtained results, and to further investigate the observed trend in kinetics.
ACKNOWLEDGEMENTS The authors would like to thank the Agency for Innovation by Science and Technology in Flanders (IWT/063374 and IWT/100196) and the Research Council of the KU Leuven (PDMK/10/121) for the financial support. This work was supported in part by Project KP/09/005 (SCORES4CHEM Knowledge Platform) of the Industrial Research Council of the KU Leuven, Project PFV/10/002 Center of Excellence OPTEC-Optimization in Engineering and the Belgian Program on Interuniversity Poles of Attraction initiated by the Belgian Federal Science Policy Office. Jan Van Impe holds the chair Safety Engineering sponsored by the Belgian chemistry and life science federation essenscia.
REFERENCES Andrews, J.F. (1969). Dynamic model of the anaerobic digestion model. Journal of the Sanitary Engineering Division; Proceedings of the American Society of Civil Engineers, 1, 95-116. Andrews, J.F. and Graef, S.P. (1971). Dynamic modeling and simulation of the anaerobic digestion process. Advances in Chemistry Series, 105, 126-162. Angelidaki, I., and Ellegaard, L. (2003). Codigestion of manure and organic wastes in centralized biogas plants – Status and future trends. Applied Biochemistry and Biotechnology, 109 (1-3), 95-105. Appels, L. Baeyens, J., Degrève, J. and Dewil, R. (2008). Principles and potential of the anaerobic digestion of waste-activated sludge. Progress in Energy and Combustion, 34 (6), 755-781. Appels, L., Degrève, J., Van der Bruggen, B., Van Impe, J. and Dewil, R. (2010). Influence of low temperature thermal pre-treatment on sludge solubilisation, heavy metal release and anaerobic digestion. Bioresource technology, 101 (15), 5743-5748. Barjenbruch, M. and Kopplow, O. (2003). Enzymatic, mechanical and thermal pre-treatment of surplus sludge. Advances in Environmental Research, 7, 715-720. Batstone, D.J., Keller, J., Angelidaki, R.I., Kalyuzhnyi, S.V., Pavlostathis, S.G., Rozzi, A., Sanders, W.T.M., Siegrist, H. and Vavilin, V.A. (2002). The Anaerobic Digestion
Model No 1 (ADM1), Water Science and Technology, 45 (10), 65-73. Batstone, D.J., Keller, J. and Steyer, J.P. (2006). A review of ADM1 extensions, applications and analysis: 2002-2005. Water Science & Technology, 54(4), 1-10. Climent, M., Ferrer, I., Baeza, M.D., Artola, A., Vazquez, F. and Font, X. (2007). Effects of thermal and mechanical pre-treatments of secondary sludge on biogas production under thermophilic conditions. Chemical Engineering Journal, 133, 335-342. De Baere, L. (2006). Will anaerobic digestion of solid waste survive in the future? Water Science and Technology, 53 (8), 187-194. Dewil, R., Appels, L., Baeyens, J. and Degrève, J. (2007). Peroxidation enhances the biogas production in the anaerobic digestion of biosolids. Journal of Hazardous Materials, 146, 577-581. Gerhardt, P., Murray, R.G.E., Wood, W.A. and Krieg, N.R. (1994). Methods for general and molecular bacteriology, ASM, Washington DC. Ghyoot, W. and Verstraete, W. (1997). Anaerobic digestion of primary sludge from chemical pre-precipitation. Water Science and Technology, 36, 357-365. Lin, T., Wang, D., Wu, S. and Wang, C. (2009). Alkali pretreatment enhances biogas production in the anaerobic digestion of pulp and paper sludge. Journal of Hazardous Materials, 170, 366-373. Neyens, E., Baeyens, J., Dewil, R. and De Heyder, B. (2004). Advanced sludge treatment affects extracellular polymeric substances to improve activated sludge dewatering. Journal of Hazardous Materials, 106 (2-3), 83-92. Rosen, C. and Jeppsson, U. (2006). Aspects on ADM1 Implementation within the BSM2 framework. IWA BSM TG Technical Report (available at www.benchmarkwwtp.org). Ripley, L.E., Boyle, W.C. and Converse J.C., (1986). Improved alkalimetric monitoring for Anaerobic Digestion of high-strength wastes. Journal WPCF, 58 (5), 406-411. Smith, P.K., Krohn, R.I., Hermanson, G.T. and Mallia, A.K. (1985). Measurement of protein using bicinchoninic acid. Analytical Biochemistry. 150, 76-85. APHA (2006). Standard methods for the examination of water and wastewater, 18th ed. American Public Health Association, Washington DC. Tiehm, A., Nickel, K., Zellhorn, M. and Neis, U. (2001). Ultrasonic waste activated sludge disintegration for improving anaerobic stabilization. Water Research, 35, 2003-2009. Zaher, U, Rodríguez, J., Franco, A. and Vanrolleghem, P.A. (2003). Application of the IWA ADM1 model to simulate anaerobic digester dynamics using a concise set of practical measurements. IWA Conference on Environmental Biotechnology - Advancement on Water and Wastewater Applications in the Tropics, December 9th – 10th, 2003, Kuala Lumpur (Malaysia).
6
*Chemical and Biochemical Process Technology and Control section, Department of Chemical Engineering, KU Leuven (University of Leuven), Heverlee, Belgium (Tel:32-16-322675; e-mail: [email protected]; [email protected]; [email protected]; [email protected]) Abstract: The disposal of sludge originating from municipal waste water treatment is a major issue and represents up to 50% of the operating costs of wastewater treatment plants (WWTP). (Appels et al., 2008) This sludge, a by-product of the treatment processes, however, has the potential to be converted into an energy rich biogas, i.e. a mixture of ca. 65% CH4 and 35% CO2, which can be utilized for the sustainable production of heat and/or electricity. Various pre-treatment methods have been suggested in literature for improving the solids reduction and biogas production rate by enhancing the digestion’s rate limiting step, i.e. organic matter hydrolysis. They all induce the solubilization of complex particulate matter so this is more rapidly and completely consumed during the anaerobic digestion process. Methods that have been shown to have a positive effect on anaerobic digestion include chemical, mechanical, biological and thermal processes. Microwave disintegration is one of the more recently applied pretreatment methods. The disintegration is caused by the combination of thermal and athermal effects and may hence provide superior results compared to a heat treatment (Eskicioglu et al., 2007). The application of mathematical models for the optimization of the digestion process is widely acknowledged. Due to the complexity of the microbial process, accurate modelling of anaerobic digestion is, however, a daunting task. A major stride countering this problem was achieved by the development of the Anaerobic Digestion Model no 1 (ADM1) by the corresponding IWA task group (Batstone et al. 2002). This model is ever since, considered to be the state of the art in modelling of anaerobic digestion and has been the platform for further refinements and numerous applications (Batstone et al. 2006). The main aim of this work is to investigate the ability of ADM1 to describe anaerobic digestion of microwave pre-treated sludge. Compared to untreated sludge, the composition and structure of the sludge is changed dramatically due to the treatment. The recorded data and digester performances were compared with the results of a modelling in ADM1 (Batstone et al., 2002). Practically, the ADM1 implementation as described by Rosen & Jeppsson (Rosen and Jeppsson (2005)) is chosen as it effectively completes the mass balance for COD, carbon and nitrogen. Keywords: anaerobic digestion, microwave, pre-treatment, ADM1
1. INTRODUCTION The use of renewable resources is of crucial importance in the current CO2-mitigation policy. Energy from biomass and waste is seen as one of the most promising future renewable energy sources, especially because a continuous power generation can be guaranteed, unlike other types such as solar and wind energy. Sewage sludge is a type of waste that can be used as a renewable energy source. Large amounts are produced during the treatment of wastewater. These amounts continue to increase due to more stringent legislation and a more efficient/thorough wastewater purification. The handling and disposal of the vast amounts of waste sludge accounts up to 50 % of the total wastewater treatment costs (Neyens et al. (2004)). Anaerobic digestion (AD) is a key factor in converting this waste into a renewable energy resource. It is a microbiological process that converts the organic fraction into an energy-rich biogas (55-70% CH4),
which can be valorized energetically (Appels et al. (2008)), e.g., in a CHP application. A major benefit is the large volume reduction of the sludge. Other beneficial features include stabilization of the sludge, inactivation and reduction of pathogens, and improvement of the sludge dewaterability (Appels et al. (2010)), which is very important for further handling after AD. As depicted in Figure 1, the process consists of four (or five if disintegration is included) subsequent stages. Each stage is governed by a different group of micro-organisms, each with their own specific operating conditions. The activity, or conversion efficiency, of the respective bacterial groups is strongly dependent on the activity of the other groups, which renders this process complex and delicate. Although this technology has been applied for several decades, there is still a lack of fundamental knowledge on the mechanism of anaerobic digestion, which is mainly due to the very high complexity of the process. As a result, the design of digester
1
systems is still generally performed by rule-of thumb (De Baere (2006)). One of the main backdraws of anaerobic digestion is the rather slow degradation of the particulate organic matter, which leads to long residence times and/or low degradation efficiencies (generally 40 % of the dry solids (DS) are degraded). This is caused by the hydrolysis step of the process, generally considered to be rate-limiting (Appels et al. (2008), Ghyoot and Verstraete (1997), Tiehm et al. (2001)). To overcome this problem, a lot of research has been conducted towards pre-treatment of the sludge, including chemical (Dewil et al. (2007), Lin et al. (2009)), mechanical (Climent et al. (2007)), enzymatic (Barjenbruch and Kopplow (2003)) and thermal (Climent et al. (2007), Appels et al. (2010)) treatments. Application of these techniques leads to a (partial) disintegration of the sludge, hereby facilitating and partly bypassing the hydrolysis step. In this way, soluble organic matter is readily available for the anaerobic micro-organisms, hence leading to a higher biogas production. In this paper, a microwave pre-treatment was applied for the enhancement of the anaerobic digestion. Microwaves are electromagnetic radiation with an oscillating frequency between 0.3 and 300 GHz and induce structural changes in the sludge flocs, micro-organisms and organic components present in the sludge, and, furthermore, generate heat (Climent et al. (2007)). According to some authors, microwaves can also induce an athermal effect in addition to their thermal effect due to dipole orientation, which results in possible breakage of hydrogen bonds and subsequently leads to the disintegration of the floc matrix (Eskicioglu et al. (2007)). In order to improve digester performance and knowledge on the process, proper dynamic modeling of anaerobic digestion is required. Since the first dynamic models of Andrews (1969) and Andrews & Graef (1971) a lot of models have been conceived. The development of the Anaerobic Digestion Model No.1 (ADM1), however, was a considerable milestone in model development for anaerobic digestion. Its structure is not revolutionary different from its predecessors, but is rather a consensus of all modeling knowledge and experience condensed in a state of the art model. The biochemical structure as it is included in ADM1, is illustrated in Figure 1.
Fig. 1. The reaction paths described in ADM1 (Batstone et al. (2002)) Since its establishment, a lot of updates and extensions have been suggested for the model. A few of them, as well as some criticisms have been noted by Batstone et al. (2006). Rosen & Jeppson (2006) discuss some issues concerning the materials balance of C and N in ADM1. They also supply a full description of implementation of ADM1 in Matlab/Simulink®. 2. MATERIALS AND METHODS 2.1 Reactor set-up and sampling In this research, 2 pilot-scale CSTR digesters, each with a 50 L working volume, were run in parallel for 50 days at 37 °C. One digester was fed with blank sludge, while the other received microwave pre-treated sludge. A photograph of the digesters is presented in Figure 2. The digesters are continuously mixed with a blade stirrer at 2 rpm. 1.25 L sludge is fed to the digesters every 12 hours, leading to a HRT of 20 days. The waste activated sludge originates from the buffer tank of a municipal wastewater treatment plant (WWTP) located in Mechelen, Belgium. After sampling, the sludge was immediately stored at 4 °C prior to further treatment and analysis. The microwave treatment was conducted in a household type microwave (Sharp). 500 mL of sludge was microwave treated at 800 W for 2.5 minutes. After the treatment, the sludge was immediately cooled in an ice bath to room temperature prior to analysis and digestion. Prior to the digestion experiment, the reactors were filled with digested sludge from the industrial scale digester of the WWTP of Antwerpen-Zuid (Belgium) and were fed during 3 HRT’s with untreated sludge to stabilize. After this period the long term experiment started.
2
detector. The operating temperatures for the injection port and the detector were 240 and 250 °C, respectively. The column temperature was: 45 °C (1 min hold), at 10 °C/min to 200 °C (3 min hold). Helium was used as the carrier gas and the inlet pressure was 10 psi. The following VFAs were analyzed: acetic acid, propionic acid, iso-butyric acid, butyric acid, iso-valeric acid, valeric acid, caproic acid. The bicarbonate content is estimated from the partial alkalinity (Ripley et al., 1986). Cations are estimated from the total alkalinity (Zaher et al., 2003). All analyses were carried out in triplicate. 2.3 ADM1 Implementation
Fig. 2. Photograph of the pilot-scale digesters 2.2 Measurements and analysis During the digestion experiment, the produced biogas was measured using drum-type gas meters (Ritter). The methane content was relatively constant at 61 vol% methane and the total methane production was calculated accordingly. Regularly, i.e. every 3 to 4 days, samples of both the influent and digester content, i.e. the effluent were taken. The dry matter content (g DM/kg sludge) of the sludge was determined as the residue after drying a sludge sample at 105 °C to constant weight. Further heating at 605 °C (to constant weight) drives off the organic dry matter (ODM), usually called ignitable matter, leaving the mineral dry matter content (MDM) as residual ash (APHA (2006)). The COD was determined using standard Nanocolor® COD 1500 test tubes (Macherey-Nagel) in a digital photometer Nanocolor® 500D (Macherey-Nagel). Ammonia/ammonium nitrogen was determined using Nanocolor® 3,50 and 200 test tubes (Macherey – Nagel). The pH of the sludge was measured using a pH-electrode (Mettler Toledo). The analysis of carbohydrates was based on the Anthrone method, described by Gerhardt et al. (1994). The amount of proteins in the sludge and the supernatant was determined using the Bicinchoninic Acid method (Smith et al., 1985). The analyses were performed on both the sludge and the supernatant (obtained by centrifugation at 12000 g and subsequent filtration through 1.3 µm glass microfibre filter paper (Merck Eurolab) to identify total and soluble fractions of each specific composition variable, respectively. The lipid content of the sludge was measured through a soxhlet extraction with petroleum ether. Prior to volatile fatty acid (VFA) analysis, the sludge samples were acidified using H2SO4. A fixed amount of an internal standard (i.e. 2methylvaleric acid) was added to the acidified sample and the VFAs were extracted with diethyl ether by shaking for two minutes. The supernatant ether-phase was transferred to a volumetric flask using a Pasteur pipette. 1 µL of the extract was subsequently injected into the GC instrument. The concentrations of each individual VFA were analyzed using a Varian 3800 gas chromatograph, equipped with a DB-FFAP (J&W Scientific) column (length 30 m, internal diameter 0.25 mm, and film thickness 0.25 µm) and a flame ionization
The equations are implemented in Matlab version 7.7 according to the approach described in Rosen & Jeppsson (2006). All reactions, apart from the calculation of pH, are implemented as ordinary differential equations (ODE). As suggested by the same authors, the acid-base equilibrium is calculated using a nested routine in which the concentrations of acetate, butyrate, valerate, propionate, ammonium and hydronium are calculated. All kinetic and stoichiometric parameters used in the model, except the ones later mentioned in the parameter estimation, are listed by Rosen & Jeppsson (2006). 2.4 Parameter estimation A basic parameter search routine was implemented in which different parameter combinations were tested. The first-order dissociation kinetic and hydrolysis constants were optimized, both for the untreated and microwave pre-treated sludge. As criterion, the sum of squared errors was used between the predicted and measured values of the biogas production.
3. RESULTS & DISCUSSION 3.1. Pre-treatment and digester performance From Table 1 it is clear that the microwave pre-treatment has a considerable effect on the solubilization of the sludge. The concentration of the soluble COD and carbohydrate concentrations increases with a factor four and 3.5 respectively. The soluble protein concentration experiences an enormous increase: up to a factor 33.5. The amount of the total organic components remains the same. This means that during pre-treatment, no components are degraded (cfr. the Maillard reaction), only transferred to the liquid phase of the sludge. Figures 3 and 4 depict the cumulative biogas production of the untreated and micro-wave treated sludge. It is clear that the treated sludge was able to produce more biogas, with an increase of 50 %.
3.2. Modeling with ADM1 3.2.1 Substrate characterization and implementation issues
3
Despite the inclusion of specific components in the model, application of ADM1 exhibits some uncertainty. This is, however, due to the incomplete characterization of the substrate or digestion mixture. To the best of the authors’ knowledge, no reports have been made of a complete characterization of all specific components included in the model. Because of the inclusion of 7 microbial groups, which can be considered as undefined selections of microorganisms, a complete characterization is in effect almost unfeasible. As a result, some assumptions are required to cope with these and other issues. Firstly, the initial biomass concentrations are estimated by simulating ADM1 through several HRTs with an average feed composition and using the default parameters given by Rosen & Jeppsson (2007). The final biomass concentrations are taken as starting points for the simulation of the runs For implementation in ADM1, all components, except for the inorganic carbon and nitrogen, are expressed in terms of COD. Conversion of the measurements is done according to Table 2, which lists the amount of COD per mg of each unit. The average deviation between the actual and the calculated sum of COD was about 10 %. The fraction ODM/DM in Table 1 is used to determine the fraction of inerts in the mixture.
concentrations. In the implementation, fractionating parameters fsI,xc, fxI,xc, fch,xc, fpr,xc, fli,xc that determine the distribution of COD and moles of C and N are made dependent on the (varying) feed.
Unit
mg COD
1 mg glucose eq
1.066
1 mg acetatic acid
1.067
1 mg butyric acid
1.820
1 mg valeric acid
2.037
1 mg propionic acid
1.512
1 mg BSAeq
1.443
1mg fats
2.90
Table 2. Conversion table
3.2.2 Results
CODtotaL (gO2/L)
45331 ± 12064
CODsoLubLe (gO2/L)
1120 ± 744
CODsoLubLe/ CODtotaL
2.9% ± 2.5%
A first observation is that the cumulative biogas production curves depicted in Figures 3 and 4 have a jagged pattern. This is attributed to the heterogeneous composition of the sludge, Microwave pre- i.e. the feed to the digesters varies in time. A few large peaks treated sludge are observed. On day 3, a large peak both for untreated and pre-treated sludge is observed. This is attributed to a 46008 ± 11853 measurement error and was omitted for the parameter estimation. After a few days, the production increased for 4401 ± 1152 both digesters. The peak at day 30 is attributed to the high 10.0% ± 3.4% COD (70 g O2/L) in the sludge feed.
DM (g/kg)
40.8 ± 7.6
40.5 ± 8.2
ODM (g/kg)
28.5 ± 5.4
28.1 ± 5.7
sugarstotaL (mg gLueq/L)
5447 ± 2835
5429 ± 2516
sugarssoluble (mg gLueq/L)
77 ± 43
285 ± 145
proteinstotaL (mg BSAeq/L)
36798 ± 12428
38986 ± 10911
proteinssoLubLe (mg BSAeq/L)
390 ± 166
13167 ± 310
Untreated sludge
Table 1. Average values of several composition variables for untreated and microwave pre-treated sludge. Another concern is the determination of the particulate matter concentration. In ADM1, the hydrolysis of sugars, proteins and lipids is preceded by a disintegration process in which particulate matter is degraded into the abovementioned particulate components. However, with the used measurements it is not possible to differentiate between both. As a consequence, the influent is considered to be completely composed of particulate matter and soluble components. The latter is of course of greater importance for the microwave solubilized sludge in which these appear in higher
After recalculating different combinations, the parameter combinations presented in Table 3 are considered to be the most optimal. They result in a RMSE of 6.0 L/d for the untreated and 9.0 L/d for the pre-treated sludge. The predictions for the untreated sludge give no exact results, but do predict the major trends in the biogas production such as the increases and declines on day 15,17 28 and 32. Additionally, the model predicts a smoother and equally distributed digestion than the observed measurements. This can be attributed to the discontinuous feeding in the set-up, which gives small but sudden surges of feed. The same phenomenon occurs for the microwave pre-treated sludge although the effect is less due to the higher overall biogas production. Visually, the simulations for the pre-treated sludge provide a better estimation of the experimental results. After the first 3 days, it follows neatly the measurements for about 35 days. After day 35 the simulation, however, overestimates the declining biogas production. No obvious explanation can be given for the discrepancy between the two (this will be the subject of further research). After recalculating different combinations, the parameter combinations presented in Table 3 are considered to be the
4
most optimal. They result in a RMSE of 6.0 l/d for the untreated and 9.0 l/d for the pre-treated sludge.
Fig. 4. Measured (-) and predicted (- -) biogas production for the microwave pre-treated sludge.
The predictions for the untreated sludge give no exact results, but do predict the major trends in the biogas production such as the increases and declines on day 15,17 28 and 32. Additionally, the model predicts a smoother and equally distributed digestion than the observed measurements. This can be attributed to the discontinuous feeding in the set-up, which gives small but sudden surges of feed. The same phenomena occurs for the microwave pre-treated sludge although the effect is lessened by the higher overall production of biogas.
US
MW
Batstone et Rosen & al., 2002 Jeppsson, 2006
kdis (d-1)
1.5
0.05
0.4
0.5
khyd,ch (d-1)
1.5
0.5
0.25
10
khyd,li (d-1)
1
0.5
0.2
10
1
0.5
0.1
10
-1
khyd,pr (d )
Table 3. Estimated parameters values for the untreated sludge (US) and the microwave pre-treated sludge (MW). As a reference, the default parameter values reported by Batstone et al., (2002) for mesophilic digestion and Rosen & Jeppsson (2006).
Visually, the simulations for the pre-treated sludge corresponds more to the measurements. After the first 3 days, it follows neatly the measurements for about 35 days. After day 35 the simulation, however, overestimates the declining biogas production. Unfortunately, no easy explanation can be given for the discrepancy between the two. The found parameter values are a bit counter-intuitive because the microwave pre-treated sludge is described by slower kinetics than the untreated sludge. However, this decrease in digestion potential is onset by the large portion of the substrate that is solubilized and as such skips this step resulting in an overall increase of the biogas production.
4. CONCLUSIONS
Fig. 3. Measured (-) and predicted (- -) biogas production for untreated sludge.
In this paper, a microwave pre-treatment was applied to waste activated sludge in order to enhance the digestion efficiency through solubilization of the organic matter present in the sludge. Various organic components were monitored: (O)DM, total and soluble COD, total and soluble protein and carbohydrate concentration, seven individual volatile fatty acids and biogas production. Subsequently, the measured data for untreated sludge and microwave-treated sludge is implemented in the ADM1. The most important conclusions that can be drawn are: x
The microwave pre-treatment was able to effectively solubilize the organic components present in the sludge. The biogas production could be increased with 50 %.
x
The substrate characterization necessary for the implementation of the ADM1 was proven to be difficult. Nevertheless, the ADM1 is a valuable tool for simulation and predicting trends, rather than exact predictions of the results.
x
The data of the microwave-treated sludge seemed to be more accurately predictable than the data of the untreated sludge.
x
The results of the parameter estimation for microwave-treated sludge indicate slower kinetics, although this effect is clearly surpassed by the large
5
solubilization of the feed. By this solubilization, the relatively slow step of disintegration and hydrolysis is skipped and should result in faster kinetics of the digestion process. The results presented in this paper are very interesting, but many questions are still unfulfilled. Future research will include the analysis of more parameters to improve the prediction results, although it should be taken into account that the identifiability of the parameters based on the measurements is doubtful. Also, the dataset of the microwave treatment results will be expanded to investigate the reproducibility of the obtained results, and to further investigate the observed trend in kinetics.
ACKNOWLEDGEMENTS The authors would like to thank the Agency for Innovation by Science and Technology in Flanders (IWT/063374 and IWT/100196) and the Research Council of the KU Leuven (PDMK/10/121) for the financial support. This work was supported in part by Project KP/09/005 (SCORES4CHEM Knowledge Platform) of the Industrial Research Council of the KU Leuven, Project PFV/10/002 Center of Excellence OPTEC-Optimization in Engineering and the Belgian Program on Interuniversity Poles of Attraction initiated by the Belgian Federal Science Policy Office. Jan Van Impe holds the chair Safety Engineering sponsored by the Belgian chemistry and life science federation essenscia.
REFERENCES Andrews, J.F. (1969). Dynamic model of the anaerobic digestion model. Journal of the Sanitary Engineering Division; Proceedings of the American Society of Civil Engineers, 1, 95-116. Andrews, J.F. and Graef, S.P. (1971). Dynamic modeling and simulation of the anaerobic digestion process. Advances in Chemistry Series, 105, 126-162. Angelidaki, I., and Ellegaard, L. (2003). Codigestion of manure and organic wastes in centralized biogas plants – Status and future trends. Applied Biochemistry and Biotechnology, 109 (1-3), 95-105. Appels, L. Baeyens, J., Degrève, J. and Dewil, R. (2008). Principles and potential of the anaerobic digestion of waste-activated sludge. Progress in Energy and Combustion, 34 (6), 755-781. Appels, L., Degrève, J., Van der Bruggen, B., Van Impe, J. and Dewil, R. (2010). Influence of low temperature thermal pre-treatment on sludge solubilisation, heavy metal release and anaerobic digestion. Bioresource technology, 101 (15), 5743-5748. Barjenbruch, M. and Kopplow, O. (2003). Enzymatic, mechanical and thermal pre-treatment of surplus sludge. Advances in Environmental Research, 7, 715-720. Batstone, D.J., Keller, J., Angelidaki, R.I., Kalyuzhnyi, S.V., Pavlostathis, S.G., Rozzi, A., Sanders, W.T.M., Siegrist, H. and Vavilin, V.A. (2002). The Anaerobic Digestion
Model No 1 (ADM1), Water Science and Technology, 45 (10), 65-73. Batstone, D.J., Keller, J. and Steyer, J.P. (2006). A review of ADM1 extensions, applications and analysis: 2002-2005. Water Science & Technology, 54(4), 1-10. Climent, M., Ferrer, I., Baeza, M.D., Artola, A., Vazquez, F. and Font, X. (2007). Effects of thermal and mechanical pre-treatments of secondary sludge on biogas production under thermophilic conditions. Chemical Engineering Journal, 133, 335-342. De Baere, L. (2006). Will anaerobic digestion of solid waste survive in the future? Water Science and Technology, 53 (8), 187-194. Dewil, R., Appels, L., Baeyens, J. and Degrève, J. (2007). Peroxidation enhances the biogas production in the anaerobic digestion of biosolids. Journal of Hazardous Materials, 146, 577-581. Gerhardt, P., Murray, R.G.E., Wood, W.A. and Krieg, N.R. (1994). Methods for general and molecular bacteriology, ASM, Washington DC. Ghyoot, W. and Verstraete, W. (1997). Anaerobic digestion of primary sludge from chemical pre-precipitation. Water Science and Technology, 36, 357-365. Lin, T., Wang, D., Wu, S. and Wang, C. (2009). Alkali pretreatment enhances biogas production in the anaerobic digestion of pulp and paper sludge. Journal of Hazardous Materials, 170, 366-373. Neyens, E., Baeyens, J., Dewil, R. and De Heyder, B. (2004). Advanced sludge treatment affects extracellular polymeric substances to improve activated sludge dewatering. Journal of Hazardous Materials, 106 (2-3), 83-92. Rosen, C. and Jeppsson, U. (2006). Aspects on ADM1 Implementation within the BSM2 framework. IWA BSM TG Technical Report (available at www.benchmarkwwtp.org). Ripley, L.E., Boyle, W.C. and Converse J.C., (1986). Improved alkalimetric monitoring for Anaerobic Digestion of high-strength wastes. Journal WPCF, 58 (5), 406-411. Smith, P.K., Krohn, R.I., Hermanson, G.T. and Mallia, A.K. (1985). Measurement of protein using bicinchoninic acid. Analytical Biochemistry. 150, 76-85. APHA (2006). Standard methods for the examination of water and wastewater, 18th ed. American Public Health Association, Washington DC. Tiehm, A., Nickel, K., Zellhorn, M. and Neis, U. (2001). Ultrasonic waste activated sludge disintegration for improving anaerobic stabilization. Water Research, 35, 2003-2009. Zaher, U, Rodríguez, J., Franco, A. and Vanrolleghem, P.A. (2003). Application of the IWA ADM1 model to simulate anaerobic digester dynamics using a concise set of practical measurements. IWA Conference on Environmental Biotechnology - Advancement on Water and Wastewater Applications in the Tropics, December 9th – 10th, 2003, Kuala Lumpur (Malaysia).
6