sign in sign up
<<
Home , Vegetable oil

ANNUAL TRANSACTIONS OF THE NORDIC RHEOLOGY SOCIETY, VOL. 22, 2014

The Effect of Waste Vegetable Addition on Pelletability and Physical Quality of Wood Pellets

Nevena Mišljenovi1, John Mosbye1, Reidar Barfod Schüller2, Odd-Ivar Lekang1 and Carlos Salas-Bringas1

1 Dep. Of Mathematical Sciences and , Norwegian of Life Sciences, P.O. Box 5003, N-1432 Ås, Norway 2 Dep. Of Chemistry, and Science, Norwegian University of Life Sciences, P.O. Box 5003, N-1432 Ås, Norway

ABSTRACT a global level. Alternative feedstock of The aim of this study is to produce particular interest are residues and side alternative wood pellets with addition of products from and food/feed waste vegetable oil and to examine how the processing2. Market integration of oil affects pelletability and pellet’s physical alternative and mixed pellets still meets properties. The oil addition was 2.2% and various constrains, but recent study showed 5.8% on dry solids. The results show that oil importance of this sector because it helps addition significantly increases energy sustainable rural development and add value content in wood pellets (p<0.05). Energy on locally available cheap raw materials3. content was increased from 19.69 MJ/kg Alternative pellets are typically (control) to 20.05 MJ/kg (2.2% of oil) and produced from sawdust, which can be 20.64 MJ/kg (5.8% of oil). Strength of considered as a carrier, and additional non- pellets and the pressure required to initiate woody biomass in smaller amounts4. The pellet discharge from the die were reduced amount of alternative depends by oil addition. on process and product quality limitations. Apart from other fibrous materials like INTRODUCTION straw, hay and grass, usage of starch and The EU is currently a leader in world lignin based materials is very common pellets market. Demands for pellets in because those materials help binding Europe have already exceeded production particles together5-7. Vegetable are capacities, thus significant amounts of mentioned in literature as an additive which pellets have been imported mainly from US. reduce die wall friction and reduce energy Moreover, European Commission predicts consumption for pelleting process due to that pelleted solid will play lubricating effect7, but there is no systematic important role in meeting study about the influence of oil on targets in Europe1. All this indicate further pelletability and physical quality of pellets. growth in pellet production and necessity In this study waste vegetable oil (WVO) for utilization of new, cheap, alternative raw was selected as a raw material for materials especially because woody biomass production of alternative biofuel pellets. is almost exclusively used as a feedstock for WVO is discarded by , food pellets production which makes it scarce on manufacturers (potato chips, breaded fish

211 sticks, doughnuts, etc.) and other facilities (Brabender, Duisburg, Germany) with (schools, , households, etc.) screen opening size of 1.5 mm. because it cannot be further used in human Adopting literature recommendation for or animal nutrition. Inappropriate disposal pelleting process11, the wood powder was of WVO can be environmentally hazardous; moisturized to 10 ± 1% moisture content by therefore its further utilization is preferable. spraying water in an intensive laboratory WVO has a considerable potential as a mixer (Diosna P1/6, Germany). The final component of pelleted biofuels, because it is moisture content in spruce powder was non-fossil oil, has high calorific value and 10.81±0.34% (Table 2.). This material does not require any pretreatment. Oil served as control sample and as a base raw addition can improve wood fuel properties, material for the mixtures with oil. The but densification of wood with added oil can moisturized powder was stored in vacuum be more challenging. Application of oil can packed plastic bags at ambient conditions reduce formation during pelleting and until further usage (1 – 2 weeks). later during pellet handling. Availability of The WVO was obtained from WVO is not a limitation since large amounts McDonald’s in Vestby, Norway. are available from restaurants, food Before spaying, the WVO was filtrated processing and households8. Price trough filter paper. Basic physicochemical of WVO oil is 2.5 – 3 times lower than properties of the WVO are presented in virgin vegetable oil9. Costs of handling Table 1. WVO before usage include just transportation, storage, and costs of Table 1. Properties of waste vegetable oil. mechanical cleaning (filtration), which is Property Value usually done at production place10. HHV, MJ/kg 37.60±0.07 The main purpose of this article is to Density*, kg/m3 910.7 produce alternative wood pellets with Dynamic viscosity12, 0.117 (at 20°C) addition of WVO and to examine how the Pa s 0.026 (at 60°C) oil addition affects pelletability and pellet’s 0.007 (at 120°C) physical properties. Oil has been added in Ash content9, % 0.006 two different amounts (2.2 and 5.8% on dry *Measured at room temperature solids). Mixed and control pellets (no oil addition) were produced by single pellet The WVO at two different levels (L1 and press method at 120°C under four different L2) was added to the powder by spraying in compacting pressures (75, 150, 275 and 300 a high shear mixer having three impellers MPa). and a tulip-form chopper (Diosna P1/6, Germany). Intensive mixing (mixer 250 MATERIAL AND METHODS rpm; chopper 500 rpm), manipulating oil Preparation of the raw materials drop size in the spraying lance (Düsen- Spruce stumps (Picea abies) were Schlick GmbH, Germany, Model 970) and selected as raw material and cut from a local preheating of the WVO to 60ºC were used forest in Ås, Norway in December 2012. In to facilitate an even of the oil in order to preserve woods physicochemical sawdust. Final content of oil in sawdust was properties the stumps were stored in a calculated from difference in the higher freezer until further usage. Prior to the heating value (HHV) of control powder and pelleting the wood was defrosted, the bark powder with oil. Powders L1 and L2 was removed, stumps were cut in smaller contained 2.2% and 5.8% of WVO on dry pieces, pre-dried and grinded in a grinder solids, respectively. The materials were pelletized at least 48 hours after the mixing

212 in order to allow WVO to penetrate from required to initiate pellet discharge from the fiber surface into pores. die i.e. the pressure at incipient flow (Pmax) was recorded to determine differences in Characterization of the raw materials friction generated on ‘die – pellet’ contact Following analyses on raw materials area among the materials. were performed: The particle size distribution of spruce Testing the pellets sawdust was measured by Mastersizer 3000 The produced pellets were stored in optical unit combined with Aero S dry sealed plastic bags at ambient conditions dispersion unit (Malvern Instruments, U.K.). until further testing. The strength of pellets The bulk density was determined by was measured 48 hours after production. measuring the mass of a known volume of Before the strength measurement, material that has been loosely poured into a dimensions (length and diameter) and mass graduated cylinder. The density of oil was of each pellet were measured to calculate determined by measuring the mass of know pellet density. The strength measurement volume of oil in a graduate cylinder (±0.2 was performed by a diametric compression ml) on analytical balance (±0.0001 g) test using a 60mm diameter probe connected The water activity value (aw) was to Lloyd LR 5K texture analyzer (Lloyd measured by a Rotronic HygroLab C1 Instruments, U.K.). Maximum normal force (Switzerland) instrument. Average during the pellet breakage was recorded. temperature during the aw measurements Compression test was performed with speed was 20.16±0.10ºC. of 1 mm/min. The moisture content was determined by overnight drying in an oven (Termaks, Data analysis Norway) at 105°C. Experiment was designed around two The content of ash and volatile matters factors; oil content at three levels (C – was determined in a muffle furnace control; L1 – 2.2%; L2 – 5.8%) and (Nabertherm D-2804, Germany) according compressing pressure at five levels (0, 75, to EN 14775 and EN 15148 standard 150, 225 and 300MPa). Minimum 3 procedure, respectively. replicates were tested per each combination of the factors. Results were analyzed using The fixed carbon (%d.b) was calculated by difference between 100 and the sum of ANOVA – General Linear Model for the volatile matter and ash content. unbalanced data (p<0.05). Tukey’s tests were employed to determine which groups The HHV was calorimetrically determined (Parr 1341 Bomb differed (95% confidence interval). Calorimeter, Moline IL) following the EN 14918 standard procedure. RESULTS AND DISCUSSION The addition of WVO caused changes in Pellet production the spruce sawdust properties. Small Pellets were produced in a single pellet changes in the powders appearance were press following the same procedure as observed. The powders with oil were less described in Mišljenovi et al.13 Detailed dusty, slightly darker and less flowable description of the pelleting unit was during handling and filling of the die channel. The changes in powder flow are previousely presented by Salas–Bringas et 14 particularly important for potential al. The pellets were produced at 120 ºC industrial application of the oily raw under four different compacting pressures - materials. 75, 150, 225 and 300 MPa. The pressure

213

Table 2. Raw material properties. Parameter Control L1 L2 Bulk density, kg/m3 119.44±1.54a 119.35±1.95a 122.05±1.81a a a b aw value 0.517±0.009 0.540±0.007 0.461±0.018 Proximate analysis Moisture content, % 10.81±0.34 10.60* 10.24* a a a Ash, %d.b. 0.28±0.02 0.27±0.05 0.28±0.04 * Fixed carbon , %d.b. 12.14 8.19 7.62

Volatile matter, %d.b. 87.58±0.75 91.54±0.15 92.10±1.08 a b c HHV, MJ/kg d.b. 19.59±0.16 19.99±0.10 20.64±0.18 abc different letters within the same raw indicate significant differences according to post – hoc Tukey’s test (95% confidence interval); *calculated values of pellets. The most important change Good flowability of powders can be caused by WVO addition is reflected in achieved by careful of silo and increased energy content. Significant equipment or by modification of powder increasing was observed (p<0.05). Energy flow properties15. The most important content was increased from 19.69 MJ/kg properties of tested materials are presented (control) to 20.05 MJ/kg (L1) and 20.64 in Table 2. The bulk density varied between MJ/kg (L2). Compared with other method of 119.35 – 122.05 kg/m3. The oil addition did biomass energy enrichment like not cause significant changes in the bulk torrefaction, addition of 2.2% oil increased density of spruce sawdust (p<0.05). final energy content in powder similarly as o 13 Aw value describes availability of free torrefaction at 225 C . This way of energy water in different products (typically food enrichment is definitely easier and less and feed). Free water affects products energy demanding. In addition to the tabular microbiological, enzymatic and chemical data which describe used materials, the stability. The aw value of tested materials volume based particle size distribution of was significantly reduced only when the spruce powder is presented in Fig. 1. 90% WVO was added in higher amount (L2). of the particles had diameter lower than This can be explained by covering the water 1750 m and 10% had higher than 166 m. with oil layer and making it in that way not The mean volume diameter of particles D accessible for the measurement. The aw [4;3] was 819 m. values of tested powders were very low indicating stability of materials in respect of 6 their microbiological activity. Microbiological activity can be a reason of 5 material self-ignition, deterioration of 4 material and health problems. In this range of aw values (0.461 – 0.540) growth of 3 microorganisms is not possible16. 2

The proximate analysis of powder % Density, Volume showed expected changes such as increasing 1 of the volatiles content with WVO addition 0 and no significant change of the ash in 10 20 30 50 70 100 200 300 500 1000 2000 5000 powder. Tested powders had low ash Size classes, m content and according to EU legislation (EN Figure 1. Particle size distribution of 14961-1) can be categorized in the A group spruce sawdust.

214 Pellets produced in single pellet press filling the fiber lumen and inter-particle have perfect cylindrical shape. In this space with incompressible and experiment the control pellets were well- because of poorer binding among the formed, with shiny surface, while the L1 and particles. L2 pellets had more porous structure and could easily disintegrate, indicating poor Table 3. Results of statistical analysis. binding between the particles. Physical Parameter df SS MS F p characterization of the pellets was Strength performed by measuring their density and Material 2 888.03 444.0 137.7 0.00 strength. Statistically, the type of material Pressure 3 894.14 298.0 90.40 0.00 did not have significant influence on pellet Interact. 6 182.91 30.48 9.25 0.00 Error 41 135.17 3.3 density while the compacting pressure and their interaction had (Table 3.). However, Density from Fig. 2 is possible to see changes in Material 2 667 334 1.75 0.18 compressibility of the oiled materials Pressure 3 62035 20678 108 0.00 compared to control. The densities of Interact. 6 29319 4886 25.56 0.00 control pellets significantly increase with Error 55 10513 191 increasing compacting pressure from 75 to DP 300 MPa, which was not the case for L1 and Material 2 0.1835 0.092 52.49 0.00 L2 pellets. The densities of control pellets Pressure 3 0.0184 0.006 3.51 0.02 changed from 1089 kg/m3 (75 MPa) to 1243 Interact. 6 0.0044 0.001 0.42 0.86 3 Error 55 0.0961 0.002 kg/m (300 MPa), while the densities of L1 and L pellets were practically pressure 2 independent (L : 1142 1196 kg/m3; L : 1 2 The strength of the pellets was measured 1164 1191 kg/m3). Additionally, L and 1 by a diametric compression test following L lines overlap indicating that the amount 2 the same procedure as described in of oil between 2.2 to 5.8 % did not have Mišljenovi et al.13 The peak force was influence on pellet density. Control pellets recorded and the results are expressed as the had higher density than L and L pellets 1 2 maximum force per length of pellet except for those produced at 75 MPa. (N/mm). Results of the statistical analysis

are presented in Table 3. Both factors and 1400 their interaction had statistically significant

1200 impact on pellet strength. Also, statistically there were no differences in strength 1000 3 between L1 and L2 pellets produced at the 800 same pressure, but both differed compared Control 600 to control (according to Tukey’s test). The L1

Density, kg/m L 400 2 strength of pellets significantly decreased with addition of WVO indicating poorer 200 binding among particles (Fig. 3). Binding 0 mechanisms during pelleting can be divided 0 30 60 90 120 150 180 210 240 270 300 Compacting pressure, MPa in five groups: solid , adhesion and Figure 2. Compressibility of spruce sawdust cohesion forces, surface tension and with different oil levels. capillary pressure, attraction forces between 17 solids and interlocking bonds . The oil The results show that WVO made the layer on the sawdust particles led to the material less compressible, probably due to surface deactivation, which resulted in

215 deactivation of all the mechanisms that exceeded the value of the pressure used to include contact sites and attraction forces produce pellets. Absolute value of this between surfaces. Similar effect was parameter does not indicate the extent of the observed by Nielsen et al.18 in sawdust with power requirements, but it can show the higher extractive content which deposition difference between materials and rank how on particle surface deactivated their binding easy or difficult material will flow through ability. Only interlocking and capillary die. The material’s resistance to be moved attraction between the particles due to in a die hole is in large extent responsible entrapped oil gave the strength to pellets. for the power requirements in commercial Even the interlocking was poor because of pellet presses. the oil’s lubricating effect which caused Results showed that Pmax values were sliding between the particles. Capillary significantly reduced by oil addition attraction due to oil layer is lower compared (p<0.05). Moreover, higher oil content led to water since the force is directly to a lower Pmax. This is a consequence of the 19 proportional to surface tension . Surface lubricating effect of WVO. Pmax was slightly tension of oil is about half of water value higher when higher pressure was used for and decreases with increasing temperature20. pelleting (Fig. 4). Both factors had significant influence on Pmax, but their interaction did not (Table 3). 35 32.5 Control L1 30 L2 27.5 25 22.5 20

Strength, N/mm Strength, 17.5 15 12.5 10 50 75 100 125 150 175 200 225 250 275 300 Compacting pressure, MPa Figure 3. Strength of pellets produced under different compacting pressures. Figure 4. Discharging pressure (Pmax) for pellets produced at different compacting Pelleting in single pellet press is a pressures. method that simulates compaction that occurs in flat or ring die pellet press. This In this experiment the pelleting was method is widely used for testing material performed 48 hours after the oil spraying in pelleting properties and can be considered order to allow better oil incorporation in as first step in process and experimental fiber particles. In an industrial case this design, but it cannot be a replacement for could be done in a silo. During pelleting, it pilot or large scale experiments. Energy was possible to observe some oil retention consumption cannot be measured in single on the bottom rod and on the bottom part of pellet press but can be indirectly the pelleting machine. Also, pellets 21 estimated . In this paper the pressure produced at higher pressure had ‘oil required to initiate pellet discharge (Pmax) gradient’ which can be a result of pressing was used as an indicator of the friction out the WVO from fiber lumen and inter- generated in the die - pellet contact area. particle space (Fig. 5). This leakage During pellet discharge, Pmax never definitely is not desirable. In context of real

216 pellet press where material continuously flows through die ‘oil gradient’ probably 2. Pollex, A. and T. Zeng. Initiators would not appear. For a large scale Handbook MIXBIOPELLS, 2012 production the main concern is about the previously mentioned powder flowability 3. Pollex, A. and T. Zeng. Summary of the and the fact that temperature in the die is MixBioPells project results, generated by friction between material and MIXBIOPELLS, 2012 the wall, so if the friction is too low temperature will not be high enough to 4. Nunes, L.J.R., J.C.O. Matias, and J.P.S. ensure good pellet quality. This article Catalão, (2014), "Mixed biomass pellets for shows the effects of oil in wooden materials, thermal energy production: A review of but an investigation in a pilot or commercial combustion models", Applied Energy, pellet press is needed to find the optimum 127(0): p. 135-140. between processing and physical quality. 5. Ahn, B.J., H.-s. Chang, S.M. Lee, D.H. Choi, S.T. Cho, G.-s. Han, and I. Yang, (2014), "Effect of binders on the durability of wood pellets fabricated from Larix kaemferi C. and Liriodendron tulipifera L. sawdust", Renewable Energy, 62(0): p. 18- 23.

6. Berghel, J., S. Frodeson, K. Granström, R. Renström, M. Ståhl, D. Nordgren, and P. Tomani, (2013), "The effects of kraft lignin additives on wood fuel pellet quality, energy use and shelf life", Fuel Processing Technology, 112(0): p. 64-69. Figure 5. Example of the ‘oil gradient’ in L1 pellet produced at 150 MPa. 7. Tarasov, D., C. Shahi, and M. Leitch, (2013), "Effect of Additives on Wood Pellet CONCLUSION Physical and Thermal Characteristics: A Waste vegetable oil has been added in Review", ISRN , 2013: p. 6. two different amounts in spruce sawdust, which has been pelletized in a single pellet 8. Toop, G., S. Alberici, M. Spoettle, and press under four compacting pressures. H.v. Steen. Trends in the UCO market, Presented results led to the conclusion that Ecofys, 2013 oil addition significantly increases energy content in biofuel, make material less 9. Demirbas, A., (2009), " from compressible, reduce pellet strength and waste oil via base-catalytic and reduce friction on the pellet – die contact supercritical methanol ", area. Energy Conversion and , 50(4): p. 923-927. REFERENCES 1. Flach, B., S. Lieberz, K. Bendz, and B. 10. Cvengroš, J. and Z. Cvengrošová, Dahlbacka. EU Biofuels Annual 2013, (2004), "Used oils and and their Global Agricultural Information Network utilization in the production of methyl esters (GAIN), NL3034, 2013

217 of higher fatty acids", Biomass and 16. Figura, L.O. and A.A. Teixeira, (2007) Bioenergy, 27(2): p. 173-181. "Food Physics Physical Properties - Measurement and Applications": Springer - 11. Leaver, R.H. "Wood Pellet Fuel and the Verlag Berlin Heidelberg Residential Market". in The Ninth Biennial Bioenergy Conference - Bioenergy 2000 - 17. Pietsch, W., (2002) "Agglomeration Moving Technology into the Marketplace. Processes - Phenomena, , 2000. Bufalo, New York. Equipment". Weinheim: Wiley - VCH Verlag GmbH 12. Mišljenovi, N., R.B. Schüller, E.-O. Rukke, and C.S. Bringas, (2013), 18. Nielsen, N.P.K., L. Nørgaard, B.W. "Rheological Characterization of Liquid Strobel, and C. Felby, (2009), "Effect of Raw Materials for Solid Biofuel storage on extractives from particle surfaces Production", Annual Transaction of the of softwood and hardwood raw materials for Nordic Rheology Society, 21: p. 61 - 68. wood pellets", European Journal of Wood and Wood Products, 67(1): p. 19-26. 13. Mišljenovi, N., Q.-V. Bach, K.-Q. Tran, C. Salas-Bringas, and Ø. Skreiberg, 19. Naidich, Y.V., I.A. Lavrinenko, and (2014), "Torrefaction Influence on V.Y. Petrishchev, (1965), "Study on the Pelletability and Pellet Quality of capillary forces between solid Norwegian Forest Residues", Energy & particles with a liquid layer at the points of Fuels, 28(4): p. 2554-2561. contact", Soviet and Metal , 4(2): p. 129-133. 14. Salas-Bringas, C., T. Filbakk, G. Skjevrak, O.-I. Lekang, O. Høibø, and R.B. 20. Esteban, B., J.-R. Riba, G. Baquero, R. Schüller, (2010), " Assesment of a new Puig, and A. Rius, (2012), "Characterization laboratory die pelleting rig attached to a of the surface tension of vegetable oils to be texture analyzer to predict process-ability of used as fuel in diesel engines", Fuel, 102(0): wood pellets. Energy consumption and p. 231-238. pellet strength.", Annual Transactions of the Nordic Rheology Society, 18: p. 77 - 85. 21. Nielsen, N., D. Gardner, T. Poulsen, and C. Felby, (2009), "Importance of 15. Schulze, D., (2013), "Shear Testing of Temperature, Moisture Content, and Species Powdera for Process Optimization", Annual for the Conversion Process of Wood Transactions of the Nordic Rheology Residues Into Fuel Pellets", Wood and Fiber Society, 21: p. 99 - 106. Science, 41(4): p. 414-425.

218