PERMEABILITY OF BULK WOOD PELLETS WITH RESPECT TO AIRFLOW by FAHIMEH YAZDANPANAH B.Sc., Amirkabir University of Technology (Tehran Polytechnic) , 2006 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF Master of Applied Science in THE FACULTY OF GRADUATE STUDIES (Chemical and Biological Engineering) THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver) June 2009 © Fahimeh Yazdanpanah 2009 Abstract Data on the resistance of wood pellets to air flow are required for the design and control of ventilation and drying of bulk pellets in storage. In this study, pressure drops versus air flows were measured for several sizes of wood pellets, with diameter of 6 mm and lengths varying from 4 to 34 mm. Air flow rates ranging from 0.014 to 0.80 m were used in the experiment. The maximum pressure drop measured was 2550 Pa m1. Three predictive models - Shedd, Hukill-Ives, and Ergun equations that relate pressure drop to air flow in bulk granular materials were used to analyze the data. The Ergun equation was found to provide the best fit to the data. Aeration of bulk pellets in storage requires a low airflow. The airflow range used for low permeability tests was from 0.0002 to 0.0220 m 5* The corresponding measured pressure drop ranged from 0.18 to 8.30 Pa m1 for low permeability tests. Three models were investigated and compared for the low permeability data. The increase in moisture content over a wider range of airflows (0.0042 to 0.7 148 m srn’) showed a slight decrease in the resistance to airflow due to increased moisture content. Broken and fines are produced when pellets are handled. The resistance to air flow for wood pellets was measured in the presence of fine materials. Fines were defined as broken pellets passed a 4 mm sieve. The average geometric diameter of the fines was 0.75 mm. The pressure drop for pellets mixed with fines ranged from (2.0 to 191.2 Pa m’) and (7.9 to 1779.0 Pa mrn’) for 1% and 20% fines content (mass basis) respectively. Coefficients of Hukill and Ives’ equation for pellets were estimated as a function of percent fines content 11 Table of Contents Abstract.ii Table of contents iii List of tables v List of figures vi List of symbols vii Acknowledgments viii Dedication ix Co-authorship statement x CHAPTER 1: Introduction 1 1-1. Background 1 1-2. Literature review 3 1-2-1. Air flow versus pressure drop along packed beds 3 1-2-2. Effect of particle size and bulk density on pressure drop airflow data 7 1-2-3. Effect of moisture content on pressure drop 8 1-2-4. Effect of fine material contents on pressure drop 9 1-3. Ventilation system design 11 1-4. Objectives 12 1-5. References 13 CHAPTER 2: Air flow versus pressure drop for bulk wood pellets 16 2-1. Introduction 16 2-2. Materials and methods 21 2-2-1. Experimental apparatus 21 2-2-2. Sample preparation and measurements of physical properties 22 2-3. Results and discussion 23 2-4. Conclusions 25 2-5. References 32 CHAPTER 3: Low air flow permeability of wood pellets 34 3-1. Introduction 34 111 3-2. Materials and methods .36 3-2-1. Materials 36 3-2-2. Methods 37 3-3. Results and discussion 38 3-4. Conclusion 42 3-6. References 50 CHAPTER 4: Permeability of wood pellets in the presence of fines 53 4-1. Introduction 53 4-2. Materials and methods 54 4-2-1. Material 54 4-2-2. Experimental apparatus 56 4-2-3. Test procedure 57 4-3. Results and discussion 57 4-4. Conclusion 60 4-5. References 66 CHAPTER 5: Discussion, conclusions and recommendations for future work 68 5-1. Discussion and conclusions 68 5-2. Recommendations for future work 74 5-3. References 75 Appendix A: Experimental and predicted data 76 Appendix B: Experimental set up 80 iv List of Tables Table 1.1 Total emission factors in g/tonne wood pellets 3 Table 2.1 Physical properties of pellets made from spruce 29 Table 2.2 Air flow rate and pressure drop data for three size categories of pellets 29 Table 2.3 Estimated constants for the various predictive models 30 Table 2.4 Constants a1 and b1 in Shedd’s equation among various materials 31 Table 3.1 Flow meters and manometers model and measurement ranges 47 Table 3.2 Properties of wood pellets used in low permeability test 47 Table 3.3 Pellet properties of samples used in tests on moisture content 48 Table 3.4 Estimated constants for Shedd (Eqn. 1), Hukill-Ives (Eqn. 2), and Ergun (Eqn. 3) 48 Table 3.5 Constants a1 and b1 in Shedd’s equation among various materials 49 Table 4.1 Flow meters and manometers model and measurement ranges 61 Table 4.2 Properties of wood pellets used in low permeability test 61 Table 4.3 Results of fitting Hukill and Ives equation to data 62 Table 5.1 Results of fitting Shedd equation to Q - AP data for clean pellets 71 Table 5.2 Results of fitting Shedd equation to Q - tiP data for pellets with 1% fines 71 Table 5.3 Results of fitting Shedd equation to Q - tiP data for pellets with 5% fines 71 Table 5.4 Results of fitting Shedd equation to Q - tiP data for pellets with 10% fines 71 Table 5.5 Results of fitting Shedd equation to Q - tiP data for pellets with 15% fines 72 Table 5.6 Results of fitting Shedd equation to Q - tiP data for pellets with 20% fines 72 Table 5.7 Relationship between A values and fine materials (Constant B) 73 Table A. 1 Measured and predicted pressure drops relevant to Figure 2.3 76 Table A.2 Measured and predicted pressure drops relevant to Figure 3.5 78 V List of Figures Figure 1.1 Blower performance curve and system curve 11 Figure 2.1 Equipment for measuring the resistance of wood pellets to air flows 27 Figure 2.2 Size distribution for sample 1 (L>6.7mm) and sample 2 (4mm<L<6.7mm) 27 Figure 2.3 Percentage error in the prediction of pressure drop for mixed sizes of wood pellets as a function of air velocity using three models 28 Figure 2.4 Resistance of white wood pellets with different sizes to air flow 28 Figure 3.1 Equipment for measuring the resistance of wood pellets to air flows 44 Figure 3.2 Size Distribution for Samples 1, 2 and 3 in moisture content tests 44 Figure 3.3 Resistance of a mix of wood pellets to very low airflow (Flowrate from 0.0002 to 0.0220 m 45 Figure 3.4 Experimental pressure drop per unit depth of bed height as a function of air flow for three different moisture content (Particle size range was between 5.9 to 36.1mm) 45 Figure 3.5 Percentage error in the prediction of pressure drop as function of air velocity for 3 models: (1) Shedd-type model (2) Hukill and Ives model (3) Ergun-type model 46 Figure 4.1 Equipment for measuring the resistance of wood pellets to air flows 62 Figure 4.2 Typical particle size distribution of fine materials in bed of wood pellets 63 Figure 4.3 Bulk density of wood pellets as a function of fine content 63 Figure 4.4 Experimental pressure drop per unit depth of bed height as a function of air flow for four different fines contents 64 Figure 4.5 Predicted pressure drop per unit depth of bed height according to Hukill and Ives Equation as a function of air flow for two different fines contents 64 Figure 4.6 Constant k of equation 2 as a function of airflow 65 Figure 5.1 Constant A of Shedd equation as a function of fine content for the air flow range of 0.024-0.06 m s1 (B 0.82) 73 Figure A. 1 Absolute error using three models (Relevant to data in page 76) 77 Figure A.2 Absolute error using three models (Relevant to data in pages 78-79) 79 Figure B. 1 Photo of experimental equipment 80 vi List of Symbols Symbol Meaning (common units Pressure drop in bed (Pa) a1 Equation constant b Equation constant V Air flow rate (m3 m2) L Bed length (m) A,B Equation constant Permeability of porous media (m2 s’ P&1) Air viscosity (Pa s) G Granular permeability R,S Equation constant Porosity of bulk solid p Density (kg m3) Shape factor d Characteristic dimension (m) X Equation constant M Moisture content (% w.b.) c3,y Equation constant Q Flow rate (m3 s m2) a,b,c,d Equation constant mc Moisture content % (w.b.) fin Fine material content (%) f Fine material content (%) k1,k2 Equation constant C1 Equation constant D Equivalent diameter of bed particles (m) SCFM Standard cubic foot per minute S, Standard deviation of estimates N Number of data points D Effective particle diameter (m) a Specific surface area of a particle (m2 m3) °C Celsius Pb Bulk density of seed (kg m3), cm bulk density (kg m3) Single pellet density (kg m3) Pk Corn kernel density (kg m3) vii Acknowledgments My sincere gratitude and appreciation goes to my academic advisers Dr. Shahab Sokhansanj and Dr. Anthony Lau for their excellent guidance, support and patience over the course of this work. Without their insight both into my research and personal development, I don’t believe I could have ever achieved all that I have. Also I extend my sincere thanks to members of the Biomass and Bioenergy Research Group (BBRG) particularly Dr. Xiaotao Bi, Dr. C. Jim Lim and Mr. Staffan Melin for their encouragement and helpful discussions. I owe much gratitude to several staff members at Chemical and Biological Engineering Department at the University of British Columbia (David Roberts, Doug Yuen, Horace Lam, Richard Ryoo, Gordon Cheng and Helsa Leong) for their administrative and technical assistance. The financial and in kind support from Natural Science and Engineering Research Council of Canada, Wood Pellet Association of Canada, Ministry of Forest and Range, Natural Resources Canada, and the Oak Ridge National Laboratory (Tennessee) is also acknowledged. Thanks are also extended to Princeton Co-generation Corporation and Fibreco Export Inc. for donating wood pellets for this study. Many thanks are expressed to my mom, Nahid and my brothers, Fardad, Farzad and Fahim for their continuous support and encouragement. Finally my special and deepest thanks and appreciation is to my husband, Hooman for his understanding, patience and continuous support and assistance during the course of this work. vi” xWq1{pu1p1ZJ1J‘p1pJ1J‘SJOqoJg‘JAT pu1 U1UTOOH‘pu1qsn‘JAI ‘pqfl‘WOJAI‘J’\I :opoipo Co-Authorship Statement Three chapters of the thesis have been submitted or will be submitted for publication in refereed journals. The co-authors include S. Sokhansanj, A. Lau, C.J. Lim, X. Bi and S. Melin. My contributions include: Planning the experiments Conducting all laboratory work and developing the predictive models; Performing experimental data analyses; Being the principal author of the manuscripts. x CHAPTER 1 INTRODUCTION 1-1. Background Wood Pellets are a form of compacted biomass to increase the density of the fuel and thus making it more economical to transport over longer distances. Most pellets are made from sawdust and ground wood chips, which are waste materials from trees used to make furniture, lumber, and other products. Resins and binders (lignin) occurring naturally in the plant material hold the sawdust together to make a pellet, so a pellet does not contain additives. Other materials, like straw, corn, nut hulls and similar can also be used to produce pellets, but those are less common. Pellets are cylindrical with a wide ranging diameter from a few mm to almost 25 mm. Pellets of interest in this thesis are 5-6 mm in diameter and 15- 25 mm in length. One of the main advantages of wood pellets compared to other commonly used biomass fuels (wood chips and logs) is their convenience: bags of pellets stack compactly and store easily, their uniform and small shape allows them to flow making the automation of fuel handling easy. The small size of pellets also allows for precisely regulated fuel feed and in turn, combustion air can be regulated easily for optimum burn efficiency. High combustion efficiency is also due to the uniformly low moisture content of pellets (typically 7-8%, compared to 30-35% moisture content in wood chips), and this means a high heat output and a very low level of unwanted emissions. As a biomass fuel, pellets offer the advantages of sustainable energy supplies through renewable raw materials. In addition, pellets are a by-product, not a primary user, of these renewable materials, and using pellets also helps reduce the costs and problems of waste disposal (Velimir egon, www.woodpellets.ca) Measurement of emissions from wood pellet burners has shown that the amount of particulate matter in the exhaust air is often less than 10 mg m3, thus meeting most stringent environmental requirements. Pellet heating systems also are CO2 neutral, because the quantity of CO2 emitted during combustion is equal to the CO2 absorbed by the tree during its growth. With the high efficiency burners developed in recent years, other emissions such as NO and volatile organic compounds are very low, making this biofuel of the most non- polluting heating options available (SWBC official website). Using sawdust as fuel in comparison with natural gas can reduce the amount of CO2 from 421529 g/tonne to 256329 g/tonne. It also reduces the amount of other pollutants like CO, NON, N20, VOC, PM, SO, and aldehyde to a great extent (Magelli et al. 2007). A comparison between the total emission factors of wood pellets versus natural gas is reported by Megalli et al. (2007) as shown in Table 1.1. 2 Table 1.1 Total emission factors in g/tonne wood pellets Production with natural gas asPollutant Production with sawdust as fuel fuel CO2 256329 421529 CO 836 854 CH4 105 1024 N20 13.7 16.6 NO 6065 6097 VOC 1510 1725 PM 907 916 SO 3780 3862 aldehyde 1.73 1.73 NH3 6.00 9.81 Because of their economics and environmental benefits, wood pellets are used increasingly as a substitute for fossil fuels. As solid fuels, pellets are handled frequently and kept in confined stores from the time of their production to the time of their use. Spontaneous combustion of long-term storage of wood pellets is a long-standing practical problem. To reduce potentials for spontaneous combustion, it is recommended to prevent the development of temperature and moisture gradients through well managed ventilation. 1-2. Literature review 1-2-1. Air flow pressure drop along packed beds For the past fifty years researchers have developed semi-empirical and empirical equations to describe pressure drop versus airflow for a specific product. ASAE D272.3 (ASABE 2007) provides airflow versus pressure drop for 40 granular products. These include data for small seeds such as clover to large particles such as alfalfa pellets, potato, and bell peppers, and ear corn. The reported data covers airfiows from 0.005 to almost 1.0 m 3 s1. (Note: this thesis uses the units for airflow either as m s1 or as m3 s m2). For some products such as wheat and corn the reported data covers airflow ranges as low as 0.0002 m s1. The plotted data are all in a log-log scale. ASABE (2007) gives the following semi- empirical equation to represent the airflow versus pressure drop data: LXP aV 1LThn(1+bV) () which is in the form of Hukill and Ives (1955) equation, AP is pressure drop (Pa), L is bed depth (m), V is airflow rate (m s’) and a, b are constants. ASAE D272.3 lists values for a and b for a specified range of airflow. The simplest and most widely used equation for expressing the relationship between the pressure drop and airflow is Shedd (1953) equation: V= A(r)B (2) where A and B are constants. Darcy’s law defines permeability as a constant to equate velocity with static pressure: LW V (3) where ..t is the permeability of porous media. (m2 s’ P&’) According to Darcy’s law the pressure drop varies linearly with the velocity. The permeability coefficient is an indication of media porosity. The permeability coefficient is a function of the pores structure and the fluid viscosity (Greenkorn, 1983). The greater the resistance coefficient is, the greater the pressure drop is created in the porous media. A higher value of permeability means lower pressure drop for the porous media. Segerlind (1982) defines a granular permeability G where airflow and pressure drop are non linear, 4 V tXP a” — _____ — Al B—1 — (tiP — (4) Equation (4) is used to develop flow streams within a bulk material. Hunter (1983) suggested using the following equation to describe the airflow-pressure drop relationship: P=RV+SV2 (5) where R and S are product-dependent constants. Compared to Hukill and Ives equation (1), this equation can be easily applied to non-uniform airflow distribution in stored grain. Hunter (1983) fitted equations (2) and (5) to airflow resistance data from the literature and mentioned that R and S values are related to a and b by the following relationships: R = 1.12 () and S = O.346b Hunter (1983) gives values for a and b for 28 products. Except the above mentioned empirical equations, there are some semi-theoretical and theoretical equations to be used in this area. Garret and Brooker (1965) used the method of Burke and Plummer (1928) for predicting the pressure drop in a bed of agricultural materials which was based on the summation of the drag forces acting on individual particles. Their predicted pressure drop didn’t agree with the measured values by Shedd (1953). Ergun (1952) developed the following fundamental equation to calculate pressure drop in a packed bed: (1—)2iV 1—e 3Ø2 .+1.75Ø3dpV (6) where e is porosity of bulk solids (fraction), V is air velocity (m s’), t is air viscosity (Pa.s), p is air density (kg m3), 4 is a shape factor and d (m) is a characteristics dimension of the 5 particle. Woodcock and Mason (1987) describe the development of equation (6) from fundamental fluid mechanics. On the right hand side of equation (6), the first term is a linear function of velocity. This term represents the viscous effect. The second term is a second order function of airflow. This term represents the kinetic effect. Equation (6) is useful for a packed bed of particles that have a well defined shape factor and characteristic dimension. These parameters are not well defined for most irregular shaped or non spherical particles. Ray et al. (2004) measured the airflow resistance of four types of commercially available pelleted feed varying in shape and dimensions. They used Equation (2) to express the relationship between the pressure drop and airflow rate. Sokhansanj et al. (1993) investigated the airflow resistance for alfalfa cubes, clean pellets, and fines mixed with pellets. Equation (2) fitted their data better than Equation (1). The constant A showed a strong dependence on the percentage of fines in the mix whereas B was not sensitive to fines and was assumed constant. Several factors affect the resistance of bulk granulated biomass to airflow including the air viscosity and density, porosity of the bulk material, orientation of the particles in bulk, size, and shape and particle surface condition. Montross et al. (2005), Chung et al. (2001), Giner et al. (1996), Sacilik et al. (2004), Kristensen et al. (2000) and Pagano et al. (2000) have reported data on resistance to airflow for corn, soybean, soft red and white wheat, baled hay, grain sorghum and rough rice, wood chips and oat seeds as affected by air velocity, bulk density, moisture, fine materials, type of wood. Ray et al. (2004) tested the effect of bin shape (circular, square or rectangular) on the 6 pressure drop in the beds of above mentioned materials and the results have indicated that the test apparatus cross section had no effect on the measured airflow resistance. 1-2-2. Effect of particle size and bulk density on pressure drop air flow data Calderwood (1973) considered the effect of increased bulk density (packing) and moisture content on the pressure drop. He found that the data from his tests were related to the data of Shedd (1953) by a Shedd’s curve multiplier (SCM). The SCM was the ratio of the pressure drop through a grain mass to the pressure drop predicted by Shedd’s curve for the grain at the same airflow rate. With a fill density of 639 and 729 kg/rn3, the SCM was 1.18 and 2.72, respectively, for rough rice at a moisture content of 15.2%. Bern and Charity (1975) studied the effect of the grain bulk density on the airflow resistance of corn. An increase in the bulk density resulted in an increase in the airflow resistance per unit depth of corn. They suggested the following empirical equations: = 1 + 2 ()2V + 3 () V2 (7) (1—a) (1—) Pk Where Pb is the corn bulk density (kg/rn3), Pk is the corn kernel density (kg/rn3),X1, X2 and X3 are constants. Molenda et al. (2005) also investigated the effect of bulk density in resistance to airflow for seeds. Montross et al. (2005) indicated the permeability of corn, soybeans, and soft red and white winter red as affected by bulk density. The resistance of corn, soybeans, soft white wheat, and soft red wheat were measured as a function of filling method, bulk density, and moisture content. Low bulk density corn and soybeans had the lowest resistance 7 coefficient, between 863 and 604 Pa s/rn2 (or permeability between 2.12 x and 3.03 x 10-8 m2, respectively). Increasing the bulk density by 7% decreased the permeability by approximately 40%. Wheat had the highest resistance, between 3112 and 1591 Pa s/rn2 (or a permeability between 5.88 x 1 and 1.15 x 10 m2, respectively) at the low bulk density. A 4.7% increase in bulk density resulted in a 41% greater resistance coefficient for white and red wheat. The effect of kernel orientation was not as significant as changes in the porosity. Darcy’s law was valid for predicting natural convection currents during non- aerated storage. 1-2-3. Effect of moisture content on pressure drop One of the major factors affecting the resistance to airflow is the moisture content of materials tested. Giner et al. (1996) did tests on the wheat with different moisture content. First they tested the three existing models (Shedd, Hukill and Ives and Ergun) and found Ergun type equation fitted the data best. To develop the equation which can express the effect of moisture content, they modified the Ergun type equation as follows: Lw = (ci. — ‘M)V + f3V2 (8) where AP is the pressure drop per meter depth of material in Pa m1, L is the bed length in m, M is the moisture content % (w.b), and V is the air velocity in m Equation (8) indicates that only the linear term of the Ergun-type equation is the function of moisture content. The second term that represents turbulent flow does not depend on moisture content. Equation 8 states that the higher the moisture content is the lower the pressure drop would be. 8 Chung et al. (2001) did the moisture content tests in beds of grain sorghum and rough rice and figured out the same trend in data. He proposed an equation including the effects of both moisture content and fine materials. = a(Q) + b(Q)2 + c(Q)(mc) + d(Q)(fm) (9) where AP is the static pressure drop per meter depth of grain (Palm), Q is the airflow rate in m3m2s’, mc is the grain moisture content (% w.b), fm is the fine material content (%) and a, b, c and d are regression factors. Static pressure drops for both grain sorghum and rough rice tended to increase with decreasing grain moisture content. They also found that among three variables (airflow rate, moisture content and fine materials), airflow had the most significant effect on the resistance to airflow. Other researchers including Li et al. (1994), Siebenmorgen et al. (1987), Sokhansanj et al. (1993), Pagano et a!. (2000) and Montross et al. (2005) have also verified a decrease in pressure drop with an increase in moisture content. 1-2-4. Effect of fine material contents on pressure drop The other major factor affecting the airflow resistance is the percentage of fine materials in the bed. Sokhansanj et al. (1993) has done tests in beds of alfalfa pellets containing fine materials in it. To be consistent with the published ASAE D272 (ASABE 2007), Sokhansanj et al. (1993) provided a correction to equation 2 for the fines content, ()ñnes = ()ciean [1 + (0.361 + 1.298v)fmj (10) where (AP/L)ciean is calculated from equation 2 for clean pellets; fm is the mass fraction of fines and broken in the mix; (AP/L)nes is the pressure drop of pellets mixed with fines. 9 Sokhansanj et a!. (1993) estimated a= 1.8 xl and b68 .72 for clean uniform size alfalfa pellets (6.2 mm diameter). Increasing the fine material content in the bed increased the resistance of bulk material to airflow. Chung et al. (2001) developed equation 9 for beds of sorghum and rough rice which includes both the effects of moisture content and fine material. In grain sorghum, static pressure drop tended to increase with increasing amounts of fine material. In rough rice, the static pressure drop decreased as fine material concentration increased, possibly because of larger particle size of fine material compared to grain sorghum. The effect of fine material was also investigated by Giner et a!. (1996) in beds of wheat. They showed the following relation between pressure drop and airflow with two constants k1 andk2. corrected =k1+2f (11) ()clean where f is the fine fraction in the bed. Sacilik (2004) also investigated equations for the effect of fine materials in beds of granular materials (poppy seeds). They also included the effects of moisture content in combination with fines to develop the following empirical equation: = C1Q2 + C2MQ + C3pbQ + C4fQ (12) where tiP is the pressure drop per unit depth in Palm; Q is the airflow rate in m s’; M is the moisture content in % d.b.; Pb is the bulk density of the seed in kgm3; F is the fines content in %; and Cl, C2, C3 and C4 are regression coefficients. 10 1-3. Ventilation systems design As mentioned earlier to prevent spontaneous combustion in storages, aeration and ventilation schemes are devised to keep the moisture content and temperature of pellets down and uniform throughout storage structure. Pressure drop data vs. airflow rate are required in order to design ventilation systems including size and power requirement for the fans and blowers. Fans or blowers are devices by which air is moved. The operation of a fan is characterized by the volume of air it moves and the total pressure associated with the volume of air (McLean, 1989). The output of the blower should be able to overcome the system pressure in order to deliver certain amount of air. Figure 1.. 1 shows a blower’s airflow-pressure curve and the system’s airflow pressure drop. The intersection of the two curves point A is the calculated operating conditions when the blower is operated in conjunction of the system to be ventilated. Pressure drop Figure 1.1 Blower performance curve and system curve. Interaction A is the operating condition. System Blower Airflow 11 1-4. Objectives The existing pressure drop data for feed pellets does not cover low airflow ranges prevalent in storing and aeration of wood pellets. Furthermore size distribution of wood pellets and the characteristics of fines are different than those for pellets feeds. The purpose of this study was (1) to determine the effects of airflow rates and particle size on pressure drop (2) to determine the permeability to low air flow of bulk wood pellets and to develop equations that relate differential static pressures in beds of bulk wood pellets subject to forced airflow (3) to investigate the effect of moisture content on the static pressure drop and (4) to develop experimental data on permeability of bulk wood pellets to air flow in the presence of fines fractions, and (5) to develop an equation to predict pressure drop as a function of airflow rate and fines content. Objective 1 is addressed in chapter 2 while objectives (2,3) and (4,5) are addressed in chapters 3 and 4 respectively. And finally the conclusions for the research and some recommendations for the future work are discussed in chapter 5. 12 1-5. References ASAE, 2007. ASAE D272.3 MAR1996 (R2007). Resistance to airflow of grains, seeds, other agricultural products and perforated metal sheets, ASABE Standards 4 (2007), pp. 627—633. Bern, C.J., Charity, L.F., 1975. Airflow resistance characteristics of corn as influenced by bulk density. Technical Report. 75-35 10. ASAE, St. Joseph, MI. Burke, S.P., Plummer, W.B., 1928. Gas flow through packed columns. Industrial Engrg. Chem. 20, 1196-1200. Calderwood, D.L., 1973. Resistance to airflow of rough, brown and mild rice. Trans. ASAE. 16, 525-527, 532. Chung, D.S., Maghirang, R.G., Kim, Y.S., Kim, M.S., 2001. Effects of moisture and fine material on static pressure drops in a bed of grain sorghum and rough rice. Transactions of the ASAE. 44(2), 33 1—336. Ergun, S., 1952. Fluid flow through packed columns. Chemical Engineering Progress. 48, 89-94. Garret, R.E., Brooker, D.B., 1965. Aerodynamic drag of farm grains. Trans. ASAE. 8, 49- 52. Giner, S.A., Denisienia, E., 1995. Pressure drop through wheat as affected by air velocity, moisture content and fines. J. Agric. Engng Res. 63, 73-86. Greenkorn, R.A., 1983. Flow phenomena in porous media. New York: Marcel Dekker. Hukill, W.V., Ives, N.C., 1955. Radial flow resistance of grain. Agricultural Engineering. 36(5), 332-342. 13 Hunter, A.J., 1983. Pressure differences across an aerated seed bulk for some common duct and store cross-sections. Journal of Agricultural Engineering Research. 28(5), 43 7-450. Kristensen, E.F., Kofman, P.D., 2000. Pressure resistance to air flow during ventilation of different types of wood fuel chip. Biomass and Bioenergy. 18, 175-180. Li, W., Sokhansanj, S., 1994, Resistance of bulk alfalfa seeds to airflow. Canadian Agricultural Engineering. 36(1), 25-28. McLean, K.A., 1989. Drying and storing combinable crops. Second Edition. Farming Press Books. Ipswich, UK. Megalli, F., Bi, X., 2007. A Streamlined Life Cycle Analysis of Canadian Wood Pellets. Presentation. Chemical and Biological Engineering Department. University of British Columbia. Molenda, A., Montross, A.D., McNeil!, S.G., Horabik, J., 2005. Airflow resistance of seeds at different bulk densities using Ergun’s equation. Transactions of the ASAE. 48(3), 1137-1145. Montross, M.D., McNeil!, S.G., 2005. Permeability of corn, soybeans, and soft red and white winter wheat as affected by bulk density. Applied Engineering in Agriculture. 21(3), 479-484. Pagano, A.M., Corzza, D.E., Nolasco, S.M., 2000. Airflow resistance of oat seeds: Effect of airflow direction, moisture content and foreign material. Drying Technology. 18(1&2), 457-468. Ray, S.J., Pordesimo, L.O., Wilhelm, L.R., 2004. Airflow Resistance of Some Pelleted Feed. Transactions of ASAE. 47(2), 513-519. Sacilik, K., 2004. Resistance of bulk poppy seeds to airflow. Biosystems Engineering. 89(4), 43 5-443. 14 ci ‘)IIOAMNjJ9-Jpuuiwdj 0U11PU1HspilosjpqL861‘Sf‘t105N‘‘>f.J‘)pOopOOp UIOOUjuoqMsMM//dUq :(uflJnsuo3ssouTsnflp1A&pIJoM,us)Juoog)sqgjoos I-LO‘(E)çEuuutuJ.rnMnouvuipiuwjoJpnoqijpssaidwoo puisijpd‘s3qn3JJJJnJOOU15IS)J66T“OO‘UIsJ‘4‘rj‘‘usuipios 619-919‘(6)E U!1UIUJJflflfl3UMOjJJTEOSpSUPU1JJO3urnS!sJEc6J‘fD‘PPqS tTT—8TT‘Q7)OE3VSVSULTP)IInqP’UO11U1O31’J AqPP°JJ°!‘JOM0UJ!VL86I‘NA‘J1puTf“f.1‘UJOUiUqig :tp‘qdsos LIO-l8°NJd1dVSVuoinbMOtJLit?WOUTj-UOUUiAJOg86I‘f]‘PUPS CHAPTER 2’ AIR FLOW VERSUS PRESSURE DROP FOR BULK WOOD PELLETS 2-1. Introduction Wood pellets are compacted ground wood materials, whereby natural resins and lignin in wood bind the loose particles together. The most common dimensions of wood pellets produced in Canada are 6 mm in diameter and up to 30 mm in length. The moisture content of pellets is in the range of 5-7% wet basis. While individual pellets have a particle density of about 1.2 g cm3, the bulk density of pellets ranges from 600 to 750 kg m3. Solid biofuels such as wood chips, bark, and grasses have low volumetric energy content due to their low bulk density. In regions especially those away from the Equator, the harvest season for biomass is short. Storage spaces are necessary to safely keep a large supply of feedstock for timely delivery to a biorefinery throughout the year. The biomass storage period may last for one day to several months. It is important to maintain the stored pellets under cool and dry conditions, and uniform temperature to prevent moisture migration and self heating. In wood pellets storage, slow oxidation seems to be the first step in a self-heating process, as the fuel is too dry to sustain any significant biological activity (Wadso, 2007). It was postulated that an initial temperature rise in storage is due to microbial activity, followed by exothermic oxidation when the temperature is greater than 60°C and finally thermal cracking takes place at temperatures higher than 200°C (Liodakis et a!., 2002). Hydrolysis, pyrolysis and oxidation A version of this chapter has been submitted for publication. Yazdanpanah, F., Sokhansanj, S., Lau, A., Lim, C.J., Bi, X., Melin, S. Airflow versus Pressure Drop for Bulk Wood Pellets. 16 reactions (Kubler et al., 1987) are among major factors contributing to heat production. Preventive methods are necessary to avoid storage fire, which is dangerous and cannot be easily controlled. Forced ventilation is a management tool to prevent excessive concentration of toxic gases such as CO, CO2 and CH4 and to prevent the possibility of spontaneous combustion. The design of a ventilating system requires pressure drop versus air velocity data for a bed of packed materials and its associated air delivery and exhaust system. These data guide the selection of an optimum size blower that would provide the necessary ventilation air flow rate (Giner, 1996). An overall pressure drop-air flow relationship is also useful for studying air flow patterns within the bulk of solid particles. These air flow streamlines can identify critical spots that remain unventilated and hence experience a moisture content and temperature increase that may lead to self-heating and spontaneous combustion. These streamlines also assist in designing and locating the ventilation ports to minimize hot spots. The objective of this research is to develop equations that relate differential static pressures in a column of bulk wood pellets subject to forced ventilation. Factors that affect the resistance of bulk pellets to airflow include air viscosity and density; porosity of the bulk material; size, shape and surface condition of particles; and orientation of the particles in bulk, These factors are highly interdependent; it is difficult to develop a unique model to relate the pressure drop to airflow as a function of these factors (Li and Sokhansanj, 1994). Sokhansanj et al. (1993) fitted Shedd’s equation (Shedd, 1953) to experimental data for alfalfa pellets and cubes, V =a1(iP)’’ (1) 17 where V is frontal air flow rate in m3 m2s1 and AP is pressure drop in Pa m* The constants a1 and b1 were estimated for one size of cubes and three sizes of pellets. They also estimated a1 and b1 for 5 to 25% fine contents in the 6.4 mm diameter pellet samples. The ASAE D272.3 standard (2007) provides data on air flow versus pressure drop for some forty materials, mostly agricultural crops. The data are presented graphically in a logarithmic scale. Hukill and Ives (Hukill and Ives, 1955) suggested the following equation that relates air flow to pressure drop for agricultural materials: AD i2a 2L — in (1 + b2V) where AP is the pressure drop (Pa), L is the depth of the bed (m), V is superficial or frontal air velocity (m s’). Constants a2 and b2 are tabulated in ASAE D272.3 (2007). The specified airflow range for alfalfa pellets is 0.0053 to 0.63 m s. Ray et al. (2004) investigated air flow versus pressure drop characteristics for feed pellets 4.0 mm, 6.7 mm, and 9.4 mm in diameter as well as cubes having a cross section area of 33.2 mnix34.9 mm. The air flow in their study ranged from 0.0093 to 0.236 m s. They investigated three storage container shapes - round, square and rectangular, and two filling methods - loose and tapped. Pressure drop versus air flow was found to be insensitive to the geometry of the container cross section. Tapping increased pressure drop by a maximum of 14% for the small (4 mm) pellets. Ray et al. (2004) presented Shedd’s constants A1 and B1 for the tested pellets and cubes. Kristensen and Kofman (2000) published pressure drop versus air flow data for cut willow chips, which were made from newly harvested material and the moisture content was about 18 55%. They classified the chips according to different size groups. Of particular interest to us are their data on chips with a nominal cutting length of 28 mm; 54% of the chips in this class had actual lengths of 7 -16 mm, and the majority of mass fractions were in the smaller size group (< 16 mm). Variations in particle size were determined to be the largest factor influencing air flow data. Table 2.4 lists the transformed values for a and b as derived from their study. Kristensen et al. (2003) published further more data on air flow and pressure drop for wood chips that included large chunks and for 8-mm size wood pellets. Air flow ranged from 0.02 to 0.3 m s. Nevertheless, Kristensen et al. (2003) did not have data on ranges in pellet lengths, bulk density, or moisture content of the pellets. Hunter (1983) made a comparison between the Hukill-Ives model (Eq. 2) and Ergun’s model (Ergun, 1952) and suggested the following modified form of Ergun’s equation, zW=RQ+SQ2 (3) where R=1 .12 a2/b and S0.346 b2. The symbols a2 and b2 are the constants in Eq. (2). Hunter (1983) reported values for R and S for 28 different grains. Gayathri and Jayas (2007) reviewed literature on the mathematical modeling of airflow distribution in bulk grain storage. They initiated their review from constituent Eqs. (1) and (2), followed by more complex equations. The Ergun equation is based on dividing the flow regime in a porous packed bed to laminar (viscous) flow and turbulent (dynamic) flow similar to Eq. (3): zPj=a3V+b2 (4) where a3 and b3 are constants associated with viscous flow and turbulent flow, 19 (1_e)2 a3 = 150 32 (5) (1— b3 = 1.75 ø3d p (6) where bed parameters are porosity e (fraction), bulk density p (kg m3); equivalent diameter of bed particles d (m); acceleration due to gravity, viscosity of the air fluid ji (Pa s) and the shape factor . Gu et al. (2000) instrumented an experimental grain bin with hot wire anemometers to measure the air flow pattern in the grain bin as a function of air entry geometries. They modeled the air stream lines using second order differential equations describing the pressure gradients within the pile (Sokhansanj et al., 1997), and used a technique developed by Brooker (1959) to linearize constants a2 and b2 in Eq. (2). The predicted flow patterns agreed reasonably well with the experimental data. From the reviewed literature it has become clear that the existing data for air flow versus pressure drop for pellets has several shortcomings. Firstly, the data for alfalfa pellets are not consistent. Secondly, the data do not cover low air flow ranges prevalent in storing and aeration of wood pellets. Lastly, the variability in particle lengths and bulk density of pellets and their effects on air flow pressure drop was not investigated. These shortcomings cast uncertainty on the adaption of existing data for feed pellets towards wood pellets. So the objectives were to find the relationship between the pressure drop and air flow as it is affected by particle size (bulk density) and to evaluate the common existing equations for air flow resistance to determine the equation which can express the data best. 20 2-2. Materials and methods 2-2-1. Experimental apparatus Figure 2.1 shows a schematic diagram of the laboratory test apparatus. The apparatus consists of a transparent acrylic cylindrical column and instruments for measuring the flow rates and static pressure. The column was 290 mm in diameter and 570 mm in height. Air was introduced at the bottom of the cylinder. A plenum that contained plastic rings was provided for uniform air entry into the cylinder. The source of air was filtered compressed air from a centralized generating station. The average air temperature and relative humidity were 20°C and 30%, respectively. Two float-type in-line flow meters that covered a wide range of flow rates were used. The tested air flows range from a low velocity to near fluidization velocity. The low range flow meter (Model FL-7313 Omega Engineering Inc.) measured the air flow rates between 9.43 x 10 to 6.60 x m3 &. The high range flow meter (Model FL-2093, Omega Engineering Inc.) measured air flow rates up to 4.71 x 102 m3 s* The measured flow rates were in English units of standard cubic feet per minute (SCFM) with a precision of ±2% full scale. Pressure drop was measured along the depth of the wood pellets column via an inclined manometer (Model 26 Mark II, Dwyer Instrument Inc.) for preliminary tests and a digital manometer (Model HHP-103, Omega Engineering Inc.) for subsequent tests. The manufacturer’s specified pressure range for Model HHP-103 was 0 to 2953 Pa. Four pressure taps were located at four levels 100 mm apart along the height of the column. The lowest tap was 50 mm from the bottom of the sample. Air flow from the compressed air line passed through a series of filters, pressure regulators and through the float-type airflow meter before entering the pellet column. 21 All tests in this study were conducted in three replicates. For low airflow range (0.0142 to 0.1072 m pressure drop was measured at 14 levels of airflow rates and for high airflow range (0.1072 m s1 to 0.7148 m s’) it was measured at 18 levels of flow rates. The ambient temperature and pressure were recorded during each test. 2-2-2. Sample preparation and measurements of physical properties The wood pellet material ranges from clean uniform sizes to a mixture of different sizes. Table 2.1 lists the characteristics of two types of white wood pellets used in these experiments. Using a sieve shaker (Model TM-5, Gilson Company, Inc.) and trays with openings of 4 mm and 6.7 mm, wood pellets were fractioned to three different size categories- L>6.7mm,4 mm<L<6.7mm, and L< 4 mm. The orientation of the pellets on the trays during the screening process is very important; they may orient horizontally which is desirable. Some pellets broke into smaller pieces due to the shaker vibration upon classifying the pellets and preparing samples. The length of sieving time was shortened to minimize the pellet breakage. Pellets with L <4 mm considered as fines, and they were discarded. The length and diameter of about 200 pellets from each of the two remaining fractions were measured using a calliper. Each pellet was also weighed using a digital scale to 0.01 g precision. The ratio of mass over volume was the pellet density. The bulk porosity was determined using the following equation (Mohsenin, 1986): (7)Pp where Pb is the bulk density and p, is the single pellet or particle density. 22 The bulk density of wood pellets was determined according to a modified ASAE Standard S269.4 DEC 01(2007). A cylindrical container 152.4 mm in diameter and 122 mm in height was used for the determination of bulk density. The pellets were poured slowly into the bucket from a height of 500 mm from the bottom of the container until the container was overflowing. Excess material was removed by striking a straight edge across the top. The weight of the material with the bucket was recorded to 0.OIg precision. For tapped density, the loosely filled container was tapped on the laboratory bench five times. Filling and tapping was repeated until the container was overflowing. The filled container was weighed to 0.01 g precision. After the net weight of the samples was obtained by subtracting the weight of the empty bucket, bulk density was calculated as the ratio of mass over volume of the bulk materials. The durability of pellets was measured using a DURAL Tester (Sokhasnsanj and Crerar, 1999). The tester uses 100 g sample subjected to a rotating knife in a sealed container for 30 seconds. The treated sample was sieved through a 4 mm wire mesh sieve. Overs and unders were weighed and their percent mass of the original sample mass was expressed as breakage. 2-3. Results and discussion Table 2.1 lists the average and range of physical properties of two samples from two different biomass suppliers. The mean diameter of pellets from the two suppliers are virtually the same, at 6.4-6.5 mm, and the mean length of pellets from supplier 1 at 11.0 mm was slightly larger than that from supplier 2 at 10.5 mm. Samples from supplier 2 had a durability of 67.2% compared to those from supplier 1 at 58.1%. Both samples were clean with no dust, though sample 1 had somewhat higher bulk density than sample 2. Samples 1 and 2 had very similar particle density, at 1.13-1.14 g cm3. 23 Figure 2.2 shows the length distribution of two samples from one manufacturer. Pellets left on the 6.7 mm screen had a size ranging from 12 to 28 mm, whereas pellets that passed the 6.7 mm screen and remained on the 4 mm screen had a size ranging from 5 to 12 mm. Eqs. (1), (2) and (4) were fitted to the experimental data for each size category using a non linear least squares regression method with two software packages - DataFit 8.2 (Oakdale Engineering) and POLYMATH 6.0 (2004). Table 2.3 contains the values of the fitted parameters, their standard errors, and standard deviation of estimates (Sr). Sf,, expresses the average deviation between experimental and predicted values and is defined as follows: — E1(prjct&1 — experimenta1)2() (8)(N—2) where N is the number of data points and (N-2) is the number of degrees of freedom. For the entire range of flow rates (0.0142 to 0.7 148 m s’), the average S, for samples 1 and 2 in Eq. (1) were 21.69 and 22.15 Pa m’ respectively, while the corresponding values for Eq. (2) were 20.49 and 21.62 Pa m1. Eq. (3) fitted the data better than the other two equations with an average S, of 18.90 and 19.65 Pa m1. Table 2.4 lists the fitted parameters to Shedd’s equation for different materials including alfalfa pellets, feed pellets, 28 mm cut willow chips and 8 mm wood pellets. Shedd’ s constants for 6.4 mm alfalfa pellets estimated by Sokhansanj et al. (1993) and those estimated by Ray et al. (2004) for 6.7 mm feed pellets were compared. The calculated pressure drops from Ray et al. ‘ s data (2004) were significantly higher those calculated from Sokhansanj et al. ‘ s data (1993). The differences between their calculated pressure drops reached 80% at the low air flow of 0.01 m3 s1, though the differences decreased to 38% at greater air flow of 0.28 m3 Ray et al. 24 (2004) recorded a bulk density of 710 kg m3 for 6.4 mm pellets whereas Sokhansanj et al. (1993) recorded a bulk density of 588 kg m3 for similar size pellets. The analyzed data from this study was also shown in Table 2.4. The fitted parameters are much closer to the results derived from 28 mm cut willow chips (Kristensen and Kofman, 2000, Kristensen et al. 2003) than those of 6.4 mm alfalfa pellets and 6.7mm feed pellets (Sokhansanj et a!. 1993, Ray et al. 2004). To observe how our selected equations predict the entire range of air flow rates, percentage error of predicted versus experimental pressure drops was calculated as a function of air flow rate, for mixed sizes of pellets. Figures 2.3 shows that Eqs. (2) and (3) had a lower error at low air flows and had a better fit to the entire range of data. Ergun equation covers laminar flow and turbulent flow regimes and it seems to explain the data best. Figure 2.4 shows a log-log plot of pressure drop versus air flow rates, ranging from 0.0 142 to 0.7148 m s1, for white wood pellets. As expected, the length of wood pellets affects the bulk density; longer pellets (larger particle size) had a lower bulk density and pressure drop and conversely, shorter pellets had a larger bulk density and larger pressure drop. The highest pressure drop for longer pellets (L>6.7mm) and shorter pellets (4 mm<L<6.7 mm) was 955 Pa m1 and 2488 Pa m1, respectively. The pressure drop for mixed sizes of pellets fell in between these two extremes. It can also be seen that the plotted data are not linear; an increase of about 32% in bulk density led to almost 1.5 times increase in pressure drop. 2-4. Conclusions Bulk wood pellets manufactured from sawdust in British Columbia were subjected to air flow in laboratory tests. The pellets had an average diameter of 6 mm and length ranging from 6 to 34 25 mm. The pellets were screened to three sizes - those with length shorter than 4 mm, between 4 and 6.7 mm, and longer than 6.7 mm. The pressure drop for each of the sample group and their mix was measured over a wide range of air flow rates. The pressure drop was observed to increase with air flow rate and such increases became more profound for higher air flows. Smaller wood pellets had the highest resistance to air flow. Increasing the size of wood pellets or using a mixture of different sizes decreased the resistance. It may be assumed that larger particles will lead to more direct and straight pathways for the air to pass, thus resulting in lower resistance to air flow. Among the three predictive equations (Shedd, Hukill-Ives and Ergun) studied, the Ergun equation appeared to provide the best fit to the data for the entire range of air flows. This could be attributed to two additive terms representing linear and quadratic function of air velocity for both laminar and turbulent flows in the Ergun Equation. 26 Figure 2.2 Size distribution for sample 1 (L>6.7mm) and sample 2 (4mm<L<6.7mm). Compressed Air ___________ Tank AirOut (d=0.29m) Cyfinder diameter cc FL- 13 i V-6 Introduced to the cylinder by this line Manometer Model I Markil, Der instrument Inc. Figure 2.1 Equipment for measuring the resistance of wood pellets to air flows Size Distribution for Sample 2 (4mm<L<6.7mm) (200 Particles Measured) 50 40 30 20 10 54 Size Distribution for Sample 1 (L>6.7mm) (200 Particles Measured) 200 41 I 60 50 40 30 20 10 0 60 >5 C? 0.1 41 a. 52 -n 8 121620242832364044 Length Size (mm) 0 4 5 6 7 8 9 10 Length Size (mm) 27 10 0 -10 - 0 30 -40 -50 -60 -70 -80 -90 Figure 2.3 0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 Air Velocity (m s1) Percentage error in the prediction of pressure drop for mixed sizes of wood pellets as a function of air velocity using three models. • mixes of sizes • L>6.7mm A4mm<L<6.7mm 1.00 0.10 0 0.01 10 100 1000 Pressure drop per unit depth (Pa rn’) 10000 Figure 2.4 Resistance of white wood pellets with different sizes to air flow ‘ ‘5*5555 1 • • x x A A A x ., • XA A )‘A A Shedd Model (Model 1) X Hukill & Ives’ Model (Model 2) A • Ergun Model (Model 3) 28 Table 2.1 Physical properties of pellets made from spruce saw dust Physical property Sample 1 Sample 2 Diameter (mm) 6.5a(6367)b 6.4 (6.3-6.5) Length (mm) 11.0 (3.6-41.2) 10.5 (4.5-36.4) Moisture content (%) 4.6 (4.5-4.8) 4.6 (4.4-4.8) Durability (%) 58.1 (56.7-59.5) 67.2 (66.0-68.6) Bulk density (kg m3) 800 (793-804) 761 (757-774) Particle density (g cm3) 1.13 (1.03-1.23) 1.14 (1.03-1.26) Porosity 0.29 0.33 a mean value; b range of values Table 2.2 Air flow rate and pressure drop data for three size categories of pellets Flow rate Pressure drop (Pa m’) (m3 s’) Mixed Sizes L>6.7mm 4mm<L<6.7mm 0.00094 8.7 3.3 13.2 0.00141 12.4 4.9 19.9 0.00943 115.8 51.4 164.4 0.01651 296.4 137.0 413.5 0.01887 387.3 172.7 526.5 0.03067 868.0 408.5 1209.1 0.04247 1556.7 743.2 2021.3 0.04719 2006.3 963.3 2488.1 29 Table 2.3 Estimated constants for the various predictive models Equation 1 (Shedd ‘s Equation) Standard Error Particle Size Sample a1 b1 a1 Mixed sizes 1 0.01 12 0.5482 6.6574E-4 8.1852E-3 25.49 L>6.7 mm 1 0.0198 0.5244 1.0000E-3 8.3829E-3 12.94 4mm<L<6.7mm 1 0.0060 0.5953 3.7568E-4 7.3010E-3 26.66 Mixed sizes 2 0.0127 0.5462 1.0168E-3 1.1319E-2 28.74 L>6.7mm 2 0.0173 0.5393 1.1082E-3 9.7420E-3 15.57 Equation 2 (Hukill- Ives Equation) Standard Error Particle Size Sample a2 b2 S, a2 Mixed sizes 1 22003.69 452.52 2931.88 337.46 24.35 L>6.7mm 1 18879.30 43872.48 5469.37 1291.51 12.79 4mm<L<6mm 1 16332.10 41.64 1019.01 8.67 24.34 Mixed sizes 2 18334.52 523.02 3557.07 582.38 28.01 L>6.7mm 2 13182.17 1300.51 2640.96 1730.79 15.23 Equation 3 (Ergun ‘S Equation) Standard Error Particle Size Sample a3 b3 S,, a3 Mixedsizes 1 337.11 3354.26 40.69 72.18 20.84 L>6.7nim 1 98.77 1687.07 22.06 39.14 11.30 4mm<L<6mm 1 742.88 3764.85 47.96 85.07 24.57 Mixed sizes 2 269.69 2731.73 49.98 88.66 25.60 L>6.7 mm 2 146.09 1731.74 26.82 47.57 13.74 30 ouaijj ISZ ’E U1MOJJ!V sIt!1pwSHOUIAuouWuoijunbppqU!‘qpu.Ji;utuorziq’L ‘qonpoij (szispxtw) (prnssqooo-ojoosco s11dwI.ui9 (Eoo)JSflpM1SWOJJ 00E0-0100L16co 11100 ps1pdROOM11W] 8 (ooo)utwjoSdq3MOJIIM1fl oc9o-cooc8coL1t00 puUSUOS!J)J ww sjpd (Eoo)I’M9EO-6OOO91T80cooo PJIJILUL9 (E66I)o8.o-coo.o0c09016000 J1fU1sUq)O1JIiJ11111111j79 2-5. References ASABE Standards, ASAE D272.3. Resistance to Airflow of Grains, Seeds, Other Agricultural Products, and Perforated Metal Sheets 2007. Brooker, D.B., 1959. Computing air pressure and air velocity distribution when airfiows through a porous media and non-linear velocity pressure relationship exist. Transactions of the ASAE 2(1), 118-120. Ergun, S., 1952. Fluid flow through packed columns. Chemical Engineering Progress 48, 89-94. Kubler, H., 1987. Heat generating processes as cause of spontaneous ignition in forest products. Forest Products Abstracts 10, 299-327. Gayathri, P., Jayas, D.S., 2007. Mathematical modeling of airflow distribution in grain bulks — a review. ASABE paper no. 076226. St. Jospeh, MI. Giner, S., 1996. Pressure drop through wheat as affected by air velocity, moisture content and fines. J. Agric. Engng Res 63, 73-86. Gu, D., Sokhansanj, S., Haghighi, K., 2000. Influence of floor air entry on grain moisture content, temperature, and bulk shrinkage during ambient air in-bin drying of wheat. Canadian Agricultural Engineering 42(4), 185-194. Hunter, A.J., 1983. Pressure differences across an aerated seed bulk for some common duct and store cross-sections. Journal of Agricultural Engineering Research 28(5), 437-450. Hukill, W.V., Ives, N.C., 1955. Radial flow resistance of grain. Agricultural Engineering 36(5), 332-342. Kristensen, E.F., Kofman, P.D., 2000. Pressure resistance to air flow during ventilation of different types of wood fuel chip. Biomass and Bioenergy 18, 175-180. 32 Kristensen, E.F., Kofman, P.D., Jensen, P.D., 2003. Counter pressure on ventilation of different types of wood chip and chunk wood. Biomass and Bioenergy 25, 399-408. Li, W., Sokhansanj, S., 1994. Generalized equation for airflow resistance of bulk grains with variable density, moisture content and fines. Drying Technology 12(3), 649-667. Liodakis, S., Bakirtzis, D., Lois, E., Gakis, D., 2002. The effect of (N114)(2HPO and (NH4)(2S0 on the spontaneous ignition properties of Pinus halepensis pine needles. Fire safety journal 37(5), 481-494. Mohsenin, N.N., 1986. Physical properties of plant and animal materials, Gordon and Breach Sci. Publishers, New York. Ray, S.J., Pordesimo, L.0., Wilhelm, L.R., 2004. Airflow Resistance of Some Pelleted Feed. Transactions of ASAE 47(2), 513-519. Shedd, C.K., 1953. Resistance of grain and seeds to airflow. Agricultural Engineering 34(9), 616-619. Sokhansanj, S., Li, W., Fasina, 0.0., 1993. Resistance of alfalfa cubes, pellets and compressed herbage to airflow. Canadian Agricultural Engineering 35(3), 207-2 13. Sokhansanj, S., Lang, W., Gu, D., 1997. Advances in mathematical modeling of heat and mass transfer during through flow-air drying of cereal grain. Food Science and Technology International 3(3), 226-234. Sokhansanj, S., Crerar, W., 1999. Development of a durability tester for pelleted and cubed animal feed. ASAE 1999-01-2830. Agriculture Machinery, Tires, Tracks, and Traction SP 1474:83-87. Wadso, L., 2007. Measuring chemical heat production rates of biofuels by isothermal calorimetry for hazardous evaluation modeling. Fire and Materials 3 1(4), 241-255. 33 CHAPTER 32 LOW AIRFLOW PERMEABILITY OF WOOD PELLETS 3-1. Introduction Currently, most wood pellets in Canada are made from compacted sawdust, a by-product of saw mill operations. A typical pellet diameter is 6 mm, its length varies up to 30 mm. Pellets are dense with a single particle density of about 1.2 g/cm3. The bulk density of pellets range from 600 to 750 kg m3. Their high density permits compact storage and low cost transport over long distance. The moisture content of pellets is in the range of 5-7% wet mass basis (w.b.). Design of cooling and aeration systems for wood pellets storage is based on pressure drop versus air velocity relationships (Giner and Denisienia, 1995). Several factors affect the resistance of bulk granulated biomass to airflow including the air viscosity and density, porosity of the bulk material, orientation of the particles in bulk, size, and shape and particle surface condition. Brooker et al. (1974), Haque et al. (1978), Hukill and Ives (1955), Sokhansanj et al. (1993), Al-Yahya and Moghazi (1998), Bakker-Arkema et al. (1969), Gunasekaran et al. (1983), Henderson (1943), Montross (2005) and Siebenmorgen and Jindal (1987) have developed airflow resistance evaluations in test apparatus with circular cross—sections bin shapes. A few researchers used test bins with square (Gunasekaran and Jackson, 1988; Shedd, 1953; Hellevang et a!., 2001) or rectangular cross—sections (Kumar and Muir, 1986). The effect of bin cross section on the airflow resistance measurements was also investigated (Ray et al. 2004). The results indicated that the test apparatus cross section had no effect on the measured airflow resistance. 2 A version of this chapter will be submitted for publication. Yazdanpanah, F., Sokhansanj, S., Lau, A., Lim, C.J., Bi, X., Melin, S. Low Airflow Permeability of Wood Pellets. 34 ASAE D272.3 (2007) provides data on airflow vs. pressure drop for about 40 mostly agricultural crops, including alfalfa cubes and pellets (Sokhansanj et al. 1993). Shedd (1953) used the following equation to express the pressure drop airflow data for grains and seeds V = ai(I.XP)b (1) where V is the air superficial velocity, AP is the pressure drop and a1 and b1 are constants. Hukill and Ives (1955) suggested the following equation to relate airflow to pressure drop in agricultural materials. = a2V (2)L ln(1+bV) where AP is the pressure drop (Pa), L is the depth of the bed (m), V is superficial (frontal) air velocity (m s’). Constants a and b are tabulated in the ASAE D272. Ergun (1952) divided the flow regime in a porous packed bed to viscous and turbulent flows and used the following equation to represent these two flow regimes. Lw T = a3V + b3V2 (3) where a3 and b3 are coefficients associated with viscous flow and turbulent (dynamic) flow, (1_e)2 ii a3150 d2 U (1— b3 = 1.75 Ø63d p (5) Bed parameters are porosity e (fraction), bulk density p (kg m3); equivalent diameter of bed particles d (m); acceleration due to gravity, and viscosity of the air fluid ji (Pa s) and the shape factor (). 35 The moisture content during ventilation, storage, and handling influences packing characteristics and thus, affects the resistance to airflow in bed. The effects that moisture content has on pressure drop in beds of alfalfa pellets were formulated by Sokhansanj et a!. (1993). The specified airflow range for alfalfa pellets was 0.0053 to 0.82 m s. No other recent data are available in the literature on airflow versus pressure drop in pelleted biomass for low air flow range. The objectives of this study were (1) to determine the permeability to low air flow of bulk wood pellets, (2) to investigate the effect of moisture content on the static pressure drop and (3) to evaluate equations that relate differential static pressures in beds of bulk wood pellets subject to forced airflow. The tests were accompanied with measuring physical properties of the tested pellets. 3-2. Materials and methods 3-2-1. Materials Three different types of fuel wood pellets from three wood pellet processing plants in BC were used. All of these pellets were made from saw dust with no bark content. The pellets varied in dimensions, moisture content, bulk density, particle density and durability. Pellet length and diameter were measured using electronic calliper for 200 pellets. Density of a pellet was calculated from its volume and mass. The moisture content of the pellets were measured by oven drying at 103°C for 24 hours or until there was no further change in weight. The moisture content is expressed in wet basis (ASAE S358.2). The measurements were made in triplicate. Bulk density of wood pellets was determined according to a slight modification of ASAE Standard ASAE S358.2. A cylindrical container 152 mm in diameter and 122 mm in height was 36 used. Wood pellets were poured slowly into the bucket from a height of 500 mm from the bottom of the container until the container was overflowing. Excess material was removed by striking a straight edge across the top. The weight of the material with the container was recorded to 0.01 g precision. For tapped density, the loosely filled container was tapped on the laboratory bench 5 times. Filling and tapping was repeated until the container was overflowing. The filled container was weighed to 0.01 g precision. The net weight of the samples was obtained by subtracting the weight of the empty bucket. The bulk porosity or void space in bed was determined using the following equation (Mohsenin, 1986): (6) Pd where Pb is the bulk density and p is the single particle density of wood pellet. The durability of pellets was measured using DURAL Tester [Sokhansanj et a!., 1999]. The tester uses 100 g sample subjected to a rotating knife in a sealed container for 30 s. The treated sample was sieved through a 4 mm wire mesh sieve. Overs and unders were weighed and their percent mass of the original sample mass was expressed as breakage. 3-2-2. Methods Experimental apparatus for airflow-pressure measurement consisted of a transparent acrylic cylindrical pellet container and instruments for measuring the flow rates and static pressure (Figure 3.1). The container was 290 mm in diameter and 570 mm in height. The container height was extended to 1140 mm for low airflow measurements. A plenum that contained plastic rings provided for uniform entry of air into the container. The source of air was from the centralized compressed air generating station. 37 Air was filtered before entering the container. The air flow from the compressed air line passed through a series of filters, pressure regulators and through the float-type airflow meter before entering the pellet column. The static pressure was measured 50 mm from the bottom of the cylinder. Four float in-line flow meters were used to provide measurements of a wide range of airflow. The first two flow meters were used for tests on measuring low flow rates (0.0002 to 0.0220 m s’) and the latter three flow meters together (0.0066 to 0.4646 m s’) were used to test and compare the pressure drop along the bed on three different sides and also for tests on moisture content. Table 3.1 lists the flow meter models and the measurement ranges. Air pressure was measured using an inclined manometer and two digital manometers. Table 3.1 lists the flow meter and manometer models and the range of measurements. Four pressure taps were located at four levels 100 mm apart along the height of the column. Air was induced from low flow. Pressure drop was measured through beds of wood pellets at air flow rate from 0.0002 to 0.0220 m The experiments for the air flow rate and moisture content were done with clean sieved wood pellets. All tests in this study were conducted in three replicates. For low air flow range (0.0002 to 0.0220 m 1) pressure drop was measured at 25 levels of airflow rates. The ambient temperature and pressure were recorded during each test. 3-3. Results and discussion Table 3.2 shows the physical properties of the sample used for the tests of low permeability. The sample used was a mixture of sizes (diameter 5.9 to 26.4 mm) with low moisture content (2.8%). The original bags of pellets contained 1.4 % of fines. The fine is defined as particles 38 that pass through 4 mm wire mesh screen. The wall effect on pressure drop measurement was neglected. According to Carman (1937), these effects on the overall pressure drop in the bed can be neglected if the diameter of the bed is greater than 10 times the diameter of the spherical particles in the column. As the materials in the bed are not spherical, based on the Geankoplis (1993), the ratio of the diameter of the column of packed materials to the effective particle diameter should be a minimum of 8:1 to 10:1 for wall effects to be small. Geankoplis (1993) proposed the effective particle diameter to be calculated by: 6 D=— (7) where a is the specific surface area of a particle in m2/m3. It is obtained by dividing the surface area of a particle by its volume. For a pellet 6 mm in diameter and 12 mm length the specific area is 0.75 and D8 mm. The required diameter of the container should be at least 80 mm. The container used for bulk density measurement has a diameter of 290 mm. Table 3.3 shows the properties of the samples used in tests on moisture content. The samples used had almost the same size distribution while their moisture content was significantly different. As shown in figure 3.2, the size distribution of sample 1 (MC = 5.4%) and sample 2 (MC = 8.1%) are almost the same and just slightly higher than that of sample 3 with MC = 2.8%. The pellet with 8.1% moisture content seemed to have lower resistance to air flow. Figures 3.3 show a log-log plot of pressure drop versus air flow for wood pellets. The flow rates used ranged from 0.0002 to 0.0220 m s. The plotted data on log-log scale are linear. In this sample, for a bulk pellet of mixed sizes where pellet length ranges from 6 mm to 26.4 mm, the pressure drop started from 0.18 Palm and went up to 8.30 Pa m1. 39 Figure 3.4 shows the air flow pressure drop data for three types of pellets. As all pellets are all mixes of sizes and their size distributions and bulk densities are almost the same, it can be concluded that the slight decrease in pressure drop with increasing air flow was due to an increase in moisture content. Similar to agricultural grains, the higher the moisture content is, the lower the pressure drop is (Chung et al., Giner et al., Pagano et al., Montross et al., Li et al. and Sacilik et al.). Although the three samples used are not the same samples, their size distribution and bulk densities are very close. Results from previous experiments (Chapter 2) to determine the relationship between bulk density and pressure drop show that 30-35% changes in the bulk density would change the airflow resistance by 20-25%. The bulk densities of the samples used in these test differ by about 5%, while the differences in pressure drops among the three treatments exceeds 30%. Therefore, it may be concluded that the differences in pressure drops are due to the different moisture contents of the material. Moreover, aside from particle size/bulk density, other factors such as bed porosity, surface roughness of the pellets and the porosity of the pellet itself could have affected the pressure drop somewhat. All the air flow pressure drop measurements have been done on one side of the column (circular bed) up to now. To assess the possible symmetry in pressure drop data in the column, measurements were then made on three sides of the bed - left, right and front. The air flow range was split up into two ranges - 0.0023 to 0.02200 m s (low airflow range) and 0.0 142 to 0.7148 m s’ (high airflow range) and the pressure drop data were compared. Results from these measurements show that pressure drops were essentially the same along the bed of wood pellets on all sides. For the low airflow range, all three sets of measurements follow a linear trend, and for the high range of air flow they all lie in the same quadratic graph. 40 Some factors which affect the quality (moisture content, durability and size distribution) of pellets can subsequently affect on the resistance of material to air flow. As mentioned by Lehtikangas et al., (2000) storage of such pelletised materials would affect the properties of the material. The average pellet length was decreased due to breakage during storage. Storage for long period seemed to have negative effect on durability of pellets but the average bulk density or the density of individual pellets did not change significantly during storage. Moreover high moisture pellets could promote microbial growth and all kinds of pellets have a tendency to obtain equilibrium moisture content during storage. In all tests the bed porosity varied between 0.3 to 0.4. The single pellet porosity was about 0.10 and up to a maximum of 0.29. Comparing the values of the pellet porosity and the bed porosity, it is found that the single pellet porosity itself is not so low to be neglected in calculating the resistance to airflow and is significantly notable. Three airflow versus pressure drop equations 1, 2 and 3 were fitted to the entire range of experimental data. Non-linear least squares regression method with two software packages (DataFit 8.2, Oakdale Engineering; Polymath 6.0 (2004)) was used. Table 3.4 contains values of fitted parameters, their standard errors, and standard deviation of estimates (Sr). S, expresses the average deviation between experimental and predicted values and is defined is as follows: — ti1(pred cted — experimenta1)2(!) (8)(N—2) where N is the number of data points and (N-2) is the number of degrees of freedom. For the flow rate range 0.00012 to 0.7148 m s1, the average S,, in equation 1 was 8.41 Pa m1. This value for equation 2.0 was 7.39 Pa m’. Equation 3 fitted data better than the other two equations with the average S of 7.12 Pa m1. 41 Table 3.5 lists fitted parameters to Shedd’s equation for different materials including alfalfa pellets, feed pellets, 28mm cut willows and 8mm wood pellets. The fitted parameters for this study are much closer to the results for 28mm cut willows from Kristensen and Kofman than those of 6.4 mm alfalfa pellets and 6.7mm feed pellets, although the air flow rate used in this study is much broader than the other researches in the table. To be able to observe how our selected equations predict the entire range of air flow rate, the error percentage of pressure drop prediction was calculated as a function of airflow rate. (predicted — experimental1) Relative Error = 100 . (9) experimental Figure 3.5 shows that equation 2 and equation 3 had a lower error at low airfiows and had a better fit to the entire range of flow rate (The whole range of flow rate covered by four flow meters was used for model fitting and error calculation (0.0002 to 0.6791 m sd). Ergun equation covers viscous (laminar flow) and turbulent flow regimes and it seems to explain the data best. The flexibility of the Ergun-type equation [model 3, Eqn 3.] can be described to its two-term nature, which reflects the transition from laminar flow (linear term) to fully turbulent flow (quadratic term). 3-4. Conclusion The resistance of wood pellets was measured as a function of wood pellet moisture content and the column side on which the pressure drop measurement was done. The low airflow permeability was also investigated. The data were fitted into Shedd, Hukill and Ives and Ergun type model to see how well they behaves and difference between the calculated pressure drop and experimental pressure drop shows a quite good fit with the Hukill and Ives model and a better one with the Ergun Model. The tests on moisture content of pellets shows a slight 42 decrease in the resistance as the moisture content increases. Measurements on different sides of bed seems the same pressure drop and showing that the pressure drop along the column is the same. 43 Air Out V 6 al Manometer Digital lO3 Figure 3.1 Equipment for measuring the resistance of wood pellets to air flows. 25 20 5 8 12 16 20 24 28 32 36 40 44 More Length Size (mm) Figure 3.2 Size Distribution for Samples 1, 2 and 3 in moisture content tests QDW gat vs e Cder deter (dO.29m) 0 0 Air Introduced to the cylinder by this line 30 28 • 1O.37mm<L<33.45mm, MC=5.4% 24 kJ11.14mm<L<36.13, MC=8.1% 20 5.95mm<L<26.44, MC=2.8% 14 1 9 10 j 15 L 76 622 22 200 1 100 040 000 0000 44 0.1000 0.0100 0.00 10 0.000 1 0.10 1.00 Pressure Drop (Pa rn’) Pressure Drop (Pa rn’) 10.00 Figure 3.4 Experimental pressure drop per unit depth of bed height as a function of air flow for three different moisture content (Particle size range was between 5.9mm to 36.1) 0) E --. .- — +* -__ 1 —.— —-V -. - —V.——.—— — V-V__ _ * — * V..#. - ..V. * Figure 3.3 Resistance of a mix of wood pellets to very low airflow (Flowrate from 0.0002 to 0.0220 m s1) )C 10.3mm<L<33.4mm, MC=54% —5.9mm<L<26.4mm, MC=2.8% + 11.lmm<L<36.lmm, MC=8.1% E Cl 1.000 0.100 0.010 0.00 1 -__ I — -— V ! —-.--H H :: f+2r ,-* -1 - -— 1 +*+ —---- ———— V.,— +1 — — : : : : zzz. ::::t.zz.. ;j. zE 1.00 10.00 100.00 1,000.00 45 9V ipowdj-un.i()pouSAIpiiHPIH(z)ipomd(,-ppq(I):sppowE (JP0pA.I!Ijouopunjsido.punssaidjouo!pIpudu.io.i.itju.ijç (1sw)iJP0IA‘!V 080OCO090oco0v00E000OFO000 08-•o :: 0I- 001- 0 Table 3.1 Flow meters and manometers model and measurement ranges Flow Meter Type Air flow range FL-i 12 0.000 1 ito 0.00058 m s FL-i 15 0.00044 to 0.02200 m s1 FL-73i3 0.0142 to 0.1072 m s FL-2093 0.1072 to 0.7148 m s’ *M1 flow meters from Omega engineering Inc. and Accuracy ±2% full scale Manometer Type Pressure difference range Model 26 Mark II 0 to 1743.574 Pa Mode1FIFIP-103 0 to 2953 Pa Model 475-000-FM 0 to 249.1 Pa *Type 26 Mark II and 475-000-FM were from Dwyer Instrument Inc. and HHP-103 from Omega Engineering Inc. Table 3.2 Properties of wood pellets used in low permeability test Physical Property Sample Diameter (mm) 6.2a (5.8 — Length (mm) 12.6 (5.9 —26.4) Moisture content (%) 2.8 (2.7-2.8) Loose bulk density (kg m ) 720 Particle density (g cm) 1.27 (1.08-1.35) Bed Porosity (s) 0.43 Durability 69.20 a mean value; b range of values 47 Table 3.3 Pellet properties of samples used in tests on moisture content* MC Bulk density Durability Length DiameterSample -3% (kg m ) (%) (mm) (mm) Sample 1 8.1 757 53.60 19.0(11.1—36.1) 6.3 (3.3-6.8) Sample 2 2.8 720 69.20 12.0 (5.9-26.4) 6.2 (5.8-6.4) Sample 3 5.4 761 56.12 17.5 (10.3-33.4) 6.3 (6.0-6.6) *Values in parenthesis show the range of data and values outside parenthesis show the mean Table 3.4 Estimated constants for Shedd (Eq.1), Hukill-Ives (Eq.2), and Ergun (Eq. 3). Standard Error Air Flow Range a1 b1 a1 0.00011 to 0.7148 m s_i 0.0105 0.5740 0.0002 0.0039 8.41 0.000lltoO.7148ms 12800 97.35 511 15.7 7.39 0.000lltoO.7148ms4 353 2550 14.7 27.5 7.12 Subscript i is index, i1 Shedd, i=2 Hukill-Ives, i=3 (Ergun) 48 617 (szspxiw) (prnssiq8ILO°HIOOOOoLc.ocoioo1jdwuy9 (Eoo)isnpMSWOJJ ZO0OO L16co11100 •IususiJ)JSTPdpooWW 8 (ooo)u1wJo ç9 ° cooc8coLI100 puUSUSTI)JSflUO 8 (Eoo)T°9EO°600091IWOciooo1pdWU’L9 (c66I)owo°coooooco9o16000sojpdmmV9 .11ofu1su1q)o z_misLW uaijjonpoij uPJMOflJ!V SIU!.TJ1ftHSflO!.IUAUOWIUO!J1flbsppqin‘qpnuI.spiijsuojciqi. 3-5. References A1-Yahya, S.A., Moghazi, H.M.M., 1998. Static pressure drop through barley grain. Can. Agric. Eng. 40(2), 213-217. ASABE Standards. ASAE D272.3. 2007. Resistance to Airflow of Grains, Seeds, Other Agricultural Products, and Perforated Metal Sheets. ASAE Moisture Measurement - Forage. ASAE S358.2 DEC1988 (R2008), 44th Edn. St. Joseph, 1997. Bakker-Arkema, F.W., Patterson, R.J., Bickert W.G., 1969. Static pressure-airflow relationships in packed beds of granular material such as cherry pits. Trans. ASAE 12(1), 134-136,140. Brooker, D.B., Bakker-Arkema, F.W., Hall, C.W., 1974. Drying Cereal Grains. Westport, Conn: AVI publishing. Carman, P.C., 1937. Fluid flow through granular beds. Trans. Inst. Chem. Engrs 15(1), 150-156. Chung, D.S., Maghirang, R.G., Kim, Y.S., Kim, M.S., 2001. Effects of moisture and fine material on static pressure drops in a bed of grain sorghum and rough rice. Transactions of the ASAE 44(2), 331—336. Ergun, S., 1952. Fluid flow through packed columns. Chemical Engineering Progress. 48, 89- 94. Giner, S.A., Denisienia, E., 1995. Pressure drop through wheat as affected by air velocity, moisture content and fines. J. Agric. Engng Res. 63, 73-86. Geankoplis, C.J., 1993. Transport Processes and Unit Operations, 3rd Ed. Englewood Cliffs, New Jersey: Prentice Hall. Gunasekaran, S., Jackson, C.Y., 1988. Resistance to airflow of grain sorghum. Trans. ASAE 31(4), 1237-1240. 50 Gunasekaran, S., Jindal, V.K., Shove, G.C., 1983. Resistance to airflow of paddy (rough rice) in shallow depths. Trans. ASAE 26(2), 60 1-605. Haque, E., Foster, G.H., Chung, D.S., 1978. Static pressure drop across a bed of corn mixed with fines. Transactions of ASAE 21, 997-1000. Hellevang, K., Osowski, C., Lee, T., Henson, R., Chang, S.K.C., Jou, H.J., 2001. Storage environment effect on pinto bean quality and resistance to airflow of pinto and navy beans. ASAE Paper No. 01-6061. St. Joseph, Michigan: ASAE. Henderson, S.M., 1943. Resistance of shelled corn and bin walls to airflow. Agric. Eng. 24(11), 367-369. Hukill, W.V., Ives, N.C., 1955. Radial flow resistance of grain. Agricultural Engineering 36(5), 332-342. Kristensen, E.F., Kofman, P.D., 2000. Pressure resistance to air flow during ventilation of different types of wood fuel chip. Biomass and Bioenergy 18,175-180. Kristensen, E.F., Kofman, P.D., Jensen, P.D., 2003. Counter pressure on ventilation of different types of wood chip and chunk wood. Biomass and Bioenergy 25, 399-408. Kumar, A., Muir, W.E., 1986. Airflow resistance of wheat and barley affected by airflow direction, filling method and dockage. Trans. ASAE 29(5), 1423-1426. Lehtikangas, P., 2000, Storage effects on pelletised sawdust, logging residues and bark, Biomass and Bioenergy 29, 287-293. Li, M., Sokhansanj, S., 1994. Resistance of bulk alfalfa seeds to airflow. Canadian Agricultural Engineering 36(1), 25-28. Mohsenin, N.N., 1986. Physical properties of plant and animal materials, New York: Gordon and Breach Sci. Publishers. 51 Montross, M.D., McNeill, S.G., 2005. Permeability of corn, soybeans, and soft red and white winter wheat as affected by bulk density. Applied Engineering in Agriculture 21(3), 479-484. Pagano, A.M., Corzza, D.E., Nolasco, S.M., 2000. Airflow resistance of oat seeds: Effect of airflow direction, moisture content and foreign material. Drying Technology 18(1&2), 457- 468. Ray, S.J., Pordesimo, L.O., Wilhelm, L.R., 2004. Airflow Resistance of Some Pelleted Feed. Transactions of ASAE 47(2), 513-519. Sacilik, K., 2004. Resistance of bulk poppy seeds to airflow. BIOSYSTEMS ENGiNEERING. 89(4), 435-443. Siebenmorgen, T.J., Jindal, V.K., 1987. Airflow resistance of rough rice as affected by moisture content, fines concentration and bulk density. Trans. ASAE 30(4), 113 8—1 143. Shedd, C.K., 1953. Resistance of grains and seeds to air flow. Agricultural Engineering 34(9), 616-619. Sokhansanj, S., Crerar, W., 1999. Development of a durability tester for pelleted and cubed animal feed. ASAE 1999-01-2830. Agriculture Machinery, Tires, Tracks, and Traction SP 1474:83-87. Sokhansanj, S., Li, W., Fasina, 0.0., 1993. Resistance of alfalfa cubes, pellets and compressed herbage to airflow. Canadian Agricultural Engineering. 35(3), 207-2 13. 52 CHAPTER 43 PERMEABILITY OF WOOD PELLETS IN THE PRESENCE OF FINES 4-1. Introduction Fine materials are continuously produced during drying, transporting and handling of wood pellets. Pellets mixed with broken and fine materials increases variation in resistance to airflow in a fixed bed of pellets causing uneven temperature and moisture in an aerated bin. The resulting localized high moisture or high temperature conditions may lead to heating and spontaneous combustion. Graphical data on airflow vs. pressure drop are available in ASAE D272.3 (ASABE 2007) for about 40 mostly agricultural crops and products including for alfalfa cubes and pellets. The graphical presentation is accompanied by Hukill and Ives (1955) equation. AD y2 — a2v (1) L — In (1 + b2V) where AP is the pressure drop (Pa), L is the thickness of the bed (m), V is superficial (frontal) air velocity (m s’). Constants a and b for clean grains and products are tabulated in the ASAE D272. For the effect of fines on resistance to airflow of shelled corn, ASAE D272 recommends equation (2) to account for the fraction of fine content in a bed of granular material. () =() (1+kf) (2)corrected clean A version of this chapter will be submitted for publication. Yazdanpanah, F., Sokhansanj, S., Lau, A., Lim, C.J., Bi, X., Melin, S. Permeability of Wood Pellets in the Presence of Fines. 53 where f is the fraction fine content in the mix. ASAE D272 gives an estimation of f as a function of airflow for shelled corn. Jayas et al. (1990) developed data on airflow-pressure drop for fine content in beds of canola grain. Similar studies were conducted by Sokhansanj et al. (1993) for alfalfa pellets and Li et al. (1994) for bulk grains. Haque et al. (1978) in studying airflow resistance of corn mixed with fines. Giner and Denisienia (1995) showed a relation between pressure drop and air flow to have to constants of k1 and k2 as: ()corrected = k1 + k2f (3) (r)clean The objectives of this study were to (1) develop experimental data on permeability of bulk wood pellets to air flow in the presence of fines, and (2) to develop an equation to predict pressure drop as a function of airflow rate and fines content. 4-2. Materials and methods 4-2-1. Materials Pellets were received in 30 kg bags from a commercial pellet producer in British Columbia. Pellets had been made from saw dust from heart wood (no bark). The moisture content of pellets was 2.8% (wet mass basis) measured according to the ASAE S358.2 (ASABE 2006) standard procedure. To prepare samples for airflow test, the contents of each bag were sieved for 5 minutes over a 4 mm wire mesh screen on a vertical Gilson TM-5 shaker. The materials left on the screen were designated “clean pellets”. The materials passed through the screen were designated as “fines”. A sample of clean pellets and fines were analyzed through a 5-sieve Tyler Rototap according to the ASAE Standard S319.3 (ASABE 2007). 54 Bulk density of wood pellets was determined according to a slight modification of ASAE Standard S269.4 (ASAE 2007). A cylindrical container 152.4 mm in diameter and 122 mm in height was used. Wood pellets were poured slowly into the bucket from a height of 500 mm from the bottom of the container until the container was overflowing. Excess material was removed by striking a straight edge across the top. The weight of the material with the bucket was recorded to 0.01 g precision. For tapped density, the loosely filled container was tapped on the laboratory bench 5 times. Filling and tapping was repeated until the container was overflowing. The filled container was weighed to 0.01 g precision. The net weight of the samples was obtained by subtracting the weight of the empty bucket. The durability of pellets was measured using DURAL Tester (Sokhansanj et al. 1999). The tester uses 100 g sample subjected to a rotating knife in a sealed container for 30 seconds. The treated sample was sieved through a 4 mm wire mesh sieve. Overs and unders were weighed and their percent mass of the original sample mass was expressed as breakage. Six mixing ratios of pellets and fines at 0%, 1%, 5%, 10%, 15 and 20% (w/w) were planned for the experiment. Roughly 20 kg biomass was needed to fill up the container. Test mixtures were prepared in small batches about 2 kg each. For instance a 2 kg of 10% mixture was prepared by mixing 200 g fines with 1800 g clean pellets. Mixing was done in a plastic container using a piece of wood stick as stirrer. A complete mixing was judged visually. Ten 2- kg batches were prepared to make up the 20-kg column. Each mixture was poured into the test column, evened out and more layers were added until the desired level was reached. The bulk density (total mass in the test column divided by volume) for each mixture was recorded. 55 Density of a pellet was calculated from its volume and mass. A manual calliper was used to measure the diameter and length of 200 randomly selected pellets from the cleaned batch. The moisture content of the pellets were measured by oven drying at 103 °C for 24 hours or until there was no further change in weight. The moisture content is expressed in wet basis (ASAE S358.2). The measurements were made in triplicate. 4-2-2. Experimental apparatus Experimental apparatus consisted of a transparent acrylic cylindrical pellet container and instruments for measuring the flow rates and static pressure (Figure 4.1). The container was 290 mm in diameter and 570 mm in height. The container was made of two sections, a plenum filled with plastic rings and the sample holder. A screen at the bottom separated sample holder from plenum. The source of air was from the University centralized compressed air generating station. Air was filtered before entering the container. Four float in-line flow meters were used to cover a wide range of flow rates. Table 4.1 shows the flow meters model and airfiows used in these sets of tests. Air pressure was measured along the depth of wood pellets column by using a digital manometers. Table 4.1 lists the model and range of manometers used. Four pressure taps were located at four levels 100 mm apart along the height of the column. The static pressure was measured 50 mm from the bottom of the cylinder. 56 4-2-3. Test procedure Once the container was filled with the test material, air was allowed to flow into the plenum and through the column. Each test was done by loading the test column with semi-packed fill. This was achieved by pouring the pellets from the height of about 0.5 m into the column of the bed gradually. This was not a fully packed fill since the material was not poured from a greater height as was done by Jayas et al. (1989), and settling was not induced after each increment of fill as was done by Beavers et al. (1973). The flow of air was started from low end. The flow was increased by manipulating a manual valve until the float in the flow meter stabilized. Pressure drop was measured in approximately 50 airflow intervals. The air flow ranged from 0.004 to 0.357 m s1. Measurements were repeated by filling the column with the same sample three times. 4-3. Results and discussion Table 4.2 shows the physical properties of wood pellets with a nominal diameter of 6.2 mm and a length of 12.6 mm. The sample used was a mixture of sizes, diameter varying from 5.8 to 6.4 mm and length varying from 5.9 to 26.4 mm. The moisture content of these samples at the time of test was low at 2.8%. The original bags of pellets contained 1.4% of fines (particles passed through a 4 mm wire mesh screen). Figure 4.2 shows the size distribution of fines. More than half of the particles that passed through 4 mm screen as fines were larger than 2 mm. About 20% of the materials were less than 0.6 mm in size. The geometric mean diameter of particles was 0.75 mm. The standard deviation of the particles was 0.57 mm. Bulk density of material in the testing column was 720 kg/rn3 (±3 kg/rn) and decreased by adding fines. Figure 4.3 shows the change in bulk density versus fine percent. Small vertical error bars in the graph 57 indicates that the standard errors were so small. Figure 4.4 shows the pressure drop vs. air velocity data for 6 samples used in tests of fine material effect. The pressure drop increased with increasing the fine content and the curves were almost parallel on the log-log scale. The pressure drop increased 50-80% times when fines percent was increased from 0 to 10% on a linear scale. The pressure drop increased 3-6 times when the percent fines increased from 0% to 20%. The wall effect on pressure drop measurement can be neglected according to Carman (1937), provided the diameter of the bed is greater than 10 times the diameter of the spherical particles in the column. For non spherical particles the ratio of the diameter of the column of packed bed to the effective particle diameter should be a minimum of 8:1 to 10:1 for wall effects to be negligible. Geankoplis (1993) defines the effective particle diameter as, D=— (4) a,, where a is the specific surface area of a particle in m2/m3 and is obtained by dividing the surface area of a particle to its volume. For clean wood pellets Dp ranges from 0.007 to 0.008 and the ratio of the column diameter to particle diameter is 34 to 39. Therefore we can neglect the wall effect. Hukill and Ives equation was fitted to experimental data using non-linear least squares regression using two software packages (DataFit 8.2, Oakdale Engineering; Polymath 6.0 (2004)). Table 4.3 lists estimated a1 and b1 and their standard errors and correlation coefficients. The Hukill and Ives equation for clean pellets is: zXP 5180Q2 L = In (1+ 19.05Q) 58 where (zXP)ciean is pressure drop in Pa m1 and Q is airflow in m3 s of volume flow per m2 of the frontal area. Figure 4.5 shows also the plot of equation 5 over the entire range of airflow. Table 4.3 lists the estimated values of al and b1 as fines content increases. The estimated a increases with increased fines content. The estimated values for b1 do not show any trend. The value of b1 was fixed at 19.05 for clean pellets and estimated new values of aj as a function of fines content, a1 = 5180 (0.83 + 20.70 R2 = 0.98 (6) Equation 6 is similar to Jayas’s equation 3. The intercept was set to y=1 at x=0 to develop the following equation similar to equation 2, a1 = 5180 (1 + 19.6Q R2 = 0.97 (7) where f is the fines content in decimal fraction. Figure 4.4 shows the predicted pressure drop using equation 7. The fit is not satisfactory. Figure 4.5 shows the predicted pressure drop versus airflow for two fines content. Compared to the experimental values, the fitting is not satisfactory. The constant k needs to be calculated with changes in flow rate to be put in the modified equation for airflow versus pressure drop relationship. The following linear equation of k as a function of airflow was fitted to the data. (Figure 4.6) k = —39.685Q + 2 3.773 (8) Substituting equation 8 in equation 2 resulted in the following equation among pressure drop, air flow rate and fine material percentage. (P)corrected = (IW)ciean [1 + (39.685Q + 23.773)fj (9) Pressure drop predicted by this equation is close to the experimental data especially in airflow range of 0.004 to 0.103 m s (Percent error of 0.003 to 0.011). The results from alfalfa pellets show a linear relation between k and Q values similar to equation 8. 59 Eqs. (1) and (9) were fitted to the experimental data for wood pellets mixed with 1% fines using a non-linear least squares regression method with software package - DataFit 8.2 (Oakdale Engineering). Deviation of estimates (Sr) was calculated for both equation predictions. S expresses the average deviation between experimental and predicted values and is defined as follows: = (l=l(PredictedI — exPerimentali)2)() (10)(N—2) where N is the number of data points and (N-2) is the number of degrees of freedom. For the range of flow rates (0.004 to 0.103 m s1), the average S for samples in Eq. (9) was 0.68 Pa m’ respectively, while the corresponding values for Eq. (1) was 1.22 Pa m1. Eq. (9) fitted the data better than the Hukill and Ives equation with a lower S,, value. 4-4. Conclusion The resistance of wood pellets to airflow was measured as a function of fine materials in beds of wood pellets. The pressure drop increased with the increase in fine percent. By increasing the fine percent from 0 to about 10%, the pressure drop increased by 50-80% depending on the airflow rate. The pressure drop increased up to 3-6 times of the initial values by increasing the fine materials in bed from 10% to 20%. Hukill-Ives model was fitted to the data to and the equation was corrected for fines content. 60 Table 4.1 Flow meters and manometers model and measurement ranges Flow Meter Type Air flow range FL-i 12 0.00011 to 0.00058 m s FL-i 15 0.00044 to 0.02200 m s FL-7313 0.0142 to 0.1072 m s FL-2093 0.1072 to 0.7148 m s1 *All flow meters from Omega engineering Inc. and Accuracy ±2% full scale Manometer Type Pressure difference range Model 26 Mark II 0 to 1743.574 Pa Model HHP-103 0 to 2953 Pa Model 475-000-FM 0 to 249.1 Pa *Type 26 Mark II and 475-000-FM were from Dwyer Instrument Inc. and HHP-103 from Omega Engineering Inc. Table 4.2 Properties of wood pellets used in low permeability test Physical Property Sample aDiameter (mm) 6.2 (5.8—6.4) Length (mm) 12.6 (5.9 — 26.4) Moisture content (%) 2.8 (2.7-2.8) Loose bulk density (kg m3) 720 Particle density (g cm3) 1.27 (1.08-1.35) Bed Porosity (6) 0.43 Durability 69.20 a mean value; b range of values 61 Table 4.3 Results of fitting HukiJi and Ives equation to data a1 Fine% a1 b1 s.e. (ai) s.e. (b1) s.e.(ai) (b1=19.05) 0% 5180 19.05 124.8 0.94 5180 16.79 1% 5206 18.38 213.1 1.38 5306 29.74 5% 8198 15.03 256.1 0.73 9490 40.46 10% 13215 15.42 445.3 1.07 14570 64.17 15% 15236 12.24 288.1 0.44 18824 84.83 20% 25412 16.37 1589.0 2.09 27300 223.25 Compressed Air Tank Air Out Cylinder diameter (d0.29m) CO FL- 093 ::::::: FL- 313 :::..........:.:. 0 3 V-6 I Air Introduced to the cylinder by this line Digital Manometer Model 26Mark 11,Model HHP-103 DwyerInstrument Inc. Figure 4.1 Equipment for measuring the resistance of wood pellets to air flows. 62 0.45 0.397 0.40 0.35 0.30 a . 0.25 0.20 0.169 r,) 0.15 0.116 0.112 0.10 I II 0.062 0.0510.037 0.026 0.0 13 0.0260.050.00 2.82 2 1.4 0.85 0.707 0.5 0.354 0.25 0.177 0.125 0.106 Nominal Seive Size (mm) Figure 4.2 Typical particle size distribution of fine materials in bed of wood pellets. 750 730 710 690 670 650 3I a 630 I ‘ 610 .0976x + 716.8590 R2= 0.993 1 570 550 0.00 5.00 10.00 15.00 20.00 25.00 Fine content (%) Figure 4.3 Bulk density of wood pellets as a function of fine content. 63 X Fines=0% — Fines=1% Fine5% A Fines=10% 0 Fine=15% + Fines2O% 0.200 a a 0 0.020 0.002 Figure 4.4 Experimental pressure drop per unit depth of bed height as a function of air flow for four different fine contents. - Experimental Pressure Drop for pellets with 0% Fines 0.200 — — — Predicted Pressure Drop for Pellets with — 10% Fines + Experimental pressure Drop for 0.020 Pellets with 10% Fines Predicted Pressure Drop for Pellets with 0% fine 0.002 1.00 1,000.00 Figure 4.5 Predicted and experimental pressure drop per unit depth of bed height according to Hukil and Ives Equation as a function of air flow for two different fine contents. 1.00 10.00 100.00 1,000.00 Pressure drop (Pa m’) 10.00 100.00 Pressure drop (Pa rn) 64 c9 MouJnJouoipunjtuoipnbJouisuo9t (1sLw)AoJJ.IIV 9F0t7FOFOOFO8009001700OO000 0081 0061 oo.oz oo.I 00E1 OOI7 ooc 4-5. References ASABE Standards. ASAE D272.3. 2007. Resistance to Airflow of Grains, Seeds, Other Agricultural Products, and Perforated Metal Sheets. ASABE Standards. ASAE S319.4. 2008. Method of determining and expressing fineness of materials by sieving. ASABE. 2006. Moisture measurement - forages. Standard S358.2. In ASAE Standards. St. Joseph, MI: ASABE. Beavers, G.S., Sparrow, E.M., Rodenz, D.E., 1973. Influence of Bed Size on Flow Characteristics and Porosity of Randomly Packed-Beds of Spheres. Journal of Applied Mechanics-Transactions of the ASME. 40(3), 655-660. Carman, P.C., 1937. Fluid flow through granular beds. Trans. Inst. Chem. Engrs. 15(1), 150- 156. Giner, S.A., Denisienia, E., 1995. Pressure drop through wheat as affected by air velocity, moisture content and fmes. J. Agric. Engng Res. 63, 73-86. Geankoplis, C.J., 1993. Transport Processes and Unit Operations, third ed. Englewood Cliffs, New Jersey: Prentice Hall. Haque, E., Foster, G.H., Chung, D.S., 1978. Static pressure drop across a bed of corn mixed with fmes. Transactions of ASAE. 21, 997-1000. Hukill, W.V., Ives, N.C., 1955. Radial flow resistance of grain. Agricultural Engineering. 36(5), 332-342. Jayas, D., Sokhansanj, S., Moysey, E.B. and Barber, E.M., 1990. Predicting airflow pressure patterns in canola (rapeseed) bins. Can. Agric. Eng. 32(2), 249-254. 66 L9 Etl-LO‘(E)cEUTJUUIEJnnno!JvuinpujMoJJJToqioq pssaIdmo3puispd‘sqiwtjpjpjoE661“00‘uutsij‘‘I]‘g‘ftiisuitp 1Og L8E8‘I7LtT -asUO!OIPU‘PJI‘!I“MI’’YO8l-IO-666IaVSVPJPU1! pqnpupoijpdJOjJSijiqiinp1JO1UUKJOPAU666t‘‘g‘1uisuipjo .8-c‘(j)cUIUU P11Wt!JVuiiptnrnMOjJIU?OspsJJJUWNJ066{‘S‘CUSUpJOg‘vj 96z-c6‘(T)vsvqjosuoisu1ij(posdij) 1OtID{IOI4M01I-1WJ00TImS!SIU-!SU6861‘5‘fUsUq3Og‘J‘sAf CHAPTER 5 DISCUSSION, CONCLUSION AND RECOMMENDATION FOR FUTURE WORK In this study, the permeability to air flow of bulk wood pellets was determined and equations were developed that relate differential static pressures in beds of bulk wood pellets subject to forced airflow. Experimental data on permeability of bulk wood pellets developed for pellets with different moisture content and also for bulk of material in the presence of fine materials and their effects on the static pressure drop was investigated. Equations were developed to predict pressure drop as a function of airflow rate and fines content. 5-1. Discussion and conclusions Experiments on wood pellets show more resistance from smaller pellets than larger pellets. Mixtures of pellet sizes show less resistance to airflow. This increase in airflow resistance can be described by the availability of larger spaces and more pathways for airflow to pass. For the airflow range of 0.014 to 0.8 m s’, the maximum pressure drop measured was 2550 Pa m’. Three models (Shedd, Hukil and Ives and Ergun models) (Hukill and Ives 1955, Shedd 1953 and Erqun 1952) have been fitted to experimental data. An error analysis showed the best fit to the entire data is achieved by Ergun equation. The effects of other factors such as moisture content of pellets and the presence of fine materials in the bulk of wood pellets on the resistance data were tested. The airflow rate used in these test was from 0.0042 to 0.7148 m s1. As shown in table 3.3, the 3 types of pellets moisture content were 2.8 %, 5.4% and 8.1% respectively. As the moisture content increased, the pressure drop decreased slightly. Although the three samples used were not the same 68 samples but analyzing their size distribution and comparing their bulk densities, their size distribution and bulk densities were close. Results from experiments on the relationship between the bulk density and pressure drop, show that even 30-35% changes in the bulk density, will change the resistance 20-25%. Comparing the bulk densities of the samples used in these test show that the difference is about 5%,. while the changes in pressure drop data exceeds 30%. It can be concluded that portion of this changes is due to the difference in moisture level of the material. Chapter 4 presented the results of experiments on the resistance of wood pellets to air flow in the presence of fine materials. The tested airflow range was from 0.004 to 0.357 m s’.Results from tests show continuous decrease in bulk density of pellets as fine content increases. Due to lower porosity in the bed as a result of fine presence, the pressure drop data shows an increasing trend as fines are added. A range of 3-6 fold increase in pressure drop was measured in beds containing 20% fine material compared when compared to beds of clean wood pellets. To evaluate the location of pressure taps, the measurements on pressure drop was done on 3 sides of bed at the same level. The air flow range was split up to 2 ranges 0.0023 to 0.0220 m (low airflow range) and 0.0 142 to 0.7 148 m (high airflow range). The pressure drop data in 3 different sides were compared. No discernible differences between pressure drops were recorded. Of all major parameters tested in the thesis research experiments, and within the ranges of values tested, airflow rate was found to have the most significant effects, followed by pellets particle size and fines content. Moisture content was found to have the least effect on resistance to airflow. 69 01. SOflJtAUvpawanooU13 uotwnbasppaqgojpougaaa&uoipastpeauotP?Uso2uei(ij2ajpuoajppuu‘MO’I)£Oa2umi MOJp3aqipDp!A!ppu1A&OjJLtU3oi(saug%O)soijadtwajojoaournsTsaluornpaqipasna Table 5.1. Results of fitting Shedd equation to Q - AP data for clean pellets Airflow Range A B s.e. (A) s.e. (B) 0.006-0.022ms’ 3.44x10 0.89 2.49x10 3.85 x102 0.024-0.06 m s 3.44x10 0.90 3.27x10 3.18 x102 0.067-0.178 m s 5.27x1O 0.75 2.26x1O 1.04 x102 Table 5.2. Results of fitting Shedd equation to Q - AP data for pellets with 1% fines Airflow Range A B s.e. (A) s.e. (B) 0.006-0.022ms’ 2.37x10 1.01 1.29x10 2.72 x102 0.024-0.06ms’ 3.07x10 0.92 2.35x10 2.51 x102 0.067-0.178ms’ 5.06x10 0.76 2.32x10 1.10x102 Table 5.3. Results of fitting Shedd equation to Q - AP data for pellets with 5% fines Airflow Range A B s.c. (A) s.e. (B) 0.006-0.022 m s1 1.77x10 0.96 1.22x10 2.82 x102 0.024-0.06 m s 2.31x10 0.85 1.99x10 2.38 x102 0.067-0.178ms’ 3.65x10 0.73 1.32x10 7.52 xiü Table 5.4. Results of fitting Shedd equation to Q - iP data for pellets with 10% fines Airflow Range A B s.e. (A) s.e. (B) 0.006-0.022 m s 1.69x10 0.86 3.84x10 8.60 xiü 0.024-0.06 m s 2.00x10 0.79 1.40x10 1.71 x1112 0.067-0.178 m s 2.67x1W3 0.72 1.37x10 8.83 xi 71 Table 5.5. Results of fitting Shedd equation to Q - AP data for pellets with 15% fines Airflow Range A B s.e. (A) s.e. (B) 0.006-0.022 m s4 1.04x10 0.91 9.21x10 2.96 x1O2 0.024-0.06 ms’ 1.87x10 0.75 2.29x10 2.80 x102 0.067-0.178 m s1 1.54x10 0.79 6.02xl0 7.06 xi(y3 Table 5.6. Results of fitting Shedd equation to Q - AP data for pellets with 20% fines Airflow Range A B s.e. (A) s.e. (B) 0.006-0.022 m s’ 6.89x10 0.93 2.57x10 1.08 x102 0.024-0.06ms1 1.34x10 0.75 l.37x10 2.13 x1O2 0.067-0.178ms 7.14x10 0.87 i.OixiO 2.46 x102 Tables 5.1 to 5.6 show the changes in A and B values as the fine material increases. As seen in the tables for all three range of airfiows (0.006-0.022 m s_I, 0.024-0.06 m s1and 0.067-0.178 m s’), A value decreases as the fine material increase and B value shows to be changing around 0.7 -1.0. Sokhansanj et al. (1993) has fitted these values for 6.4mm alfalfa pellets for the airflow range of 0.0053 - 0.82 m s which falls within the first two ranges of airflow I tested. Inspection for A and B values for wood pellets mixed with fines showed that while A values changed test to test, changes in B values was insignificant. By assigning a constant value for B, new values for A were obtained. New A values were plotted versus fine % to find the trend and relationship. Table 5.7 summarizes the obtained results. Figure 5.1 also shows a typical graph for changes in A value as fine content increases for the flow rate of 0.024-0.06 m 72 Table 5.7. Relationship between A values and fine material (Constant B) Airflow Range Equation for A versus f B 0.006-0.022 m s A = -0.000lf+ 0.0029 0.93 0.92 0.024-0.06 m s1 A = -0.000lf+ 0.0033 0.82 0.94 0.067-0.178 m s1 A = -0.0002f+ 0.0046 0.77 0.89 4.OOE-03 — 3.50E-03 + 3.OOE-03 2.50E-03 y-O.000lx+O.0033 R4O94362.OOE-03 1.50E-03 +1 .OOE-03 - 5.OOE-04 O.OOE+OO —________ __________ ____ 0.00 5.00 10.00 15.00 20.00 25.00 Fine Content (%) Figure 5.1 Constant A of Shedd equation as a function of fine content for the air flow range of 0.024-0.06 m s (B0.82). 73 5-2. Recommendations for future work 1- Buoyancy induced airfiows due to differential temperature within a storage structure may be prevalent in wood pellet storage. The effect of surface roughness of pellets and single pellet porosity of pellets on these low airflow regimes need to be investigated. 2- All tests have done in this study have been in vertical direction. The same test can be done with the air flowing in the horizontal direction to compare the results in the resistance of wood pellets to airflow. 3- Expand research on that effects the ventilating air humidity have on the moisture content of pellets including 74 cL EI-LO‘()çEuuuTu[innnvurnpiuoqJo1{ possaIdwo3pusipd‘sqn3jijijoousisj66I“oo‘UtJ‘‘Fl‘‘ftUsU1pIoS 619-9t9 ‘(6)tuuouiui’minov°ijospspunsuiniJounsTsjç]‘>i‘ppqg ‘(ç)UI1UIUI1FL!VUTnJJODUSIS.1MOJJP!PIcc61‘YM‘SAI‘AM‘IIPTH t76 -68‘8t7ssao1autouuJnMW31J3suwnlo3ppndqno.iqA&OIJpInI1c6J‘s‘‘‘ SU.IJ)Jc- APPENDIX A: EXPERIMENTAL AND PREDICTED DATA Table A.1. Measured and predicted pressure drops relevant to Figure 2.3 Experimental Predicted by Hukill Predicted by Ergun Predicted by SheddAirflow (m3 s m2) Pressure Drop and lye Equation Equation Equation(Pa m’) (Pa m1) (Pa m’) (Pa m’) 0.0142972 0.0214458 0.0285944 0.035743 0.04289 16 0.0500402 0.0571888 0.0643374 0.071486 0.0786346 0.0857832 0.09293 18 0.1000804 0.107229 0.142972 0.1787 15 0.214458 0.250201 0.285944 0.32 1687 0.35743 0.393 173 0.428916 0.464659 0.500402 0.536145 0.571888 0.60763 1 0.643374 0.679117 0.71486 8.7178 12.454 16. 1902 18.681 21.1718 26.1534 3 1. 135 36.1 166 42.3436 49.8 16 54.7976 62.27 68.497 79.7056 115.8222 174.3 56 241.6076 296.4052 387.3194 488.1968 570.3932 635.154 764.6756 868.0438 981.3752 1127.087 1281.5 17 1399.83 1556.75 1820.775 2006.339 2.236723 4.268792 6.828406 9.886311 13.42172 17.41873 21.86462 26.74885 32.06249 37.7978 43.94802 50.507 13 57.46971 64.83089 107.4717 159.5345 220.6975 290.7133 369.3826 456.5394 552.0426 655.7696 767.6122 887.4742 1015.269 1150.9 18 1294.349 1445.496 1604.297 1770.697 1944.642 5.505496 8.772478 12.38228 16.33491 20.63036 25.26863 30.24972 35.57364 41.24038 47.24994 53.60233 60.29753 67.33556 74.7 1641 116.763 167.3802 226.5679 294.3262 370.655 455.5544 549.0244 651.0649 761.676 880.8576 1008.61 1144.933 1289.826 1443.29 1605.324 1775.929 1955. 105 1.549944 3.247076 5.487453 8.243791 11.49601 15.22835 19.42789 24.08373 29.18648 34.72791 40.70071 47.09831 53.91476 61. 14461 103 .3324 155.2361 216.4776 286.76 365.8401 453.5126 549.6009 653.9498 766.4216 886.8926 1015.25 1 115 1.394 1295.227 1446.665 1605.625 1772.033 1945.8 18 76 A Shedd Model (Model 1) + Hukill & Ives’ Model (Model 2) • Ergun Model (ModeL 3) 60 401 * + • 4 20 ± ± 0 . 4. . : 0 + A .0 A + • -20 A A A -40 4. -60 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 Air Velocity, m s’ Figure A.1 Absolute error using three models (Relevant to data in page 76) 77 Table A.2. Measured and predicted pressure drops relevant to Figure 3.5 Airfi Experimental Predicted by Hukill Predicted by Ergun Predicted by Shedd ( 3 Pressure Drop and lye Equation Equation Equationm S m (Pa m’) (Pa m’) (Pa m’) (Pa m’) 0.0002374 0.188679 0.03165 0.084108 0.001337 0.0003057 0.283019 0.040886 0.108352 0.002076 0.0003784 0.377358 0.050784 0.134191 0.003011 0.0004491 0.471698 0.060471 0.159338 0.004057 0.0005168 0.471698 0.069801 0.183431 0.005181 0.0005809 0.566038 0.078698 0.206287 0.006352 0.0023112 0.943396 0.337689 0.830966 0.070419 0.0028881 1.132075 0.4319 1.042621 0.103817 0.0042539 1.698113 0.669976 1.550513 0.203815 0.0048446 1.792453 0.779325 1.773144 0.255634 0.0062407 2.358491 1.052558 2.306358 0.397374 0.0074676 2.641509 1.309417 2.783186 0.54324 0.0078816 2.830189 1.399542 2.945826 0.596783 0.0089369 3.207547 1.636971 3.364318 0.742816 0.0104541 3.962264 1.997399 3.97599 0.976156 0.0127035 4.622642 2.571919 4.904447 1.370752 0.0138648 5.09434 2.886812 5.393897 1.596392 0.0157658 5.943396 3.428459 6.20997 1.99687 0.0164929 6.132075 3.644071 6.526972 2.160029 0.0187145 6.981132 4.331273 7.512316 2.691944 0.0194668 7.358491 4.573542 7.851702 2.883264 0.020729 7.924528 4.990732 8.427634 3.216748 0.0220065 8.301887 5.426425 9.018761 3.569933 0.0142972 7.169811 3.007181 5.5779 1.684124 0.0214458 9.811321 5.233539 8.75829 3.412992 0.0285944 11.98113 7.882979 12.19964 5.633597 0.035743 15.37736 10.93035 15.90195 8.310147 0.0428916 18.67925 14.35712 19.86522 11.41687 0.0500402 22.45283 18.14887 24.08945 14.93387 0.0571888 26.22642 22.2939 28.57464 18.84506 0.0643374 31.79245 26.7825 33.32079 23.13707 0.071486 36.60377 31.60637 38.3279 27.79845 0.0786346 42.35849 36.75834 43.59597 32.81928 0.0857832 46.98113 42.23211 49.125 38.19081 0.0929318 53.86792 48.02208 54.91498 43.90526 0.1000804 58.96226 54.12323 60.96593 49.95559 78 AIf Experimental Predicted by Hukill Predicted by Ergun Predicted by Shedd 31 102 Pressure Drop and lye Equation Equation Equation(m s m ) -1 -1 -1 -1(Pa m ) (Pa m ) (Pa m ) (Pa m ) 0.1072305 65.37736 60.5324 67.27919 56.33683 0.142974 92.16981 97.03817 102.754 92.99144 0.1787175 133.4906 140.7073 144.7529 137.1721 0.214461 188.3962 191.2435 193.276 188.4535 0.2502045 243.3962 248.4201 248.3233 246.5071 0.285948 315.0943 312.0558 309.8948 311.0676 0.3216915 366.9811 382.0003 377.9904 381.9139 0.357435 480.1887 458.1261 452.6103 458.8574 0.3931785 546.2264 540.3231 533.7543 541.7341 0.428922 626.4151 628.4946 621.4225 630.3998 0.4646655 730.1887 722.5548 715.6148 724.7257 0.500409 795.283 822.4265 816.3314 824.5961 0.5361525 929.2453 928.04 923.5721 929.9058 0.571896 1032.075 1039.332 1037.337 1040.559 0.6076395 1155.66 1156.244 1157.626 1156.467 0.643383 1266.981 1278.721 1284.439 1277.548 0.6791265 1430.189 1406.716 1417.777 1403.727 40 30 20 10 -10 -20 -30 -40 + Hukill & Ives Model (Model 2) • Ergun Model (Model 3) 0 0.1 0.2 0.3 0.4 AirVelocity, rn/s A Shedd Model (Model I) 0.5 0.6 0.7 0.8 Figure A.2 Absolute error using three models (Relevant to data in page 78-79) 0 L.. 0 . A + • • A • • • + 79 Appendix B: EXPERIMENTAL SET TIP 80 II Figure B.1 Photo of experimental equipment
- Library Home /
- Search Collections /
- Open Collections /
- Browse Collections /
- UBC Theses and Dissertations /
- Permeability of bulk wood pellets with respect to airflow
Open Collections
UBC Theses and Dissertations
Featured Collection
UBC Theses and Dissertations
Permeability of bulk wood pellets with respect to airflow Yazdanpanah, Fahimeh 2009
pdf
Notice for Google Chrome users:
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.
Page Metadata
Item Metadata
Title | Permeability of bulk wood pellets with respect to airflow |
Creator |
Yazdanpanah, Fahimeh |
Publisher | University of British Columbia |
Date Issued | 2009 |
Description | Data on the resistance of wood pellets to air flow are required for the design and control of ventilation and drying of bulk pellets in storage. In this study, pressure drops versus air flows were measured for several sizes of wood pellets, with diameter of 6 mm and lengths varying from 4 to 34 mm. Air flow rates ranging from 0.014 to 0.80 m s⁻¹ were used in the experiment. The maximum pressure drop measured was 2550 Pa m⁻¹. Three predictive models - Shedd, Hukill-Ives, and Ergun equations that relate pressure drop to air flow in bulk granular materials were used to analyze the data. The Ergun equation was found to provide the best fit to the data. Aeration of bulk pellets in storage requires a low airflow. The airflow range used for low permeability tests was from 0.0002 to 0.0220 m s⁻¹. The corresponding measured pressure drop ranged from 0.18 to 8.30 Pa m⁻¹ for low permeability tests. Three models were investigated and compared for the low permeability data. The increase in moisture content over a wider range of airflows (0.0042 to 0.7 148 m ⁻¹) showed a slight decrease in the resistance to airflow due to increased moisture content. Broken and fines are produced when pellets are handled. The resistance to air flow for wood pellets was measured in the presence of fine materials. Fines were defined as broken pellets passed a 4 mm sieve. The average geometric diameter of the fines was 0.75 mm. The pressure drop for pellets mixed with fines ranged from (2.0 to 191.2 Pa m⁻¹) and (7.9 to 1779.0 Pa m⁻¹) for 1% and 20% fines content (mass basis) respectively. Coefficients of Hukill and Ives’ equation for pellets were estimated as a function of percent fines content. |
Extent | 1971230 bytes |
Genre |
Thesis/Dissertation |
Type |
Text |
FileFormat | application/pdf |
Language | eng |
Date Available | 2009-11-18 |
Provider | Vancouver : University of British Columbia Library |
Rights | Attribution-NonCommercial-NoDerivatives 4.0 International |
DOI | 10.14288/1.0058670 |
URI | http://hdl.handle.net/2429/15231 |
Degree |
Master of Applied Science - MASc |
Program |
Chemical and Biological Engineering |
Affiliation |
Applied Science, Faculty of Chemical and Biological Engineering, Department of |
Degree Grantor | University of British Columbia |
GraduationDate | 2009-11 |
Campus |
UBCV |
Scholarly Level | Graduate |
Rights URI | http://creativecommons.org/licenses/by-nc-nd/4.0/ |
AggregatedSourceRepository | DSpace |
Download
- Media
- 24-ubc_2009_fall_yazdanpanah_fahimeh.pdf [ 1.88MB ]
- Metadata
- JSON: 24-1.0058670.json
- JSON-LD: 24-1.0058670-ld.json
- RDF/XML (Pretty): 24-1.0058670-rdf.xml
- RDF/JSON: 24-1.0058670-rdf.json
- Turtle: 24-1.0058670-turtle.txt
- N-Triples: 24-1.0058670-rdf-ntriples.txt
- Original Record: 24-1.0058670-source.json
- Full Text
- 24-1.0058670-fulltext.txt
- Citation
- 24-1.0058670.ris
Full Text
Cite
Citation Scheme:
Usage Statistics
Share
Embed
Customize your widget with the following options, then copy and paste the code below into the HTML
of your page to embed this item in your website.
<div id="ubcOpenCollectionsWidgetDisplay">
<script id="ubcOpenCollectionsWidget"
src="{[{embed.src}]}"
data-item="{[{embed.item}]}"
data-collection="{[{embed.collection}]}"
data-metadata="{[{embed.showMetadata}]}"
data-width="{[{embed.width}]}"
data-media="{[{embed.selectedMedia}]}"
async >
</script>
</div>
Our image viewer uses the IIIF 2.0 standard.
To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.24.1-0058670/manifest