Available online at www.sciencedirect.com
ScienceDirect Energy Procedia 105 (2017) 256 – 262
The 8th International Conference on Applied Energy – ICAE2016
Mathematical Modelling and Statistical Approach to Assess the Performance of Anaerobic Fixed Bed Reactor for Biogas Production from Piyungan Sanitary Landfill Leachate a Hanifrahmawan Sudibyo a,b, Zata Lini Shabrina , Lenny Halima, Wiratni a,b, Budhijanto * a
Chemical Engineering Department, Gadjah Mada University, Jalan Grafika No. 2 Yogyakarta 55281, Indonesia b Center for Energy Studies, Gadjah Mada University, Jalan Sekip UGM K-1A Yogyakarta 55281, Indonesia
Abstract Sanitary landfill method to treat solid waste offered cheaper operational cost but created environmental problems, one of which was leachate accumulation. Treatment of leachate to meet the environmental standard for disposal to the water bodies was mandatory by regulation. One way to treat leachate is using anaerobic bioreactor to digest the leachate and to produce biogas besides cleaner water to be disposed. To achieve stability in the bioreactor performance, zeolite was as microbial immobilization media. Zeolite was proven to be useful in stabilizing the growth of the microorganism and reducing the inhibitors in the leachate. The growth behavior of the microorganism before and after the zeolite addition was identified through the mathematical modelling and the statistical approach. © 2017 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
© 2016 The Authors. Published by Elsevier Ltd. (http://creativecommons.org/licenses/by-nc-nd/4.0/). Selection peer-review of under responsibility of ICAE Peer-reviewand/or under responsibility the scientific committee of the 8th International Conference on Applied Energy. Keywords: leachate; anaerobic fixed bed reactor; mathematical modelling; immobilization; zeolite; statistical approach
1. Introduction The modernizat ion of lifestyles, continuous development of industry, and commercial growth in economic sector has been accompanied by rapid increases in both the municipal and industrial solid waste production. The sanitary landfill method for the final d isposal of solid waste material continues to be widely accepted and used due to its economic advantages. Co mparative studies of the various possible means of eliminating solid u rban waste (landfilling, incinerat ion, etc.) have shown that the cheapest, in term of exp loitation and capital costs, is landfilling [1]. Unfortunately, this method results in the *Corresponding author. T el.: +6281328183160. E-mail address:
[email protected].
1876-6102 © 2017 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Peer-review under responsibility of the scientific committee of the 8th International Conference on Applied Energy. doi:10.1016/j.egypro.2017.03.311
257
Hanifrahmawan Sudibyo et al. / Energy Procedia 105 (2017) 256 – 262
undesirable side product, which was leachate. Leachate is defined as the aqueous effluent generated as a consequence of rainwater percolation through wastes, biochemical processes in waste’s cells and the inherent water content of wastes themselves. The leachate generated from a landfill site containing organic, inorganic, and heavy metal co mpounds (Zn and Hg) has a co mplex mixt ure with a foul odour. The flo w rate and the co mposition of the sanitary landfill leachate vary depending on site, season, and age of the landfill [2]. Leachate generated from Piyungan Sanitary Landfill, Yogyakarta, Indonesia contains high organic load (sCOD 2,800-3,500 mg.L-1 ), h igh volatile fatty acid (VFA) content (970-1,000 mg.L-1 ), and high ammonia content (500 – 800 mg.L-1 ). Leachate treat ment using anaerobic digestion offers some advantages [3] and also the barrier wh ich was the slow growth of microorganisms so that the using of conventional anaerobic digester needed huge volume o f digester wh ich was not economical. Besides, wash out often happened at the conventional anaerobic digester due to the high flowrate. The solution for stabilizing and maxi mizing the microorganis m’s growth in line with preventing the wash out problem is to use media to immob ilize the bacteria [4,5]. By using fixed bed anaerobic digester, the possibility of the bacteria to be washed out can be minimized. One potential material as the immobilization media is natural zeo lite. Natural zeolite’s pore diameter was 3-10 Ǻ, with the average surface area is 24.9 m2 /g appro ximately, and the void volu me is about 29 – 39% [6]. The attachment of the microorganisms to the zeolite surface t o form b iofilm is supported by the nature of natural zeolite which is rich in Na +, Ca2+, and Mg 2+ cation concentration [7]. Nomenclature μm1
[day -1]
maximum specific growth rate of acidogenic cell
μm2
[day -1]
maximum specific growth rate of methanogenic cell
KSX1
[mg sCOD/mg acidogenic cell]
half-saturation constant associated with sCOD
KSX2
[mg VFA/mg methanogenic cell]
half-saturation constant associated with VFA
YX1/sCOD
[mg acidogenic cell/mg sCOD]
yield of cell formation per mg sCOD reduction
YX2/VFA
[mg methanogenic cell/mg VFA]
yield of cell formation per mg VFA reduction
YCH4/X2
[mL CH4/g sCOD o]/[mg methanogenic cell/L]
yield of cumulative CH4 formation per cell formation
YVFA/X1
[mg VFA/mg acidogenic cell]
yield of VFA formation per mg acidogenic cell
KI
[mg VFA/L]
inhibition constant associated with VFA
The reactor performance was studied by varying the weight of zeo lite used inside the fixed bed. Afterward, the both mathematical and statistical approaches were required to scale-up the process in the bioreactor design in the future. Contois and Haldane kinetics were used in the mathematical approach to simu late the behaviour of the microbial growth rate because these equations were commonly used in the wastewater treat ment systems [5,8,9]. In the statistical approach, Pearson correlation coefficient (Eq . 1) was chosen as the tool. Fro m each weight of zeolite used in the fixed bed, there would be some constants derived fro m the mathemat ical equations. The correlation between the zeolite additions to the reaction kinetic can be identified fro m the Pearson correlat ion coefficient. This work aimed to identify the correlation between the various weights of natural zeo lite as the immob ilization media in the anaerobic fixed bed digester with the reactor performance and the Contois’s and Haldane’s kinetic constants.
r
^ n.¦ x
n.¦ x. y ¦ x.¦ y
2
¦ x . n.¦ y ¦ y 2
2
2
`
0,5
(1)
258
Hanifrahmawan Sudibyo et al. / Energy Procedia 105 (2017) 256 – 262
To determine the kinetics constants, a set of ordinary differential equations for the organic matter concentration as acidogenic cell (X1 ), methanogenic cell (X2 ), substrate (CsCOD ), volatile fatty acid (CVFA ), and methane (CCH4 ) was developed in Eq. 2, Eq.3, Eq. 4, Eq . 5, and 6. These differential equations were solved numerically and the corresponding reaction rate constants was determined by minimizing the Su m of Square of Erro r (SSE) between calculated and experimental data of sCOD, VFA, and CH4 in batch reactor.
dX1 dt dX 2 dt
dCsCOD dt dCVFA dt dCCH 4 dt
P m1.CsCOD . X 1 ; (Contois equation) K SX1 . X 1 CsCOD P m 2 .CVFA . X 2 K SX 1 . X 2 CVFA
1 YX 1 / sCOD
YVFA / X 1 .
.
; (Haldane equation)
C KI
dX1 dt
dX 2 dt
(3)
2 VFA
dX 1 dX 1 . 2 dt YX 2 / VFA dt
YCH 4 / X 2 .
(2)
(4)
(5)
(6)
2. Experimental Fresh leachate was obtained from Piyungan Sanitary Landfill, Yogyakarta, Indonesia. Starter, in form of active digester effluent, was supplied by the cow-manure -based biogas min i-p lant located at Gadjah Mada University’s PIAT (Pusat Inovasi Agroteknologi) at Berbah, Sleman. Immob ilization media was produced from Lampung natural zeo lite supported by b entonite as the adhesive agent. To produce the immob ilization media, the raw natural zeolite powder (undersize 100 mesh) was mixed with bentonite with the ratio of 1:1. Afterward, the zeo lite-bentonite mixture was molded to form Raschig ring with the size of 1 cm inside diameter, 5 mm th ickness, and 2 cm long by the extruder. Lastly, the mo lded mixture was heated at 110o C for 12 hours by using furnace Thermolyne Tube Heater F21100. Anaerobic digestion was operated in batch system using a vertical-cylinder-formed digester made of acrylic equipped with vertical tube gasometer (Fig. 1). The certain volu me o f leachate was added to the digester as substrate and so was the inoculu m. The total volu me of mixture was consecutively 2.4 L, 2.25 L, and 2.0 L fo llo wed by the zeolite addition as many as 110 g, 225 g, and 450 g to form the ratio of 150, 240, and 350 g zeolite/g sCOD. Anaerobic d igestion without zeolite as the immob ilization media was set as control (0 g zeo lite/g sCOD). The analysis of sCOD, VFA, and ammonia during the experiment wou ld follow the standard procedure by APHA [9]. The sCOD analysis was conducted as the closed reflu x colorimet ric method. The VFA analysis used the titrimetric method and the ammon ia measurement used ion selective electrode (ISE) measurement. The gas volume was measured using the gasometer meth od outlined by Walker [10] while the methane content was analys ed by using Gas Chromatography (GC).
Hanifrahmawan Sudibyo et al. / Energy Procedia 105 (2017) 256 – 262
2
1 4
Annotation: 1. Anaerobic batch bioreactor 2. Gas sampling port 3. Bubbler 4. Gasometer
3 Fig. 1. Experimental set up 3. Results and Discussion The comparison among various treatments in this experiment was analysed in terms of digestion rate (as dCsCOD /dt in Eq. 4), the VFA accumu lation rate (as dCVFA /dt in Eq. 5), and methane production rate (as dCCH4 /dt in Eq. 6). The rates can also be identified by v isual observation of the slopes at the graphs in Fig. 2. Based on the aforementioned analysis, it was observed that all of the zeolite variations and control showed similar t rends. The trends of sCOD digestion rate (the slopes in Fig. 2 (a)) were not exactly the same among the bioreactors. Bioreactor with no med ia, and with med ia of 150 g/g sCOD, 240 g/g sCOD, and 350 g/g sCOD showed the highest digestion rate in the different periods of day 21-35, day 14-28, day 7-21, and day 7-14 respectively. The percentages of total sCOD decrease, which were calculated based on the lowest sCOD value reached, were 53.33% (0), 60.91% (150), 60.34% (240), and 62.26% (350) consecutively for the increasing of the zeolite content inside the bioreactor. In the first seven days, the VFA concentration was decreasing, while on the period of day 7 to day 21, the VFA concentration tended to be stagnant (Fig. 2 (b)). It indicated that beginning on day 7, the total VFA produced by the conversion of sCOD was immediately consumed by methanogenic cells. It was confirmed by the fact that methane production rate (the slopes in Fig. 2(c)) also increased significantly fro m day 7 to day 21 until it reached the highest accumulative volu me of methane produced in the period of day 28-35. On the other hand, the rate of methane production was also better by adding immobili zation media to the bioreactor than the one without it. The rate of methane production was the highest at the 350 g/g sCOD. The trend of ammon ia concentration change (Fig. 2 (d)) varied among the bioreactors. The ammonia concentration change for bioreactor without zeolite and b ioreactor with 240 g/g sCOD zeolite was first increasing until day 14 before it was decreasing afterwards. They were totally d ifferent fro m b ioreactor with 350 g/g sCOD zeolite content which shown consistent decreasing trend and 150 g/g sCOD zeolite content which shown fluctuating decreasing trend. However, the increasing of ammonia concentration was reasonable because ammonia could also be produced through the degradation of nitrogen -based chemicals, such as protein and urea, into amino acid and then ammonia. If the ammon ia produced was not able to be adsorbed totally by the zeolite, there would be the accumu lation of ammonia then. This result was similar to the study by Schnurer and Jarvis [11] showing that higher ammonia concentration resulted in lower methane production. Therefore, the optimu m zeo lite content to be added to the bioreactor was 350 g/g sCOD because this number of zeo lite was capable of reducing the ammonia through adsorption, producing more methane, and reducing the sCOD up to 62%.
259
260
Hanifrahmawan Sudibyo et al. / Energy Procedia 105 (2017) 256 – 262
There were nine constants to be defined (Table 1). After the constants value defined, the correlation between the number of zeolite content and the constants value was verified. The calculated correlation coefficient (using Eq. 2) was then transformed into absolute value and was ready to be co mpared with the critical value o f Pearson correlat ion coefficient. This work studied four variations for the zeolite rat ios so that the degree of freedo m was two. Therefore, the crit ical value of Pearson correlat ion coefficient fo r this work was 0.950 (level of significance 0.05). The absolute value of the calculated correlation coefficient must be greater than the critical value to show there is a co rrelation. After being verified there is a correlation, the integer value of the correlation coefficient (either positive or negative) shows what the correlation was. It could be linearly correlated, inversely correlated, or not correlated.
(a)
(b)
(c) (d) Fig. 2. Anaerobic digestion experimental data of: (a) sCOD (b) VFA (c) CH 4 (d) NH 3 (◊: no media; □: 150 g zeolite/g sCOD; ': 240 g zeolite/g sCOD; x: 350 g zeolite/g sCOD) Table 1 Defined constants of Contois and Haldane equations generated from M ATLAB simulation Constants
No Media
150 g/g sCOD
240 g/g sCOD
350 g/g sCOD
Correlation Coefficient (r)
Indication
μ m1 μ m2 KSX1 KSX2 YX1/sCOD YX2/VFA YCH4/X2 YVFA/X1 KI
4.2652 2.3111 494.0023 52.5642 0.9076 1.9249 10.4698 0.1806 20.0577
8.3111 2.5587 495.8256 52.0442 0.8304 1.9264 18.2752 0.1319 21.1357
8.3511 2.8757 486.9423 51.2988 0.8018 1.5181 19.3727 0.1833 21.7501
13.9329 3.5595 499.9532 51.3011 0.7829 2.2079 33.1370 0.1833 24.1114
0.952 0.951 0.220 -0.948 -0.973 0.193 0.951 0.205 0.952
Correlated Correlated No correlation No correlation Correlated No correlation Correlated No correlation Correlated
SSE sCOD SSE CVFA SSE CCH4
411,590 78,732 12.07
17,978 17,427 129.42
208,420 68,476 10.57
232,750 28,177 280.21
-
-
Hanifrahmawan Sudibyo et al. / Energy Procedia 105 (2017) 256 – 262
The correlation of the zeo lite addition with the constants led to important interpretations regarding the future reactor design. Based on the calculation, there were five constants correlated with the zeolite addition (Table 1). Fro m those five constants, there were four constants linearly correlated with the zeolite ratios, while the other was inversely correlated. This statistical approach supported the theoretical analysis. For instance, the value of YVFA/X1 and YX2/VFA were not correlated with the zeolite addition. It represented the trend of the VFA concentration along the time during the experiment wh ich was almost the same for every nu mber of zeo lite. Both variables existed in the d ifferential equation of dCVFA/dt. Meanwhile, the correlat ion between μ m1 and μ m2 with the zeolite addition was in line with the fact that zeolite addit ion took role in reducing the soluble ammonia in the substrate. That is why μ m1 and μ m2 value increased, showing that the cell could grow better with the presence of the zeolite. Besides, KI value also increased which meant the zeolite increase could decrease the inhibition of VFA. As the consequence of a better growth of the methanogenic cell because of the zeolite, YCH4/VFA increased too. This constant controlled the number of CH4 produced in the bioreactor. The last correlated constant, YX1/sCOD , became lower when the zeolite increased. The YX1/sCOD correlation represented the percentage of sCOD consumed. Based on Eq. 5, the decreasing YX1/sCOD would result in the greater percentage of sCOD decrease. 4. Conclusions The imp rovements after immobilization med ia was added were the h igher sCOD consumption and methane production rate. Zeolite was capable of adsorbing ammonia produced during the leachate digestion. In the range of zeolite ratio studied in, the optimu m zeo lite rat io added was 350 g/g sCOD. There were also improvements of μ m1 , μ m2 , YX1/sCOD , YX2/VFA , KI, and YCH4/VFA with the addition of zeo lite. Kinetic constants resulted from mathemat ical modelling and statistical approaches were in good agreement with the theoretical concept. Acknowledgements The study was conducted under the CLEA N Project financially supported by USAID PEER -Science Research Grant (NAS Sub-Grant Award Letter Agreement Nu mber 2000004934 and Sponsor Grant Award Nu mber AID-OAA-A-11-00012). The authors also exp ressed the highest appreciation to the Office of Civ il Work and Energy/Mineral Resources of Yogyakarta and the bureau of Waste Treatment, Infrastructure, and Municipal Water Supply of the Governmen t of Yogyakarta Province. References [1] [2] [3] [4] [5] [6]
Knox K, Jones PH,” Complexation characteristics of sanitary landfill leachates,” Water Res 1979; 13, pp. 839-846. Silva AC, Dezotti M, Sant'Anna Jr GL,”Treatment and detoxification of a sanitary landfill leachate,” Chemosphere 2004; 55 (2), pp. 207-214. Parawira W. Anaerobic Treatment of Agricultural Residues and Wastewater Application of High-Rate Reactors, Doctoral Dissertation, Department of Biotechnology. Swedia: Lund University; 2004. Mshandete AM, Björnsson L, Kivais AK, Rubindamayugi MST , Mattiasson B, “ Performance of biofilm carriers in anaerobic digestion of sisal leaf waste leachate,”Electronic Journal of Biotechnology 2008; 11 (1), pp.1-9. Shuler M, Kargi F. Bioprocess Engineering Basic Concepts. 2 nd ed. New Jersey: Prentice Hall; 2002. Wirawan SK, Sudibyo H., Setiaji MF, Warmada IW, Wahyuni ET, “Development of natural zeolites adsorbent: chemical analysis and preliminary TPD adsorption study ,” Journal of Engineering Science and Technology 2015; Special Issue 4 on SOMCHE 2014 & RSCE 2014 Conference, pp. 87-95.
261
262
Hanifrahmawan Sudibyo et al. / Energy Procedia 105 (2017) 256 – 262
[7]
Montalvo S, Guerrero L, Borja R, Sánchez E, Milán Z, Cortés I, de la Rubia MA, “ Application of Natural Zeolites in Anaerobic Digestion Processes: A Review,” Appl Clay Sci 2012; 58, pp. 125–133. [8] Wresta A, Budhijanto W,” The Effect of the Addition of Active Digester Effluent for Start-up Accelerator in Anaerobic Digestion of Soybean Curd Industry Waste Water (Basic Research for Biogas Power Generation),” Journal of Mechatronics, Electrical Power, and Vehiculrar Technology 2012; 3 (2), pp. 81-86. [9] APHA, Standard Methods for the Examination of Water and Wastewater. 20th . New York: American Public Health Association; 2005. [10] Walker M, Zhang Y, Heaven S, Banks C, “ Potential Errors in the Quantitative Evaluation of Biogas Production in Anaerobic Digestion Processes,” Bioresource Technol. 2009; 100, pp. 6339-6346 [11] Schnürer A, Jarvis A, Microbiological Handbook for Biogas Plants. Sweden: Svenskt Gastekniskt Center AB; 2010.
Biography Hanifrahmawan Sudibyo is the young researcher at Centre for Energy Studies, Gadjah Mada University focusing on mathematical modelling, kinetic study, waste treatment, and biomass utilization. He is officially the new faculty member of Chemical Engineering Department, Gadjah Mada University in 2016.
