Fuel 85 (2006) 1535–1540 www.fuelfirst.com
Pyrolysis and gasification of pellets from sugar cane bagasse and wood Catharina Erlich a,*, Emilia Bjo¨ rnbom b,1, David Bolado c,2, Marian Giner d,3, Torsten Torsten H. Fransson Fransson a,4 a
b
Department Department of Energy Technology, Technology, Royal Institute of Technology, Technology, 10044 Stockholm, Stockholm, Sweden Department Department of Chemical Engineering and Technology, Royal Institute of Technology, Technology, 10044 Stockholm, Stockholm, Sweden c School of Industrial Engineering and Telecommunications, Cantabria University, 39005 Santander, Spain d School of Industrial Engineering, Polytechnic University of Valencia, 46022 Valencia, Spain
Received 22 April 2005; received in revised form 6 December 2005; accepted 7 December 2005 Available online 3 January 2006
Abstract
Wood pellets have become become a popular form of biomass biomass for power generation and residential residential heating due to easier handling both for transportation and for feeders in the treatment units, improved conversion and storage possibilities. The research on wood pellets as fuel has also been intensified during the past decade. However, other biomass sorts in pellet form, such as sugar cane bagasse, have not yet been extensively studied, especially not physical effects on the pellets during thermal treatment. Bagasse and wood pellets of different origin and sizes, shredded bagasse and wood chips have been studied in a thermogravimetric equipment to compare the effects of sort, origin, size and form of biomass during slow pyrolysis and steam gasification. Physical parameters such as decrease of volume and mass during treatment, as well as pyrolysis and gasification rates are of primary interest in the study. An important observation from the study is that for pellets the char density decreased during pyrolysis to a minimum around 450 C, but thereafter increased with continued heating. The wood chips behaved differently with a continuous char density decrease during pyrolysis. Another conclusion from the work is that the size of the pellet has larger impact on the shrinkage behaviour throughout the conversion than the raw material, which the pellet is made of. q 2006 Elsevier Ltd. All rights reserved. 8
Keywords: Pyrolysis; Char density; Steam gasification
1. Introduction
The energy policy during the past decade in Sweden with taxation of fossil fuels has resulted in a large bio fuel market for heat and power generation. Wood pellets are a commercialised product with an annual production of about 1 million ton in Sweden [1] [1],, and has become become very attract attractive ive for reside residenti ntial al heating in houses that are not connected to a district-heating network. Wood pellets can be obtained in common gasoline stat statio ions ns and and are are thus thus easi easily ly avai availa labl ble. e. The The cost cost of pell pellet et production is about 61 Euro/ton where the raw material stands
* Corresponding author. Tel.: C46 8 790 7468; fax: C46 8 20 4161.
erlich@energy. rgy.kth.s kth.see (C. Erlich), Erlich), emilia@ke
[email protected] t.kth.se .se E-mail addresses: addresses: erlich@ene (E. Bjo¨ rnbom),
[email protected] [email protected] ( D. Bolado), mariansita17@hotmail. mariansita17@hotmail. com (M. Giner),
[email protected] (T.H. Fransson). 1 Tel.: C46 8 790 8256. 2 Tel.: C34 94 220 1055. 3 Tel.: C34 96 387 7170. 4 Tel.: C46 8 790 7475; fax: C46 820 4161. 0016-2361/$ - see front matter doi:10.1016/j.fuel.2005.12.005
q
2006 Elsevier Ltd. All rights reserved.
for half of the cost [2] [2].. Wood pellets are are also used in large-scale heat power plants for electricity generation and district heating. With the increased interest for pellets, the research in the field has also also been been intens intensifie ified. d. Emiss Emission ionss from from wood wood pellet pelletss combustion has been studied [3,4] [3,4],, as well as pyrolysis and combustion behaviour [5–7] and physical effects on the wood pellet during pyrolysis pyrolysis [5,8] [5,8];; comparisons are also made with wood in chip form [8] [8].. The advantages with biomass pellets are multiple; pelletisation facilitates storage, transport, and feeding in the treatment units units.. The The lowe lowerr mo moist isture ure conten contentt and and the decr decreas eased ed hetero heterogen geneit eity y of the densifi densified ed bioma biomass ss also also contrib contribute ute to improve improved d conversi conversion on technol technologie ogies. s. The developm development ent in Sweden Sweden shows shows that that the pellet pelletisa isatio tion n techno technolog logy y can be economically profitable if done in large scale [9] [9].. Another biomass which is available in large quantities is sugar cane bagasse, which without densification is very bulky and inhomogeneous making it difficult to introduce in modern conversion technologies. Presently, most bagasse is used in inefficient combustion devices connected to a steam cycle with low steam parameters producing heat and power needed for the sugar process. Pelletisation of bagasse would make possible the
1536
C. Erlich et al. / Fuel 85 (2006) 1535–1540
Table 1 Composition of raw materials used in the experiments Material
Cuban Bagasse Brazilian Bagasse Softwoodc a b c d
Composition wt% m.a.f.a
wt% d.b.b
C
N
H
O (calc)
Ash
46.9 47.0 52.7
0.2 0.2 0.1
5.5 6.0 6.0
47.4 46.8 41.2
1.7 5.5 d !0.3
m.a.f.: moisture- and ash free basis. d.b.: dry basis. Average composition of softwood [15]. According to the pellet manufacturer.
use of high efficient energy conversion processes with low emission levels and significant increase in electricity production. Furthermore, pelletisation makes it possible to efficiently add ash-controlling components decreasing the risk for ash-melting in the conversion devices [10]. Compared to wood pellets, there is very little experience on the thermochemical behaviour of different bagasse pellets (origin and size). Large bagasse pellets have been tested in fluidised bed gasifiers by Olivarez et al. [11], Waldheim et al. [12] and Filippos et al. [13], but very few parametric conversion studies exists [14]. The objective with the present study is to look at those parameters for pellets that are of importance for application in a packed-bed gasifier, but some results from the study are as well applicable to other conversion technologies. If not pelletised, bagasse will not be applicable as a packed-bed material or in fluidised bed technology. Gasified biomass permits better control of combustion than a solid fuel as well as an increased combustion temperature. Additionally, with gasification technology, it is possible to introduce higher efficiency power cycles or IC-engines for electricity generation. The present study compares the conversion behaviour during slow pyrolysis in oxygen-free environment and steam gasification of pellets from softwood (two pellet sizes) and sugar cane bagasse (two sorts and pellet sizes), wood chips and shredded bagasse to give a qualitative base of comparison between the different materials and forms. Of large importance is how the pellets of different materials behave physically during pyrolysis and gasification (mass- and volume decrease) in comparison with chip and shred form. Of particular interest is also how the pellet size affects the reactivity in gasification. 2. Experimental 2.1. Raw materials
Sugar cane bagasse (from Cuba and Brazil) and stem wood (Swedish spruce) have been used in the experiments. The bagasse was represented in pellet form, with : 6 mm for Cuban bagasse and : 12 mm for the commercially available Brazilian bagasse, as well as in shredded form. Some experiments were performed with : 6 mm pellets of Brazilian bagasse, prepared in Sweden [14]. The wood was studied both in pellet and chip forms. Wood pellets with : 6 and : 8 mm have been selected for this study, because these sizes seem to be
the most suitable for practical application [14]. Tables 1 and 2 present the chemical composition and other relevant properties of the raw materials. The mass interval of the pellets is due to both differences in length and in density. 2.2. Equipment
Thermogravimetric equipment (TG) built for pyrolysis and gasification studies of large samples of bulky biomass have been used in the experiment (Fig. 1) [16,17]. As this TG is designed for large biomass samples, whole pellets can be used, and thus the effect of form and size of the pellets can be studied. In contrast, most commercial TG equipment take samples in the size of milligrams [13,18,19]; therefore, few studies on the effects of forms and sizes of larger biomass particles can be found in literature. The biomass samples are placed on a circular stainless steel net plate (sample holder), which is suspended to a balance measuring the mass. A thermocouple (type K) has been introduced on the holder to register the sample temperature during treatment. Electrical heaters are placed both above and under the net plate, enable to heat the sample up to 900 C. Nitrogen gas at atmospheric pressure passes downwards the 8
Table 2 Physical properties of the materials used in the experiments Material (and pellet diameter where applicable)
Physical properties Moisture (105 C) (wt%)
Density (each piece) (kg/m3)
Mass (each piece) (10K3 kg total basis)
Length (mm)
Bagasse Cuba, : 6 mm Large shred Bagasse, Cuba Bagasse Brazil, : 12 mm Bagasse, Brazil, : 6 mm Small shred Bagasse, Brazil Wood, : 8 mm Wood, : 6 mm Wood chips
6.6
700–1200
0.20–0.45
8.5–14
5.0
150–270
0.15–0.30
–
4.9
1000–1250
1.1–1.6
9.5–13
4.6
1050–1150
0.25–0.30
9–11.5
5.3
Bulk: 190
–
–
5.6 6.3 9.7
1100–1250 1100–1250 280–450
0.55–0.85 0.35–0.55 0.2–0.4
9.5–14 10–15 Width: 17–26, 25–40
8
C. Erlich et al. / Fuel 85 (2006) 1535–1540
1537
Fig. 1. Thermogravimetric equipment used for the experiments.
column from the balance to the plate, in which the biomass sample is placed, to assure oxygen-free environment and to evacuate the produced gases during treatment. Nitrogen flow of 2 L/min is used in the pyrolysis and an additional steam flow of
0.12 kg/h for the gasification experiments. The gases produced during the thermochemical treatment of the samples are evacuated from the net plate downwards, and cooled with water before exiting in a flue gas channel.
1538
C. Erlich et al. / Fuel 85 (2006) 1535–1540
Table 3 Shrinkage, density and yield of char from wood and bagasse after pyrolysis in nitrogen at a heating rate of 20 C/min 8
Biomass samples, form and size
Final pyrolysis temp. ( C)
Bagasse, Cuba (: 6 mm) Wood chips Wood (: 8 mm) Bagasse, Brazil (: 12 mm) Bagasse, Cuba (: 6 mm) Wood chips Wood (: 6 mm) Wood (: 8 mm) Bagasse, Brazil (: 12 mm)
450
79.7 G1.7
450 450 450
a b
d / d0 (%)
l / l0 (%)
Char yield, m / m0 (% m.a.f.)
Char density (kg/m3)
DY/DY0 (%)
71.9G2.3
25.7G1.3
494G70
59G2.7
70.1 aG1.4 76.5 G1.2 80.6 G0.4
84.5bG1.3 83.4G2.7 84.1G1.2
29.0G1.0 28.7G0.4 28.9G0.9
247G21 672G17 637G22
67G2.0 59G1.5 57G1.7
800
70.9 G1.7
63.8G1.6
18.3G0.7
524G65
62G4.2
800 800 800 800
70.3 aG8.8 68.6 G1.4 69.0 G0.8 75.5 G2.1
76.2bG0.6 71.8G0.9 71.4G0.7 79.0G1.9
21.5G1.6 21.3G0.2 21.5G0.1 23.0G1.1
230G43 694G39 716G29 649G44
63G8.4 64G2.6 64G1.6 58G2.9
8
Width perpendicular to the fibre length direction. Fibre length direction.
A temperature programmer set on a heating rate of 20 C/min is connected to the heaters of the sample plate (temperature regulator 2). Temperature regulator 1 controls the heating of the nitrogen gas and the superheating of the steam for gasification. The steam used in the gasification is produced in a separate steam generator at atmospheric pressure. 8
initial pellet length (l / l0) and char mass and initial pellet mass (m / m0 [% m.a.f.]) are used to illustrate the shrinkage behaviour. Similarly, the change of the density during the pyrolysis is shown using the ratio between char density and initial particle density (DY/DY0). For wood chips the variances for shrinkage in direction perpendicular of the wood fibres are much larger than this along the length of the fibres (see the results for 800 C). For pellets the variances for decrease in length and diameter are, however, very small, indicating that the pelletisation has favoured homogeneous shrinkage. For wood, the relative decrease in length and diameter is practically the same for : 6 and : 8 mm pellets. The char yield is also almost the same for the pellets and chips, even though there are significant differences in initial density and mass. Comparing all pellets, the larger (: 12 mm) Brazilian bagasse pellets showed less decrease in length and diameter during the pyrolysis. The results concerning char density at 450 and 800 C suggest that the shrinkage behaviour of wood pellets is rather 8
2.3. Procedure
The pellets used for the experiments were cut and polished to nearly perfect cylinders and thereafter weighed and measured for diameter and length. The wood chips were prepared to uniformly rectangular slabs. A batch of 10 pellets/chips was placed on the net plate of the TG equipment and suspended onto the balance via the thermocouple, to register mass decrease and sample temperature during treatment. Different final temperatures of the pyrolysis were chosen to study the temperature effect on various physical parameters of the residual char such as volume decrease, density change and char density. For the gasification experiments the samples were first pyrolysed up to 800 C, thereafter the gasification took place at isothermal conditions (800 C). The temperature of 800 C was chosen to represent an average value for the reduction temperature found in a gasifier. For pellets, the gasified char samples were collected after 50–60% conversion (based on ash-free mass), to study the shrinkage during gasification.
8
8
8
8
3. Results and discussions 3.1. Shrinkage and char density of wood and sugar cane bagasse in pyrolysis
The shrinkage, the yield and the density of char from pyrolysis of pellets from wood and bagasse are shown in Table 3 as average values with variances. The ratios between char diameter and initial pellet diameter (d / d 0), char length and
Fig. 2. Pyrolysis of pellets of bagasse and wood, shredded bagasse and of wood chips in nitrogen with a heating rate of 20 C/min. Br bag: Brazilian bagasse; Cu bag: Cuban bagasse; pell: pellets; w: wood. 8
1539
C. Erlich et al. / Fuel 85 (2006) 1535–1540
Table 4 Shrinkage, density and yield of char from bagasse pellets ( : 6 mm) after pyrolysis in nitrogen at a heating rate of 20 C/min, depending on initial pellet density 8
Initial pellet density (kg/m3)
Final pyrolysis temp. ( C)
d / d0 (%)
l / l0 (%)
Char yield, m / m0 (% m. a.f.)
Char density (kg/m3)
700–800 900–1000 750–870 960–1100
450 450 800 800
79.2 G1.7 80.0 G1.3 70.3 G0.9 71.6 G2.1
70.6G2.3 71.8G1.0 62.6G0.9 65.2G0.8
24.5G0.5 26.4G1.1 18.1G0.5 18.5G0.8
415G24 547G20 479G44 568G59
8
different compared to that of chips. During the initial course of pyrolysis, the shrinkage of pellets is counteracted by the intensive release of gas from the pyrolysing material in temperatures between 250/300 and 450 C (Fig. 2). Continued heating up to 800 C results in a very low mass loss rate but relatively larger volume decrease. Thus the densities of the char samples obtained at final pyrolysis temperature 800 C are higher than those obtained after 450 C. This was observed in different extent in pyrolysis of pellets, but not at all for wood chips, which decreased continuously its char density during heating to 800 C. The dependence of shrinkage on the initial density of Cuban bagasse pellets (: 6 mm) is shown in Table 4. The decrease in length and diameter is slightly dependent on the initial pellet density, with less decrease for the higher density pellets. The char yield is higher for the higher density pellets, both after pyrolysis up to 450 and 800 C. 8
bagasse, the char of : 12 mm pellets have significantly longer conversion time than the char of : 6 mm pellets. It is also observed that char from wood undergoes faster conversion during steam gasification than char from bagasse.
8
8
8
8
8
3.2. Removal of volatiles in pyrolysis of the biomass samples
The relative mass (m / m0 [% m.a.f.]) as function of temperature during pyrolysis of wood and bagasse pellets, shredded bagasse and wood chips is presented in Fig. 2. For the Brazilian bagasse, the initial mass loss in the pyrolysis occurs at a higher temperature, 300 C, for the larger pellets (: 12 mm) than for the shred form and the smaller pellets (: 6 mm), for which both the decomposition starts at around 250 C. The char yield is also dependent on the size; it is higher for the larger pellets (: 12 mm) compared to the smaller pellets (: 6 mm), which in turn have slightly higher char yield than shred at 450 C. The loss of material in pyrolysis of the wood samples shows similar temperature profile as that for bagasse however, the form of wood samples, the density and the size of the pellets and chips have not affected the pyrolysis temperature profile or the char yield. 8
3.4. Shrinkage of chars from pellets during gasification
The shrinkage of pellets in length and diameter after pyrolysis to final temperature 800 C and gasification at the same temperature with about 50–60% conversion of the initial char mass is presented in Table 5. The decrease in length and diameter is presented both in comparison to the initial pellet measures (d / d 0 and l / l0), and to the shrinkage obtained for the char pellets after pyrolysis to 800 C (d / d 0,char and l / l0,char ). The latter has been calculated using the average values from Table 3 for the respective material. The shrinkage during gasification of char from pellets is rather small compared to the decrease during pyrolysis. It is observed that the volume of chars from larger pellets (: 8 and : 12 mm) decreases more than this of the smaller pellets (: 6 mm), and the char : 12 mm pellets shrink most, while in pyrolysis the : 12 mm pellets had the lowest volume decrease. It seems thus that the smaller the diameter of the initial pellet, the more stable char structure after pyrolysis. 8
8
8
8
3.3. Steam gasification rate of chars obtained from wood chips, bagasse shred and pellets of wood and bagasse
The relative mass (m / m0,char [% m.a.f.]) during steam gasification at 800 C of char samples obtained in pyrolysis of pellets, chips, and bagasse shred at final temperature 800 C is presented as a function of time, Fig. 3. Origin, pellet size and material form have impacts on the gasification rate. For wood, char from chips is gasified slightly faster than char from pellets, and smaller char pellets (: 6 mm) have a slightly shorter conversion time than the larger (: 8 mm). For Brazilian 8
8
Fig. 3. Relative mass (m / m0,char ) as a function of time for steam gasification of char at 800 C; for the abbreviations, see Fig. 2. 8
1540
C. Erlich et al. / Fuel 85 (2006) 1535–1540
Table 5 Shrinkage of chars, obtained from pyrolysis up to 800 C, after steam gasification at the same temperature 8
Pellet material and diameter
d / d 0 (%)
l / l0 (%)
Char conversion (% m.a.f.)
d / d0 ,char (%)
l / l0,char (%)
Wood (: 6 mm) Wood (: 8 mm) Cuban bagasse (: 6 mm) Brazilian bagasse (: 6 mm) Brazilian bagasse (: 12 mm)
64.4 62.8 66.4 61.6 65.8
67.8 64.8 61.0 62.9 70.5
58 62 46 58 49
94 91 95 n.m. 87
94 91 96 n.m. 89
4. Conclusions
Pyrolysis and steam gasification of pellets from softwood (two pellet sizes) and sugar cane bagasse (two origins and pellet sizes), wood chips and shredded bagasse have been compared. The major conclusions from this study are as it follows: – In a set of pellets of the same size and biomass type, the shrinkage in diameter and length during pyrolysis takes place with very small variances among the pellets, suggesting homogeneous physical behaviour in a packedbed conversion technology. In untreated biomass the variances in direction perpendicular to the fibres length are much larger than those along the fibres length. – Higher density pellets of bagasse result in higher char yield and smaller shrinkage during pyrolysis. – In pyrolysis, bagasse pellets with larger diameter had less volume decrease and higher char yield compared to smaller pellets. – Pellets undergo different volume decrease during pyrolysis compared to chip form. Pellets have higher char density after pyrolysis up to 800 C than after 450 C, while the char density of biomass chips decrease continuously with increased temperature. – The pellet size has impact on gasification rate but very slightly on pyrolysis rate. The larger the pellet, the slower the gasification. – In gasification, bagasse is less reactive than wood. – The volume decrease of chars from pellets during gasification is very small, suggesting that a char structure stabilisation of the pellet takes place during pyrolysis. – The smaller the pellet, the smaller is the relative volume decrease of char during gasification. – The size of the pellet has larger impact on the shrinkage behaviour throughout the conversion than the raw material, which the pellet is made of. 8
8
Acknowledgements
The financial support of the Swedish International Development Agency (Sida) is greatly appreciated. The present work is part of Sida project SWE 2000-189B. Many thanks to CETER at ISPJAE in Havana, Cuba, for the supply of the Cuban bagasse. Thanks to TPS Thermal Processes, Sweden for the
¨ energi, supply of the Brazilian bagasse and to Glommers Miljo Sweden for the supply of the wood pellets. Technical assistance from Mikael Lundgren, Department of Chemical Engineering and Technology, KTH, Sweden, is also greatly appreciated. References [1] Olsson M, Kja¨llstrand J, Petersson G. Biomass Bioenergy 2003;24:51–7. [2] Hirsmark J. Densified biomass fuels in Sweden: country report for the EU/INDEBIF project. MSc thesis no. 38. Swedish Institute for Agricultural Sciences; 2002. [3] Kja¨llstrand J, Olsson M. Biomass Bioenergy 2004;27:557–61. [4] Olsson M, Kja¨llstrand J, Petersson G. J Anal Appl Pyrolysis 2003;67: 135–41. [5] Palchonok G, Leckner B, Tullin C, Martinsson L, Borodulya, A. Combustion characteristics of wood pellets. Pellets 2002 international conference, Stockholm, Sweden, September 2–6, 2002. p. 105–9. ¨ hman M, Hedman H, Neme´th S. Combustion of pelletised hydrolysis [6] O residues from ethanol production in residential pellet appliances (! 20 kW). ETC report no. 2002-2, Energy Technology Centre, Pitea˚, Sweden; 2002. [7] Andersson S, Salman H, Kjellstro¨m B. Utilisation of hydrolysis residues as gas turbine fuel. STEM report no. P11452-1, Swedish Energy Agency, Sweden; 2003. [8] Hansson KM. Principles of biomass pyrolysis with emphasis on the formation of the nitrogen-containing gases HNCO, HCN and NH 3. PhD Thesis. Chalmers University of Technology, Gothemburg, Sweden. ISBN 91-7291-349-5; 2003. [9] Thek G, Obernberger I. Biomass Bioenergy 2004;27:671–93. [10] Bentzen JD, Hindsgaul C, Henriksen U, So¨rensen LH. Straw gasification in a two-stage gasifier. In: Proceedings of the 12th European conference and technology exhibition on biomass for energy, industry and climate protection, Amsterdam, June 2002; 2002. p. 577–80. [11] Olivares Gomez E, Barbosa Cortez LA, Silva Lora E, Glauco Sanchez C, Bauen A. Energy Convers Manage 1999;40:205. [12] Waldheim L, Morris M, Regis Lima Verde Leal M. Biomass power generation: sugar cane bagasse and trash, progress in thermochemical biomass conversion, Tyrol, Austria, September 17–22, 2000. [13] De Filippis P, Borgianni C, Paolucci M, Pochetti F. Biomass Bioenergy 2004;27:247–52. ¨ hman M, Bjo¨rnbom E, Fransson T. Fuel 2005;84:569–75. [14] Erlich C, O [15] Nordin A. Biomass Bioenergy 1994;6:339. [16] Rensfeldt E, Blomkvist G, Ekstro¨m C, Engstro¨m S, Espen B-G, Liinanki L. Basic gasification studies for development of biomass medium-BTU gasification process. Energy from biomass and wastes, IGT, paper no. 27, 14–18 August, 1978. p. 466–94. [17] Zanzi R. Pyrolysis of biomass. PhD Thesis. Royal Institute of Technology, KTH, Stockholm, Sweden; 2001. [18] DeGroot WF, Shafizadeh F. Kinetics of wood gasification by carbon dioxide and steam. Fundamentals of thermochemical biomass conversion. Amsterdam: Elsevier; 1985. p. 275–92. ISBN 0-85334-306-3. [19] Klose W, Wo¨lki M. Fuel 2005;84:885–92.
