Dry anaerobic digestion of differently sorted organic municipal solid waste: a full-scale experience

D. Bolzonella*, P. Pavan**, S. Mace*** and F. Cecchi*

*Department of Science and Technology, University of Verona. Strada Le Grazie,15 37134 Verona, Italy **Department of Environmental Sciences, University of Venice, Dorsoduro 2137, 30123 Venice, Italy ` ***Department of Chemical Engineering, University of Barcelona, Mart i Franques 1, plta.6; 08028 Barcelona, Spain Abstract This paper presents a comparison of dry anaerobic digestion reactors fed with differently sorted municipal organic solid wastes. One reactor was fed with source sorted organic wastes and a second reactor was fed with mixed organic wastes consisting of grey wastes, mechanically selected municipal solid wastes and sludge. The two reactors utilised the same process (Valorga) and operational conditions at full scale. The results of the study emphasise the inuence of the kind of treated material on the process performances, especially in terms of biogas and methane production, thus, energy reclamation. The reactor treating the source sorted organic waste and the reactor treating the mixed organic wastes generated some 200 m3 and 60 m3 of biogas per ton of waste treated, respectively, while the specic methane production was some 0.40 and 0.13 m3CH4/kgTVS, respectively. The mass balance and the nal fate of the digested material from the two reactors were also clearly different. As for the costs, these were some 29 e per ton of treated waste (50% for personnel) and 53 e/ton for disposing of the rejected materials. Incomes were some 100 e/ton (on average) and an other 15 e/ton came from green certicates. The initial investment was 16 million Euros. Keywords Dry process; energy; grey waste; mechanically selected municipal organic solid waste; mesophilic; source sorted municipal organic solid waste

Water Science & Technology Vol 53 No 8 pp 2332 Q IWA Publishing 2006

Introduction

The need for processes in the eld of conservation of resources and energy reclamation has become more than clear in recent years, thus interest in anaerobic digestion processes for the production of energy has increased. In fact, the claimed hydrogen economy is now under construction (Turner, 2004) and still dependent on fossil fuels (Muradov and Veziroglu, 2005). Therefore, in the meanwhile, the exploitation of alternative renewable energy sources is of particular importance, especially within the European Union. In fact, it is a declared target of the EU of reaching the production of some 15 million of equivalent tons of petrol per year by means of renewable sources. Anaerobic digestion of organic wastes for biogas production, which is now a reliable technology as conrmed by the growth of full-scale applications in Europe in the last decade (De Baere, 2000; Mata-Alvarez et al., 2000; Mata-Alvarez, 2002), can be the driving force for reaching this result. At present, the global treatment capacity of anaerobic digestion processes is some 2.5 million ton/year of municipal organic wastes. A half of this capacity has been gained in the last three years. Some 1.2 million ton/year of treated waste originates from separate and source collection and the same amount originates from mechanical separation from unsorted municipal solid wastes, demonstrating that anaerobic digestion processes are suitable for both situations (De Baere, 2004). With specic reference to the strategy used for waste collection, it has to be noted that this can be considered the rst treatment step for the waste (Cecchi et al., 2002) as it determines different yields of
doi: 10.2166/wst.2006.232
23

the anaerobic digestion processes (Pavan et al., 2000; Saint-Joly et al., 2000; Bolzonella et al., 2003a,b; Mace et al., 2003; Bolzonella et al., 2005) and a different nal fate for the treated material, i.e., composting and land application rather than incineration or landlling. High yields, in terms of biogas production, and good compost production are generally associated with the treatment of separately collected (SC-) or source sorted (SS-) municipal organic waste, while the mechanically sorted organic fraction of municipal solid wastes (MS-OFMSW) gives poorer biogas production and a residual material which should be disposed of in landlls or incinerated (CITEC, 2004). With specic reference to the full-scale industrial application of the AD processes, at rst, because of the historical background, more reactors adopting the wet processes (, 10% dry solids in the reactor) were applied; since then, dry digestion (more than 25% dry solids in the feed) has prevailed because of the reduced volume of reactors and wastewater production. Most applied technologies for dry processes are Dranco, Valorga, Linde and Kompogas, all working in the range 30 40% of total solids in the reactor feeding. Figure 1 shows the different digester designs for these rms (from Lissens et al., 2001). The experience presented in this paper deals with the study of the yields and process performances of two full-scale reactors applying the mesophilic dry process (Valorga) in the Bassano treatment plant (north-east Italy) and treating two different substrates: one reactor treated the source sorted organic fraction of municipal solid wastes (SS-OFMSW) while the second reactor treated the residual organic fraction originating from the mechanical treatment of grey waste mixed with mechanically selected organic waste (MS-OFMSW) and sludge. Both the reactors were monitored for some one year of operation in steady-state conditions and a comparison between the yields of the two reactors and other full-scale applications was determined.
Experimental

D. Bolzonella et al.
24

The design capacity of Bassano plant was some 52,000 tons per year: 44,000 t/y of MSW to be treated in a mechanical selection line for the separation of the organic fraction to be digested and 8,000 t/y of selected organic municipal solid waste, both originating from the city. Moreover, the treatment of 3,000 t/y of sewage sludge was considered. The plant is constituted of a sorting line for the removal of inert material, three 2,200 m3 reactors and a dewatering section followed by a composting area for the complete

Figure 1 Different digester designs used in dry systems (A illustrates the Dranco design, B the Kompogas and Linde design, and C the Valorga design) (from Lissens et al., 2001)

stabilization of the digested organic material. Biogas is burnt to produce heat and power in co-generation units. The reactors work in the mesophilic range of temperature adopting the Valorga process. When the plant was still under construction the strategy of waste collection in the area was changed and the differentiation of waste streams was introduced. Because of the difference between the design and the actual situation, in these two years, the plant has been treating some 16,000 tons per year of source sorted organic fraction of municipal solid waste (SS-OFMSW) in one reactor and some 22,000 t/y of grey waste (the residual MSW after the separated collection of the organic waste), together with some 9,500 t/y of mechanically selected organic wastes (MS-MSW) originating from other treatment plants and 3,000 t/y of sludge in the other two reactors. The grey fraction still contains organic matter which contaminates other inert elements like plastics, making very difcult their possible recycling or disposal in landlls. If the grey fraction has to be disposed of in landlls or incinerated or used as packing material for soil restoration, then the organic matter should be stabilized prior to the nal disposal (Mace et al., 2003). The typical composition and the characteristics of the two main wastes treated at Bassano plant are reported in Tables 1 and 2. In particular, Table 1 reports the typical composition of the two main streams of wastes, while Table 2 shows the total solid and total volatile solid content of the two wastes. With reference to the data reported in Table 2, it has to be emphasized that the high content of volatile matter for the grey wastes is due to the presence of paper/cardboard and plastic material rather than to organic biodegradable material. The values in the tables are in the typical ranges reported for these types of materials (Cecchi et al., 2002). Because of their characteristics, the two wastes underwent different treatments prior to being fed to the anaerobic digesters. In particular, the SS-OFMSW is treated using a simpler sorting line: bags containing waste are broken and the organic material is sent to a rst riddle for the separation of inert material, then to a magnetic separator for the removal of metals and is then minced (nal size of some 10 mm) prior to being fed to the digester. The grey MSW is sent to a breaker for opening the bags and then to a rst riddle, then to a metal separator, to a cutter, to a secondary drum-screen and is then treated in a densimetric separator for a further removal of inert material prior to being fed to the digester. Metals are then disposed of and the other inert material is used as refuse derived fuels (RDF) in incineration plants. Despite its characteristics the grey waste can be still considered for the AD process: in fact, beside the organic material, which can be sorted from the inert material, the paper and cardboard also show a certain anaerobic biodegradability with biogas production (Saint-Joly et al., 2000). During the observation period, both the digesters were monitored through a complete set of analysis, considering mass balance parameters (TS, TVS, biogas production) and stability parameters (temperature, pH, alkalinity, VFA).
Table 1 Composition of the two treated wastes (wet weight)
Fraction, % Source sorted-OFMSW Grey MSW

D. Bolzonella et al.

Organic Paper and cardboard Wood Metals Glass and inert Plastic Textiles

70 83 7 12 1 6 0 1 1 4 4 9 1 3

10 16 34 50 2 4 1 9 2 14 24 34 3 8

25

Table 2 Characteristics of the two raw wastes (average in brackets)

Parameter SS-OFMSW Grey MSW

Total solid, % wet weight Total volatile solid, %TS

27 47 (33) 55 90 (77)

48 72 (60) 35 91 (70)

Results and discussion

D. Bolzonella et al.
26

The two reactors were fed with some 40 tons per day of SS-OFMSW and some 40 tons per day of mixed organic material consisting of grey wastes (50% on average), MS-OFMSW (35% on average), and sludge (15% on average). In both the reactors part of the efuent streams and some warm water were added in order to reach a total solid concentration of some 30% in the fed material and the requested temperature in the inuent material. The typical biogas productions compared to the feeding loadings for the two reactors are reported in Figures 2a and 2b. With specic reference to Figure 2a, where the biogas production is reported for the reactor treating the mixed wastes, it is important to note that the variability of biogas production has to be ascribed to the variability of the MS-OFMSW present in the feeding: in fact, the organic material from the grey wastes showed a constant rate of some 20 tons per day and a biogas production of some 1,600 2,000 m3/d, while when feeding the MS-OFMSW the biogas production varied, reaching peaks of some 4,0005,000 m3/d. The peaks of biogas production are related to the quality of the inuent material rather than to the amount of fed material. On the other

Figure 2 Load of material (tons per day) and biogas yields for the two reactors treating the grey material (2a) and the SS-OFMSW (2b)

hand, when considering Figure 2b, which shows the data for load and biogas yields in the reactor treating the sole SS-OFMSW, it is evident how this material can give yields as high as 10,000 m3/d when feeding some 50 tons per day for a couple of weeks. That is a specic production with peaks of some 200 m3/ton of wastes. Here, it is evident the relation between the peaks of loading and the peaks in biogas production. Table 3 reports the typical yields observed in steady-state conditions in the two reactors during a period of one year of operation as well as the operational parameters of the reactors. As for the loads and feed characteristics, in terms of total and volatile solids and for the operational parameters (temperature, HRT, OLR), the two reactors worked in the same conditions. Also the typical process parameters for process monitoring, pH, alkalinity determined at pH 4 and 6 and VFA concentration, showed similar values. Alkalinity, in particular, showed very high values compared to those reported for the anaerobic digestion of differently sorted organic solid wastes (Bolzonella et al., 2003a), while the concentration of the total VFAs were in the typical range for this kind of wastes when operating in steady-state conditions (Mace et al., 2003). However, with specic reference to the treatment of the SS-OFMSW, which shows a larger tendency to biodegradation (Bolzonella et al., 2005), the variations in VFAs concentration in the reactor can be very important. Figure 3 shows the typical variation of VFA concentrations related to different loadings of organic material in the reactor treating the SS-OFMSW. It is evident, here, that the highest concentrations are related to peaks in loading or after the weekend, when the feeding was stopped on Sunday. The re-start of the feeding determined disturbing conditions. In fact, when the VFAs concentration increased, the relative presence of acetic acid diminished and the propionic acid increased while in ordinary conditions acetic acid was constantly 90% of the total VFAs concentration. This evidence is in good agreement with previous studies reported in the literature (Ahring et al., 1995; Bolzonella et al., 2003a; Bouallagui et al., 2005).
Table 3 Operational conditions and yields observed in the two reactors
Reactor R402 R403

D. Bolzonella et al.

Feed characteristics Material Average feed, tons/day TS feed, % TVS feed, %TS Operational conditions OLR, kgTVS/m3 reactord (typical) HRT, d (typical) Digesters temperature, 8C Process parameters pH Alkalinity (@ pH 6), gCaCO3/l Alkalinity (@ pH 4), gCaCO3/l Volatile fatty acids, mgCOD/l (avg in brackets) TS reactor, % TVS reactor, %TS Yields GP, m3 biogas/t waste 3 GPR, m3 biogas/mreactord 3 SGP, m CH4/kgTVSfeed CH4, % VS removal, %

Grey MS-OFMSW sludge 40 38 67 38 50 70 38.6 8.0 11.0 18.2 100 2,500 (450) 28 55 60 1.1 0.14 55 35 40

SS-OFMSW 40 33 78 46 40 60 36.7 7.9 10.6 18.5 200 2,000 (500) 23 50 180 3.2 0.40 56 40 45
27

OLR: organic loading rate; HRT: hydraulic retention time; GP: biogas production; GPR: biogas production rate; SGP: specic biogas production

2000 VFA, mgCOD/l 1500 1000 500 0 29/02/2004

100 Load, ton/day

50

D. Bolzonella et al.
28

0 14/04/2004 29/05/2004 Days

VFA Load

13/07/2004

Figure 3 Effect of different loadings of SS-OFMSW on the VFA concentration in the reactor

With specic reference to yields reported in Table 3 it is evident how the different characteristics of the treated wastes gave completely different yields in terms of biogas production: in particular, even though operating at the same organic loading rate (in the range 38 and 46 kgTVS/m3 reactord, respectively) the biogas production per ton of treated waste was completely different, passing from 60 to 180 m3 biogas/t of waste using the source sorted organic waste. Consequently, the specic methane production per kg of fed volatile matter was nearly three times higher (0.40 vs 0.13 m3CH4/kgTVSfeed) in the reactor treating the waste originating from source separated collection. The values observed in this study can be compared with typical yields reported for other full-scale applications of dry processes treating different types of substrates (see Table 4). From data reported in Table 4 it can be seen that typical yields obtained at full scale for wastes originating from different collection strategies are, in any case, greater than 100 m3 biogas per ton of treated waste. On the other hand, in Bassano, where a blend of grey, MS-OFMSW and sludge is treated, the biogas production dropped down to 60 m3 biogas/t, while when treating the SS-OFMSW the observed yields are as high as 180 200 m3 biogas/t, a value greater than the ones observed in other plants. Further, the specic methane production is also highly affected by the kind of treated material. In Bassano, where only organic wastes from separate collection were fed to one reactor, the specic yields reached values of 0.4 m3CH4/kgTVS. On the other hand, when treating the mechanically selected organic fraction from MSW, like in Amiens, the yield values were in the range 0.20 0.25 m3CH4/kgTVS while when treating kitchen and garden wastes, like in Tilburg and Engelskirchen, the specic productions are in the range 0.17 0.35 m3CH4/kgTVS. In particular, with specic reference to these plants, the higher the presence of garden wastes the lower the biogas production (Saint-Joly et al., 2000). This value goes down to 0.13 m3CH4/kgTVS when treating grey wastes, as in one reactor in the Bassano plant. Thus, a general agreement with other full-scale experiences reported in literature about the dry AD process can be observed. When considering different collection strategies, not only the gas yields, but also the nal fate of the digested material is strictly affected. In particular, digested material undergoes an aerobic composting process, but when coming from reactors treating SS-OFMSW and kitchen wastes this can be conveniently used as compost for agricultural purposes, while material coming from a mechanical selection, because of the presence of plastics and other inert materials, can be then used as material for covering landlls or other similar purposes or burnt in incineration plants. As a consequence of the characteristics of the treated materials also the mass balances gave different responses. Figure 4 shows the mass balances determined for the two

Table 4 Operational conditions and yields observed in different dry processes (de Laclos et al., 1997; Saint-Joly et al., 2000; De Baere, 2000; Wellinger et al., 1993).
Place, year Design capacity ton/year Waste (type) TS % TVS, %TS HRT d OLR8 kgTVS/m3d Avg. GP m3/t waste SGP m3CH4/kgTVS

Bassano (Italy), 2003 Amiens (France), 1987 Bassum (Germany), 2003 Brecht (Belgium), 2000 Salzburg (Austria), 1993 Tilburg (Holland), 1994 Engelskirchen (Germany), 1997

52,000 85,000 60,000 50,000 20,000 52,000 35,000

SS-OFMSW Mixture MSW MSW Garden Kitchen and garden Kitchen and garden Kitchen and garden

33 38 60 57 40 31 37 55 35 45

78 67 65 51 55 70 32 65 55 75

40 60 50 70 17 25 na 15 na 20 55 25 30

46 38 7.5 11.5 na 15 18 na 7.5 8.6 8.4 10.0

180 60 145 147 103 135 92 126

0.40 0.13 0.20 0.25 na 0.26 na 0.17 0.32 0.22 0.35

V V V D D D V V

8calculated on the basis of the typical feed rate and characteristics of wastes; D: Dranco, V: Valorga; na: not available

D. Bolzonella et al.

29

D. Bolzonella et al.
Figure 4 Mass balances for the two reactors at Bassano plant. (a) Mass balance for the reactor treating SS-OFMSW; (b) mass balance for the reactor treating the mixture of grey, MS-OFMSW and sludge.

reactors operating at Bassano: in particular, Figure 4a shows the mass balance for the reactor treating the SS-OFMSW, while Figure 4b shows the typical mass balance for the reactor treating the mixed wastes (grey, MS-OFMSW and sludge). Mass balances were determined on the basis of one-year operation at steady-state conditions. It is evident, when considering Figure 4a, that the residual materials originating from the treatment of the SS-OFMSW can be conveniently treated and recycled. In particular, the treatment of 1 ton of SS-OFMSW can give some 270 kWh of electric energy and 178 kg of compost of good quality to be used for agricultural purposes, while produced wastewaters are some 450 kg and the reject materials to be disposed of in landlls are only 15 20% of treated wastes. On the other hand, when considering the treatment of the mixed material (Figure 4b), the energetic yield is clearly lower (some 77 kWh of electric energy for any ton of treated material) and the rejected materials from the sorting line are some 600 kg. The major part of those can be recovered as refuse derived fuel (RDF) and burnt in incineration systems. Then, also 200 kg of bio-stabilised material, sent to landlls, and 160 kg of wastewaters are produced. Energetics and economics. From the energetic standpoint, the specic energy input in the treatment line during the observation period was some 72 kWh per ton of treated waste, that is a consumption of some 425 MWh on a monthly basis. Part of this is recovered from energy self-production (some 485 MWh per month) while the other 62 MWh per month is acquired and 130 MWh/month is sold to the national net by means of the green certicates. In terms of energy efciency, the plant shows an index of 1.4 and 4.3 kWhproduced/kWhconsumed for the mixed waste and the SS-OFMSW, respectively. At this stage, the income from the energy selling is some 15 e per ton of treated waste. As for the other economic items, the initial investment cost for the plant was some 16 million Euros (public investment). Incomes deriving from treatment (fees and taxes) were some 115 e/ton, 79 e/ton and 58 e/ton, for grey, MS-OFMSW and SS-OFMSW, respectively. These gave an average price of some 100 e per ton of treated waste. With specic reference to managing costs, these were some 29 e per ton of treated waste, 50% due to personnel costs (14 e/ton). Costs for disposal in landlls of rejected materials were some 53 e/ton while capital cost for investment was some 13 e/ton. The global economic balance was positive.

30

Conclusions

According to the evidence of this study the following conclusions can be summarised: rst, the strategy of waste collection affects the characteristics of the organic wastes and determines the yields of the anaerobic reactors, not only in terms of biogas production (thus energy reclamation) but also in terms of nal disposal (agricultural elds or landll/incineration) of the efuent streams from the process. Secondly, under the same operational conditions, reactor yields can be very different, depending on the characteristics of the treated waste: the reactor treating the SS-OFMSW showed specic biogas production up to 200 m3 per ton of treated waste, with a specic methane production of some 0.4 m3 of methane CH4 per kg of volatile solid fed to the reactor, while the reactor treating the mixed material consisting of the grey fraction of MSW, the MS-OFMSW and sludge showed a biogas production of some 60 m3 per ton of treated waste with a specic methane production of some 0.13 m3 of CH4 per kg of volatile solid fed to the reactor. That is a factor 3 of difference in terms of biogas yields. Thirdly, the energetic yield was 1.4 and 4.3 kWhproduced/kWhconsumed for the mixed waste and the SS-OFMSW, respectively. And nally, as for the managing costs, these were some 29 e per ton of treated waste (50% for personnel) and 53 e/ton for disposing of the rejected materials. Incomes were some 100 e/ton (as an average) and an other 15 e/ton came from green certicates. The initial investment was 16 million Euros.

D. Bolzonella et al.

References
Ahring, B.K., Sandberg, M. and Angelidaki, I. (1995). Volatile fatty acids as indicators of process imbalance in anaerobic digestors. Appl. Microbiol. Biotechnol., 43, 559 565. Bolzonella, D., Battistoni, P., Mata-Alvarez, J. and Cecchi, F. (2003a). Anaerobic digestion of organic solid wastes: process behaviour in transient conditions. Wat. Sci. Technol., 48(4), 1 8. Bolzonella, D., Innocenti, L., Pavan, P., Traverso, P. and Cecchi, F. (2003b). Semi-dry thermophilic anaerobic digestion of the organic fraction of municipal solid waste: focusing on the start up phase. Bioresource Technol., 86(2), 123 129. Bolzonella, D., Pavan, P., Fatone, F. and Cecchi, F. (2005). Anaerobic fermentation of organic municipal solid wastes for the production of soluble organic compounds. Ind. Eng. Chem. Res., 44(10), 3412 3418. Bouallagui, H., Touhami, Y., Cheikh, R.B. and Hamdi, M. (2005). Bioreactor performance in anaerobic digestion of fruit and vegetable wastes. Process Biochem, 40(3 4), 989 995. Cecchi, F., Pavan, P., Battistoni, P., Bolzonella, D. and Innocenti, L. (2002) Characteristics of the organic fraction of municipal solid wastes in Europe for different sorting strategies and related performances of the anaerobic digestion process. VII Latin American Symposium on Anaerobic Digestion. MeridaYucatan, Mexico, 22 25 October 2002, pp. 439 446. CITEC (2004). Guidelines for the design, production and running of high technology plants for the disposal of urban waste. Geva Editions, Italy. De Baere, L. (2000). Anaerobic digestion of solid waste: state-of-the-art. Wat. Sci. Technol., 41(3), 283 290. De Baere, L. (2004). The role of anaerobic digestion in the treatment of MSW: state of the art. In preprints of IWA Conference Anaerobic Digestion 2004, AD 10th, Montreal, Canada, August 29 September 2, 2004, pp. 395 400. de Laclos, H.F., Desbois, S. and Saint-Joly, C. (1997). Anaerobic digestion of municipal solid organic waste: Valorga full-scale plant in Tilburg, the Netherlands. Wat. Sci. Technol., 36(6 7), 457 462. Lissens, G., Vandevivere, P., De Baere, L., Biey, E.M. and Verstraete, W. (2001). Solid waste digestors: process performance and practice for municipal solid waste digestion. Wat. Sci. Technol., 44(8), 91 102. Mace, S., Bolzonella, D., Cecchi, F. and Mata-Alvarez, J. (2003). Comparison of the biodegradability of the grey fraction of municipal solid waste of Barcelona in mesophilic and thermophilic conditions. Wat. Sci. Technol., 48(4), 21 28. Mata-Alvarez, J., Mace, S. and Llabres, P. (2000). Anaerobic digestion of organic solid wastes. An overview of research achievements and perspectives. Bioresource Technol., 74, 3 16. Mata-Alvarez, J. (ed.) (2002). Biomethanization of the Organic Fraction of the Municipal Solid Wastes. IWA Publishing, London, UK.

31

Muradov, N.Z. and Veziroglu, T.N. (2005). From hydrocarbon to hydrogen-carbon to hydrogen economy. International J. Hydrogen Energy, 30(3), 225 237. Pavan, P., Battistoni, P., Mata-Alvarez, J. and Cecchi, F. (2000). Performance of thermophilic semi-dry anaerobic digestion process changing the feed biodegradability. Wat. Sci. Technol., 41(3), 75 81. Saint-Joly, C., Desbois, S. and Lotti, J.P. (2000). Determinant impact of waste collection and composition on anaerobic digestion performance: industrial results. Wat. Sci. Technol., 41(3), 291 297. Turner, J.A. (2004). Sustainable hydrogen production. Science, 305, 972 974. Wellinger, A., Wyder, K. and Metzler, A.E. (1993). Kompogas a new system for the anaerobic treatment of source separated waste. Wat. Sci. Technol., 27(2), 153 158.

D. Bolzonella et al.
32