Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Measurement and modeling of the effect of fines content on the transverse permeability of oriented strand… Fakhri, Hamidreza 2005

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_2005-0436.pdf [ 14.87MB ]
Metadata
JSON: 831-1.0075068.json
JSON-LD: 831-1.0075068-ld.json
RDF/XML (Pretty): 831-1.0075068-rdf.xml
RDF/JSON: 831-1.0075068-rdf.json
Turtle: 831-1.0075068-turtle.txt
N-Triples: 831-1.0075068-rdf-ntriples.txt
Original Record: 831-1.0075068-source.json
Full Text
831-1.0075068-fulltext.txt
Citation
831-1.0075068.ris

Full Text

M E A S U R E M E N T A N D MODELING OF THE EFFECT OF FINES CONTENT ON THE TRANSVERSE PERMEABILITY OF ORIENTED STRAND B O A R D (OSB) by H A M I D R E Z A F A K H R I B. Mechanical Engineering, Sharif University of Technology, Tehran, Iran A THESIS SUBMITTED IN PARTIAL F U L F I L M E N T OF THE REQUIRMENTS FOR THE DEGREE OF M A S T E R OF APPLIED SCIENCE in THE F A C U L T Y OF G R A D U A T E STUDIES ("Wood Science) THE UNIVERSITY OF BRITISH C O L U M B I A May 2005 Hamid Reza Fakhri, 2005 Abstract Most Oriented Strand Board (OSB) panels produced today contain up to 30% fines in the core. Since substantial quantities of fines are generated during the stranding process, there is a strong economic incentive for incorporating as much of this material into finished panels as possible while still meeting product quality standards. The inclusion of fines into the core of an OSB mat affects the permeability of the mat, which in turn may affect the press cycle of the boards. There is some work on the permeability of particleboard and OSB made from only strands or without resin in the literature, but none on resinated mixtures of fines and strands. For this work, OSB furnish was collected from a commercial OSB plant. The species mix in the furnish was approximately 60% aspen, 30% pine, and 10% birch. The strands and fines were separated by screening. A set of 45 boards was made at 3 levels of target density and 5 different strands and fines ratios and transverse permeability of the full thickness, surface, and core layers were measured. Vertical density profile (VDP) and internal bond strength (IB) of the boards were also tested. The results indicate that board density has a major effect on transverse permeability of the full thickness of the OSB panel as well as individual layers within a board (top, bottom, and core). Fines content in the core layer has a statistically significant effect on its permeability. In contrast, there was no statistically significant effect of fines content on permeability of the full thickness samples; this implies that permeability of the core layer was heavily masked by the denser impermeable surface layers. The permeability of the core layer increases with fines content at each panel density level with the rate of increase being higher for the high density panels. Based on these results, a model was developed to predict the permeability of the core of OSB containing a mixture of strands and fines. The model considers the furnish mixture as layers of only strands and only fines arranged in series and parallel configurations. The inputs to the model are the mass fraction of the fines in the layer, the known permeability of the composite comprised of 100% fines and 100% strands .layers as functions of density, and an empirical coefficient ' a ' which expresses the contribution from each ii model. It was found that the upper bound of the permeability data was approximately coincident with the parallel model while the lower bound corresponded with the series model. The permeability data were well described by a rule of mixtures combination of the parallel and series models for a - 0.48. i i i Table of Contents Abstract •••• i i Table of Contents iv List of Tables vii List of Figures ix List of symbols, Nomenclature, and Abbreviations xiv Acknowledgement xvii Chapter 1: Introduction 1 1.1 Rationale 2 1.2 Objectives and structure of thesis chapters 3 Chapter 2: Literature Review 5 2.1 Overview 5 2.2 Background 5 2.3 Permeability 8 2.3.1 Definition 8 2.3.2 Movement of fluids in wood 9 2.3.3 Bulk flow 9 2.3.4 Diffusion 11 2.3.5 Derivation of Darcy's law for permeability of wood 12 2.4 Heat and mass transfer in hot-pressing 14 2.4.1 Heat transfer mechanism 14 2.4.2 Mass transfer mechanism 16 2.5 Mat deformation during pressing 18 2.6 Factors affecting permeability of solid wood 19 2.7 Factors affecting permeability wood composites 21 2.8 Summary 28 Chapter 3: Pilot Study of Manufacturing OSB Panels with Various Fines Content 29 3.1 Introduction 29 3.2 Experimental design 30 3.3 Materials and methods 31 3.3.1 Materials 31 3.3.2 Board preparation 32 3.3.3 Resination 33 3.3.4 Mat formation 34 3.3.5 Hot-pressing 35 3.3.6 Sample preparation 36 3.3.7 Measurement of permeability 38 3.4 Results and discussion 40 3.4.1 Transverse permeability 40 3.4.2 Vertical density profile 42 3.5 Conclusions 43 Chapter 4: Effect of Fines Content on Transverse Permeability and Internal Bond Strength 45 4.1 Introduction 45 4.2 Experimental design 46 4.3 Materials and methods 46 4.3.1 Furnish 46 4.3.2 Furnish characteristics 49 4.3.3 Resination and Mat formation 51 4.3.4 Sample preparation 52 4.3.5 Appearance of boards and their core layers 52 4.3.6 Measurement of internal bond (IB) strength 53 4.3.7 Statistical analysis 54 4.4 Results and discussion 55 4.4.1 Vertical density profile 55 4.4.2 Internal bond strength 57 4.4.3 Comparison of measured permeability values with previous works 58 4.4.4 Comparison of permeability of full thickness, faces, and core of OSB 59 4.4.5 Effect of density and fines content on permeability 61 4.5 Conclusions 64 ChapterS: Permeability Model 66 5.1 Introduction 66 5.2 Model development 66 5.2.1 Parallel model development 68 5.2.2 Series model development 70 5.3 Results and discussion 73 5.3.1 Description of Kf and Ks as functions of density, p 73 5.3.2 Contribution of series and parallel models to system permeability 76 5.3.3 Determination of the series coefficient, a 77 5.3.4 Sensitivity analysis of the models 79 5.4 Conclusions 84 Chapter 6: Summary, Recommendations and Future Work 85 6.1 Summary 85 6.2 Recommendation and Future Work 87 References 88 Appendix A: Preliminary Observations of Transverse Permeability of Faces and Core Layers of Commercial Panels (OSB, PB, and MDF) 93 A.l Introduction 93 A.2 Materials and methods 93 A.2.1 Materials and sample preparation 93 A.2.2 Measurement of transverse permeability 96 A.2.3 Measurement of vertical density profile 97 A.3 Results and discussion 98 v A.3.1 Vertical density profile 98 A.3.2 Relationship between permeability and layer density 99 A.3.3 Effects of board type and layer position on permeability and density 99 A.4 Conclusions 101 Appendix B: Press Cycles and PressMan Output for Each Board 102 Appendix C: Length, Width, and Thickness Classifications of Strands and Fines .107 Appendix D: Board Constituents 109 Appendix E: Summary of Analysis of Variance and Raw Data 110 List of Tables Table 2.1: Mat internal conditions and physical properties that are needed for modeling 7 Table 3.1: Factors and responses of the pilot experiment 31 Table 3.2: Frequency, mass, mass fraction, and mean lengths and widths of the strands classified by length 31 Table 3.3: The treatment combination for density and fines contents of the nine boards made in the pilot study. Samples were only cut from those boards whose label appears in italics 33 Table 3.4: Constituent mass required to produce one board 34 Table 4.1: Variable factors and experiment responses for the laboratory made OSB 46 Table 4.2: Average length, width, and thickness of strands for each replication 50 Table 4.3: Overall core density and IB strength for low, medium, and high density 58 Table 4.4: Mean density of full thickness samples and top, bottom, and core layers 60 Table 4.5: Mean permeability of full thickness samples and top, bottom, and core layers. 60 Table 4.6: p-values for density, fines, and their interaction on permeability of the top, bottom, core layers, and full thickness samples 61 Table 5.1: Comparison of the permeabilities of the two models at the densities of 500 and 600 kg/m 3 and their ratios at the indicated fines content 76 Table 5.2: Comparison of the sum of square errors of the predicted permeabilities of the system for different a values 78 Table 5.3: Comparison of the variation on permeability models caused by ±15% changes in each parameter at 30% fines content and density of 520 kg/iri 83 Table B . l : Press schedule for high density board (target thickness 19 mm) 102 Table B.2: Press schedule for medium density board (target thickness 22 mm) 102 Table B.3: Press schedule for low density board (target thickness 27 mm) 103 Table C . l : Weight, average length, average width, and average thickness of the strands of the first replicate among approximately 390 strands 107 Table C.2: Weight, average length, average width, and average thickness of the strands of the second replicate among approximately 880 strands 107 vn Table C.3: Weight, average length, average width, and average thickness of the strands of the third replicate among approximately 1100 strands 107 Table C.4: Weight, average length, average width, and average thickness of the fines for the first replication among over 5000 fines 108 Table D . l : Constituent mass required to produce one board of the comprehensive experiment 109 Table E . l : Summary of two-way A N O V A for the permeability of full thickness samples 110 Table E.l: Summary of two-way A N O V A for the permeability of core layers 110 Table E.3: Summary of two-way A N O V A for Internal bond (D3) 111 Table E.4: Probability values of the t-test at the 0.05 level, applied to permeability and density of top and bottom layers for each density level I l l Table E.5: Raw data for full thickness board 112 Table E.6: Raw data for top layer 113 Table E.7: Raw data for bottom layer 115 Table E.8: Raw data for core layer 117 List of Figures Figure 2.1: Modeling approach for hot-pressing, where the mat, press, and press controller are simulated (adapted from Hubert and Dai, 1998) 6 Figure 2.2: A schematic diagram of the interactions between various processes during the hot-pressing of wood composites (adapted from Bolton and Humphrey, 1988) 7 Figure 2.3: A Permeable specimen and Darcy's law parameters (adapted from Siau, 1995) 12 Figure 2.4: A typical mat pressure as a function of time during hot-pressing. (adapted from Dai et al., 2000) 19 Figure 2.5: Effect of board density on superficial gas permeability of boards (adapted from Hata, 1993) , 22 Figure 2.6: Specific permeability as a function of M D F panel density adapted from Garcia and Cloutier (2005) 23 Figure 2.7: Permeability of longitudinal direction, K,, and transverse direction, K,, as a function of mat density (adapted from Carvalho and Costa, 1998) 24 Figure 2.8: Effect of particle length on superficial gas permeability of boards (adapted from Hata, 1993) 25 Figure 2.9: Effect of particle thickness on superficial gas permeability of boards (adapted from Hata, 1993) 25 Figure 2.10: Cross-sectional area of a laboratory made OSB panel with an average density of 520 kg/m3, scanned at Wood Composites lab, The University of British Columbia 26 Figure 2.11: Surfaces of boards made by steam-injection pressing at low and high density from Aspen (from Bolton and Humphrey, 1994) 28 Figure 3.1: Frequency distribution of strands used in pilot study 32 Figure 3.2: Fines used for the pilot study: (a) appearance of a bulk sample of fines, and (b) the weight portion of fines falling in a length classification 32 Figure 3.3: The 150 cm (5 ft) diameter blender used to resinate the strands and fines separately. The blender has the depth of 61 cm (2 ft), and rotated in counter-clockwise direction at 34 R P M . The front cover removed to demonstrate tumbling trajectory of strands 34 Figure 3.4: Schematic of mat formation in 3 layers. Core layer contains the mixture of fines and strands, and face layers are aligned unidirectional. L is the thickness of the board (dimensions in mm) 35 ix Figure 3.5: Steps of mat formation: (a) unidirectional strands were laid up in bottom layer by an orienter, (b) randomly oriented combined strands and fines in core layer, (c) the entire mat after being formed 35 Figure 3.6: The (a) Wabash press was used to press the mats in the Wood Composite Lab of Wood Science Department at U B C and (b) board just after being pressed and laid on the floor of the lab 36 Figure 3.7: Cutting pattern of the samples for measuring permeability and VDP tests which were selected randomly (dimensions are in mm) 37 Figure 3.8: Preparation of test specimens: (a) initial board specimens, the nine pieces of 65mm x 65mm, (b) sanding specimens to required thickness, and (c) cutting cylindrical permeability samples 37 Figure 3.9: Schematic diagram of gas permeability apparatus by the falling water displacement method (adapted from Siau, 1995). The apparatus was set up in the Wood Physics Laboratory in the Wood Science Department at the University of British Columbia 38 Figure 3.10: Relationship between permeability and density: (a) full thickness board samples, (b) face, and (c) core layers 40 Figure 3.11: Comparison of the effect of fines on: (a) mean density versus fines content for different density levels in core layer and (b) relationship between specific permeability and levels of fines content for different density levels in core layer 41 Figure 3.12: Vertical density profile, low density: (a) 0% and (b) 50%, medium density: (c) 0%, (d) 50%, and (e) 100%, high density: (j) 50% and (g) 100%). The position of each sample on the board is shown in parentheses 43 Figure 4.1: Schematic diagram of the strands flow from the strander to the mat former at OSB plant. Note that the strands were collected between the dryer and rotary screener 47 Figure 4.2: Mixture of strands and fines falling from silo 48 Figure 4.3: Mechanical shaker table for screening furnish into 4 size classes in the Composite lab of Wood Science Department at U B C 48 Figure 4.4: Schematic diagram of size classes of screened furnish 49 Figure 4.5: Comparison of: (a) frequency distribution and (b) mass fraction of strands in each length class for the 3 replicates 50 Figure 4.6: Appearance of a fines class showing: (a) a bulk sample of fines and (b) the length range of the fines. Note that the magnification of both images is the same 51 Figure 4.7: Comparison of the averages of: (a) frequency distribution and (b) mass fraction in each length class for fines and strands 51 Figure 4.8: The 15 boards of the replicate 2< (dimensions in mm) 53 x Figure 4.9: Comparison of typical appearance surfaces of the core layers of high density boards with increasing fines contents 53 Figure 4.10: Testing IB samples on the Sintech 30D load frame: (a) test set-up, and (b) a close-up of a sample just after failure 54 Figure 4.11: Comparison of vertical density profiles from two different locations within a board for each fines content (y-axis) and density level (x-axis). VDPs are for the boards of first replicate and the position of each sample on the board is shown in parentheses 56 Figure 4.12: IB as a function of density for all density levels 58 Figure 4.13: The permeability of full thickness, top, core, and bottom samples. The total number of the samples for each category is shown on the figure 60 Figure 4.14: Comparison of permeability as a function of density for all fines contents: (a) full thickness sample and (b) core layer 62 Figure 4.15: Comparison of the permeability of the core layer as a function of fines content for all density levels: (a) log scale and (b) linear scale; note that the LSD bar is asymmetrical on a linear scale and is therefore not shown 63 Figure 4.16: Schematic structure of the flakes in an OSB board: (a) random position of flakes and the void spaces in between (b) layers of the OSB mat consisting of strands and fines (adapted from Dai et al., 1997) 64 Figure 5.1: Cross-section of OSB panel: (a) mixture of strands and fines in the core layer and (b) concentration of fines and strands oriented in different positions 67 Figure 5.2: Schematic of the models: (a) mixture of strands and fines in the core layer of OBS panel, (b) a complex of 100% fines and 100% strands is a parallel model, and (c) a complex of 100% fines and 100% strands is a series model , 68 Figure 5.3: Schematic diagram of the parallel model consisting two layers; one of 100% fines and the other of 100% strands 68 Figure 5.4: Schematic diagram of the series model consisting two layers; one of 100% fines and the other of 100% strands .'. 71 Figure 5.5: Permeability as a function of density and best fit curves for: (a) 100% strands layer and (b) 100% fines layer 74 Figure 5.6: Model results for permeability as a function of density for 0, 25, 50, 75, and 100% fine contents: (a) parallel sub-model, and (b) series sub-model ; 75 xi Figure 5.7: Comparison of the predicted permeability of parallel and series models for 0-100% fines contents at the densities of: (a) 500 and (b) 600 kg/m 3 76 Figure 5.8: Upper and lower bounds produced by parallel and series model for: (a) 25%, (b) 50%, and (c) 75% fines contents in the core layer 77 Figure 5.9: Curve fit to the permeability data using a rule of mixtures approach, combined parallel and series models using an a of 0.48 for: (a) 25%, (b) 50%, and (c) 75% fines contents 79 Figure 5.10: Comparison of the predicted permeability of the system model (left column), series model (center column), parallel model (right column) for ±15% changes in, by row from top to bottom, Mf, Ks, Kf, pf I ps, and a , respectively 81 Figure 5.11: Comparison of the predicted permeability variation of the system model (left column), series model (center column), parallel model (right column) for ±15% changes in, by row from top to bottom, Mf, Ks, Kf, pf I ps, and a , respectively 82 Figure A . l : Location of the three sub-panels in each commercial panel. Dimensions are in mm 94 Figure A.2: Sanding of sub-panels to leave top (SI), core (S2), and bottom (S3) layers. The gray areas were sanded away with a wide belt sander, and all dimensions are in mm 94 Figure A.3: Cutting templates for permeability samples (circles), and VDP samples (square); dimensions in mm 95 Figure A.4: Preparation of permeability specimens showing: (a) sub-panels after sanding, (b) specimen cutting using a hole saw, and (c) permeability specimens 95 Figure A.5: Measurement of permeability: (a) apparatus for measuring transverse permeability, (b) cylindrical holder for placing samples 96 Figure A.6: Schematic diagram of permeability apparatus; O M , open mercury manometer to environment for measuring the inlet pressure upstream; M M , differential mercury manometer; F M , air flow rate meter (rotometer); TS, test specimen; D, desiccant; PR, pressure regulator; N V , needle valve; TWV, three-way valve, and R1-R4 are the readings from manometers in mmHg, and R5 is the reading from rotometer 96 Figure A.7: The QMS vertical density profiler 98 Figure A.8: Typical Vertical density profiles of tested %" commercial panels 98 Figure A.9: Relationship between transverse permeability and density for top, core, and bottom layers of commercial panels: (a) OSB, (b) PB, and .(c) M D F 99 Figure A. 10: Interaction between board type and layer position for (a) density, and (b) permeability 100 Figure B . l : Mat pressure, temperature, and gas pressure as a function of time for replicate 1. Columns show boards with different density levels (left to right) and rows indicate fines content (top to bottom) 104 Figure B.2: Mat pressure, temperature, and gas pressure as a function of time for replicate 2. Columns show boards with different density levels (left to right) and rows indicate fines content (top to bottom) 105 Figure B.3: Mat pressure, temperature, and gas pressure as a function of time for replicate 3. Columns show boards with different density levels (left to right) and rows indicate fines content (top to bottom) 106 xiii List of Symbols, Nomenclature, and Abbreviations Symbols Description Units A cross sectional area of the specimen perpendicular to flow direction m A f cross sectional area of core layer consisting 100% fines m 2 cross section area of 100% strands m 2 c correction factor -cM molecules concentration mole/m Dc diffusion coefficient by molecules concentration kg • m /mole/s diffusion coefficient by partial vapor pressure kg/m/s/Pa Dm mat gas diffusivity kg/m/s/Pa generated energy in control volume J/m3/s energy input into the control volume J/m3/s energy output from the control volume J/m3/s K energy accumulation within in the control volume J/m3/s j total mass flux kg/m2/s vapor mass flux kg/m2/s K specific permeability m /m effective permeability m 3/m K parallel permeability of parallel model m3/m Kpred predicted permeability m3/m If series permeability of series model m /m system permeability of the system m3/m ]/• ^500 permeability at the density of 500 kg/m3 m 3/m ^600 permeability at the density of 600 kg/m3 m 3/m L length of the specimen in direction of flow mm Mf fines content % N number of observations -P total pressure Pa Palm atmospheric pressure mHg P average pressure Pa inlet pressure upstream of sample Pa Pi outlet pressure downstream of sample Pa Pi pressure of flow at interface Pa Pv vapor pressure Pa Q flow rate m3/s Q, total flow of fluid m3/s T absolute temperature ' ' ' K xiv Symbols Description Units Tc temperature °C V volume m 3 a, b, c constants ai,bj,ci regression coefficients cm mat specific heat J/kg/°C c v steam specific heat J/kg/°C k permeability m 3 (liquid)/m/s/Pa kg superficial gas permeability m 3 (gas)/m/Pa/s km mat apparent gas permeability s kz thermal conductivity of the mat in the thickness direction J/s/°C/m m mass kg ms mass of core layer composed of 100% strands kg mf mass of core layer composed of 100% fines kg n number of moles of gas pa air pressure Pa pv partial pressure of steam Pa q heat flux J/s/m3 r radius m t time s v velocity of steam m/s vd void fraction in the mat x distance in Cartesian coordinate axis m z distance in Cartesian coordinate axis m a series coefficient Af width of core layer consisting 100% fines m Xs width of core layer consisting 100% strands m p density kg/m3 pf density of core layer composed of 100% fines kg/m3 pm mat density kg/m3 ps density of core layer composed of 100% strands kg/m pv steam density kg/m3 <j) designated value of each M f, Ks, Kf, pf I ps,or a p dynamic viscosity of the fluid Pa s pa viscosity of air Pa • s p viscosity of the gas phase, air and vapor Pa • s xv Symbols Description Units FSP fiber saturation point HD high density HDD horizontal density distribution IB internal bond strength L D low density LSD least significant difference M C M million cubic meters M D medium density M D F medium density fiber board M O E module of elasticity M O R module of rupture OSB oriented strand board PB particleboard VDP vertical density profile atm atmosphere ns not applicable ns not significant od oven dried s significant wt weight Subscripts a air / fines g gas phase / longitudinal m mat s strands t transverse v steam x Cartesian coordinate axis z Cartesian coordinate axis Constants R universal gas constant, 8.31 J/mol/K Ha viscosity of air, 1.846x 10"5 Pa • s xvi Acknowledgements I would like to acknowledge number of individuals whose help and guidance throughout completing this thesis was remarkable and supportive: Dr. Gregory D. Smith my direct supervisor and also my committee members: Drs. Ian D. Hartley and Phil Evans. I would also like to thank Dr. Kate Semple, Emmanuel Sackey, Bob Myronuk, and Stephane Gautier for their assistance in the completion of this work. A special thanks to Ainsworth Lumber Co. Ltd, for providing the wood strands, to Borden Chemical Inc. for providing the required PF resin, to Dr. Stavros Avramidis for use of the permeability apparatus in the Wood Physics lab, and to Forintek Canada Corp. (Western laboratory) for the use of their facilities and equipment. xvn CHAPTER 1 Introduction Wood-based panel products can be categorized into two major groups: structural panels such as plywood and oriented strand board (OSB) that are typically used for sheathing, and non-structural panels such as medium density fiberboard (MDF) and particleboard (PB) that are mainly used in furniture and cabinets. One of the key differences between these panel products is the geometry of the furnish: plywood and laminated veneer lumber (LVL) are composed of veneer sheets, OSB uses strands, PB is composed of small flakes and slivers, and M D F is made of wood fibers. OSB is sold as a structural panel for the residential construction market (RISI, 2002). Production capacity of OSB in North America is expected to increase from 21 million cubic meters (MCM) in 2001 to about 27.9 M C M 2006 (Wood Market, 2002). The worldwide capacity of OSB is predicted to reach almost 34.5 M C M in 2006, which is more than double that produced in 1997 (Wood Market, 2002). By 2006, OSB is expected to account for 63% of the North America's structural panel consumption market share (Wood Market, 2002). Strong competition is forcing OSB producers to improve plant efficiency and profitability in order to lower production costs, and there is a need to create specialty OSB products for new markets. Two key components of the OSB manufacturing process are the production of wood elements and the hot-pressing of an unconsolidated mat into a solid panel. The addition of undersized strands, commonly referred to as fines, into the core of the OSB mat is an important area for research since it could benefit the economics of production by utilizing a waste stream, and also may increase the mat permeability, thus reducing press cycle times. Fines are small wood elements that are produced during the stranding process. They may consist of small wood slivers, cubic chunks, and occasionally pieces of bark. 1 Spelter et al. (1996) noted that fines generation during stranding is affected by many factors such as knife speed, log condition, species, and log size. Fines generation rises exponentially with knife speeds above a certain level. Frozen logs are more brittle and the wood breaks more easily and produces more fines during stranding than warm logs. Small diameter logs generally produce more fines than larger logs (Spelter et al., 1996). Although, strands are usually screened to remove fines, the mesh size of these screens is not standardized and therefore the size of fines varies from plant to plant. Fines are also considered an inferior component of the furnish and plants usually add this material into the core of the panels in order to reduce their adverse effect on mechanical properties. It is common for plants to add as much as 30% of fines into the core furnish of a panel. Therefore, it is important to characterize their effect on panel properties. The reported research shows that wood element size strongly affects the mechanical properties of PB and OSB. Moreover, studies by Turner (1954) and Brumbaugh (1960) on the effect of flake size on the properties of particleboard show that the bending properties (MOE and MOR) increase as the length of flakes increases for a constant flake thickness. The work by Barnes (2001) suggests that such an increase in mechanical properties is the result of the increase in the length of the glue line between the wood elements. Based on this, one would expect that incorporating fines into OSB would reduce its strength due to the shorter glue lines between adjacent wood elements. Fines may have a larger effect, since it is well known that smaller wood elements have larger surface area and tend to absorb a larger portion of the resin during blending (Maloney, 1993). Cafferata (2003), however, found that it was possible to increase the use of fines while retaining acceptable strength properties. 1.1 Rationale The increased use of fines in OSB has the potential to significantly reduce production cost. Based on a plant with an annual production of 0.35 M C M (400xl0 6 ft2/year, 3/8" basis) and a wood cost of $US 55/m3 (Spelter et al., 1996 and 1997), it is estimated that 2 approximately one million US dollars per year could be saved in furnish costs by adding 5% fines into the OSB panel. Previous studies (Hata, 1993; Bolton and Humphrey, 1994; D'Onofrio, 1994; and Hood, 2004) have found that flake size in particleboard and OSB strongly affects the permeability of mats and the finished composites. Increasing the fines content in OSB may also increase mat gas permeability which could lead to decreased pressing times. In this thesis the effect of adding fines to the core of OSB on its transverse permeability is examined. In order to understand how fines affect gas flow through the mat, OSB boards with various fines contents and densities were manufactured and the permeabilities of full thickness samples, both faces and core layers were measured. There have been a few reports on the characterization of strand size in OSB. However, no published work has addressed the issue of the permeability of mats made of mixed fines and strands during mat consolidation. The work presented in this thesis is the first on this topic and has of the following structure. 1.2 Objectives and structure of chapter thesis 1. Literature review on permeability in wood-based composites and the factors affecting it such as wood element size and void spaces within the mat during hot-pressing (Chapter 2). 2. Development of experimental procedures for manufacturing OSB boards consisting of mixtures of fines and strands. Identification of the minimum sample densities containing various fines content that have sufficient integrity that samples may be cut from and then their permeability measured (Chapter 3). 3. Evaluation of the effect of core fines content on transverse permeability of full thickness samples and the top, bottom, and core layers of partially consolidated OSB mats; examination of the effect of fines content on the IB strength of partially 3 densified boards and whether the VDP of these boards are comparable to commercial OSB (Chapter 4). Development of a permeability model of core layer based on the Rule of Mixtures for the transverse permeability composed of two hypothetical parallel and series layers consisting of 100% fines and 100% strands (Chapter 5). CHAPTER 2 Literature Review 2.1 Overview This literature review covers the hot-pressing process of wood-based composite mats and examines why modeling of various phenomena during hot-pressing has been of interest to researchers over the last two decades. In this chapter, the mechanisms of fluid movement in wood and wood composites and factors such as flake size are summarized. The concepts of permeability and its theory as one of the transport properties of a porous medium are reviewed. 2.2 Background The hot-pressing of wood-based composites panels is one of the critical processes in panel production. A loosely formed mat is compressed to its target thickness under high pressure and temperature. As the press closes and the surfaces of the mat come into contact with the hot press platens, the temperature increases and absorbed water in the furnish begins to evaporate. The resulting steam migrates to the core of the mat and condenses on the interior strands, transferring the latent heat of the steam to the strands and raising their temperature. As the core temperature of the mat gradually increases, the moisture in the core starts to evaporate again. The internal gas pressure increases and the steam moves towards the edges of the mat (Kamke and Casey, 1988). The rate of the moisture and heat transfer strongly depends on the structure of the mat, and how this structure changes during hot-pressing. During consolidation, there is a complex set of interactions that occur between the wood flakes, air, and water, which influences the rate of heat and mass transfer, the resin cure kinetics, and the visco-elastic deformation of the wood (Bolton and Humphrey, 1988; Hubert and Dai, 1998). Since the volumes throughput of an OSB plant is usually restricted by the press, reducing the press time is a key issue for wood composite plants. There have been various attempts at modeling and predicting hot-pressing phenomena in order to optimize press cycle time. 5 Published models have generally been one or two-dimensional (Harless et al. 1987; Hubert and Dai, 1998; Zombori, 2001), but more recently three-dimensional heat and mass transfer models (Garcia, 2002; Dai and Yu, 2004) have been published. These studies have attempted to develop accurate models for heat and mass transfer through the mat thickness as well as in-plane directions. The hot-pressing mechanisms are similar for PB, MDF, and OSB, and a hot-pressing model for one type can usually be modified and applied to other wood-based composites. The key components of an integrated hot-pressing model are shown schematically in Figure 2.1. Mat properties Material properties Mat Heating element -Press specification Press O O Q O O O O O O Press Press cycle z Controller Figure 2.1: Approach for hot-pressing model where the mat, press, and press controller are simulated (adapted from Hubert and Dai, 1998). In general, the objectives of a model are to predict the dynamics of densification and heat and moisture transfer and their effect on panel density, bonding strength, and panel performance (Garcia, 2002). For modeling the hot-pressing processes, knowledge of the mat physical properties and the internal conditions, such as those listed in Table 2.1, is required. An integrated model attempts to capture the important interactions between various processes during the hot-pressing of wood composites; the key interactions are shown schematically in Figure 2.2 (Bolton and Humphrey, 1988). 6 Table 2.1: Mat internal conditions and physical properties that are needed for modeling. Mat internal conditions: Mat physical properties: 1. Vapor pressure 1. Structure 2. Phase changes 2. Permeability 3. Gas flow 3. Heat capacity 4. Heat conduction 4. Thermal conductivity 5. Heat convection 5. Wood element rheology Figure 2.2: A schematic diagram of the interactions between various processes during the hot-pressing of wood composites (adapted from Bolton and Humphrey, 1988). The layers of a composite mat during hot-pressing are subjected to a non-linear stress-strain compression process (Dai and Steiner 1993; Lang and Wolcott 1996, and Dai, 2001). Generally, the hot-pressing of wood composites involves simultaneous heat, mass, and momentum transfers that are affected by mat properties such as thermal conductivity, specific heat, compression ratio, and permeability. Chemical reactions and adhesive polymerization also take place during hot-pressing and contribute to the energy input into the system. \ 7 Humphrey and Bolton (1989) presented a heat and mass transfer model for particleboard based on a modified two-dimensional finite difference approach. It predicted transverse and in-plane vapor flow and heat conduction, using cylindrical coordinates. The model consisted of: 1. Describing the heat conduction in the different directions of the mat heated from faces. 2. Predicting the distribution of water within the mat during hot-pressing as it undergoes various phase changes. 3. Predicting water vapor pressure within the mat and the vapor loss from the mat edges due to moisture transfer. Dai and Y u (2004) noted that more work is still needed to establish an accurate model for heat and mass transfer in wood-based composites during the hot-pressing process. Specifically, to understand and describe the interactions between mechanical pressure on the mat, heat and mass transfer, and chemical processes adequately during hot-pressing. In order to do so, more knowledge is required on the permeability and heat conductivity of the composite mat during hot pressing. 2.3 Permeability 2.3.1 Definition Permeability is a measure of how easily a fluid can move through a porous medium such as sand, soil, wood, or wood-based composites due to a pressure gradient. In wood, permeability is a function of the availability of interconnecting pits between cells and perforated plates in vessels (Siau, 1995) or in wood composites connectivity of the pathways between the void spaces (Bolton and Humphrey, 1994). In order to distinguish porosity and permeability, it should be noted that porosity is the measure of void space in the material, whereas permeability is a measure of the fluid flow through those voids. Not all porous bodies are permeable, but a solid material must be porous in order to be permeable. For solid wood, knowledge of permeability permits drying processes to be better controlled and the treatment of wood with preservatives to be more easily predicted (Siau, 1984). For wood-based composites, the permeability of the mat is important 8 because it controls the rate at which steam moves through the mat and its water vapor escapes from the edges of the mat during hot-pressing and thus effects the length of the press cycle (Zombori, 2001). The permeability of a composite panel through the thickness is referred to as the transverse permeability, Kt, while that in the plane of the panel is referred to as the longitudinal permeability, K,. The transverse permeability determines the rate of flow of heat and moisture from the press platens toward the mat core, and thus the rate at which cell walls soften and the resin cures (Bolton and Humphrey 1994). The longitudinal permeability determines the rate of the flow gas from the center of the panel towards the edges or the degas time. 2.3.2 Movement of fluids in wood The movement of liquids and gases through solid wood under steady state conditions occurs by two principal modes: bulk flow and diffusion (Siau, 1995). The relevant equations for describing these phenomena are Darcy's law and Fick's first law both of which can be expressed in their general form as, flux = coefficient x gradient (2.1) The term flux is the rate of volume or mass of fluid movement per unit of cross-sectional area perpendicular to the flow. In Darcy's law, the term gradient is considered as total pressure difference between the two ends (inlet and outlet) of the specimen and causes the flow; the term coefficient is referred to as the permeability. In Fick's first law, there are several alternative expression for gradient and the corresponding coefficient. These alternative gradients are noted as concentration of molecules (e.g. moisture content), partial vapor pressure, etc (Skaar, 1988). 2.3.3 Bulk flow The types of bulk flow that can occur in a porous medium fall into four categories: (a) laminar, (b) turbulent, (c) slip flow, and (d) non-linear flow (Dinwoodie, 1989; Siau, 1995). The differences between these flow types are discussed below: " " 9 a. Laminar or viscous flow: Viscous flow requires the application of a force that causes a smooth, streamlined flow with no turbulence to overcome the fluid's internal friction. The flow rate is directly proportional to pressure gradient and is described by Darcy's law. For example, the flow of fluids through wood cell lumens usually takes place as laminar flow (Dinwoodie, 1989). Humphrey and Bolton (1989) noted that in wood composites, laminar flow is the major flow mechanism through the mat in conventional hot-pressing. b. Turbulent Flow: As the flow velocity is increased, laminar flow is disrupted and the movement of the molecules in the fluid becomes irregular or turbulent and Darcy's law is no longer applicable because the flow rate is no longer directly proportional to the pressure gradient (Siau, 1995). The flow rate is approximately proportional to the square root of the pressure differential. It results in a significant increase in the energy required to transfer a given quantity of fluid compared with laminar flow. The Reynolds' number, defined as the ratio of inertial force to viscous force, is typically used to determine whether flow is laminar or turbulent. For values of the Reynolds' numbers over 2000 the flow is considered turbulent and for those below, flow is laminar. Siau (1995) points out that it is very difficult to achieve turbulent flow in wood due to small capillary size; it may only occur in the largest earlywood vessels of red oak (Quercus rubra). An example of a pressing process that may have predominantly turbulent flow between particles is steam injection pressing where pressurized steam is forced into the mat (Zombori, 2001). c. Slip Flow: Molecular slip flow, also known as Knudsen diffusion, occurs along with bulk flow of gases through very small capillaries in wood when the capillary dimensions are smaller than the mean free path of gas molecules. The mean free path is defined as the average distance a molecule travels between intermolecular collisions (Siau, 1995). 10 d. Non-linear flow: Non-linear flow occurs when kinetic energy is lost at the entrances of very small and short capillary openings, resulting in a larger energy requirement for the same rate of flow (Siau, 1995). As an example, Siau (1995) noted that non-linear flow in wood can occur at Reynolds' numbers between 1 and 16 where fluids enter pit openings. 2.3.4 Diffusion The predominant mechanism of water vapor movement through wood with a moisture content below fiber saturation point (FSP) is diffusion (Dinwoodie, 1989). By definition, the fiber saturation point is defined as the moisture content where the lumens are filled with saturated vapor and the cell wall is fully saturated with bound water only, and there is no liquid water within lumens. Fick's first law expresses diffusion as a flow of molecules from a region of high to low concentration with no necessity for a static pressure gradient. Fick's first law is given as: j = - D c ^ L (2.2) dx where: J = mass flux (kg/m2/s), CM = molecules concentration (moles/m3), Dc = diffusion coefficient by molecules concentration (kg • m2/mole/s), and x = distance in Cartesian coordinates (m). Fick's first law can be expressed in terms of partial pressure gradient as follows, •r = -DP^r (2-3) dx where, Dp is the diffusion coefficient by partial vapor pressure, Pv, in units of (kg/m/s/Pa). The two kinds of diffusion existing through cells below the fiber saturation are defined as bound-water diffusion and inter-gas diffusion. Bound-water diffusion occurs within the 11 cell walls of wood due to chemical potential as driving force (Zomboro, 2001). Inter-gas diffusion consists of the transfer of water vapor or gases through the lumens and interconnecting pit openings due to difference in partial vapor pressure of each gas (Siau, 1995). 2.3.5 Derivation of Darcy's law for permeability of wood The rate of the steady-state flow of fluids caused under a constant pressure gradient can be described by Darcy's law, as follows: Permeability = (2.4) where, Q is the volumetric flow rate of the fluid through the cross-sectional area of the specimen (m3/s) and AP the pressure gradient between the inlet and outlet ends of the specimen causing flow (Pa), shown schematically in Figure 2.3. Figure 2.3: Wood specimen and Darcy's law parameters (adapted from Siau, 1995). The following assumptions are typically used when applying Darcy's law to wood (Siau, 1995): 1. Flow is viscous and linear; therefore the volumetric flow rate and velocity are directly proportional to the pressure gradient. 2. The fluid is incompressible and homogeneous. 3. The porous medium is homogeneous. 4. There is no interaction between fluid and the medium. 5. Permeability is independent of the length of the specimen in the flow direction. L 12 Although some of these assumptions are not strictly true, when Darcy's law is applied to wood the basic equation provides a useful relationship between the flow rate and the pressure gradient to compute the permeability of the wood material (Siau, 1984), i.e., A L where: k = permeability (m 3 (liquid)/m/s/Pa) L - length of the specimen in direction of flow (m) A = cross-sectional area of the specimen perpendicular to flow direction (m2) AP =PX-P2 Px = inlet pressure upstream of sample (Pa) P2 = outlet pressure downstream of sample (Pa) Since the gas pressure changes continuously throughout the specimen, it is more convenient to write Darcy's law for gases in the derivative form when it is applied to gaseous flow as follows, A-dL where, kg is the superficial gas permeability (m3(gas)/m/s/Pa). The ideal gas law can be used to describe the relationship between the expansion of the gas due to decreasing pressure along the direction of flow, i.e., PV = nRT (2.7) where, n, R, and T are the number of moles of gas, universal gas constant (8.31 J/mol/K), and absolute temperature (K), respectively. Substituting Equation 2.7 into the Equation 2.6 and integrating over PandL produces the following expression: 13 (28) s AAPP where, P is the pressure at which volume is measured, AP is given by Pl -P2, and P , average pressure, is defined as: (P2 + P,) / 2. Specific permeability is the product of permeability and the viscosity of the fluid. Thus, specific permeability is independent of the fluid used and is a function of the porous structure of the medium only, i.e., K = k-fi (2.9) where, K specific permeability, (m /m), and // dynamic viscosity of the fluid, (Pa • s). The specific permeabilities for fluids can be derived from Darcy's law by substituting Equation 2.9 into Equation 2.5 to give: K = tLQ± (2.10) A-AP 2.4 Heat and mass transfer in hot-pressing 2.4.1 Heat transfer mechanism As mentioned previously, the permeability of wood composites is heavily influenced by the mat structure, wood element shape, size, degree of alignment, and density. A l l of these factors in turn affect the heat and mass transfer mechanisms in a panel. The main sources of heat are the platens, but the polymerization of the resin and the heat generated internally during mat compaction are small can be neglected for modeling purpose (Zombori, 2001). The mechanisms of heat and mass transfer during hot-pressing can be broken down into the following categories (Kamke, 2004): 1. Conduction from platen to mat surface 2. Conduction within the mat 3. Convection between the gas and particles in the mat 4. Convection at the mat edges 14 5. Bulk flow of gas within the mat 6. Bulk flow of gas out of the mat at the boundaries For wood composites, the main mode of heat transfer after the press platens come into contact with the mat is conduction across the interface between platen and mat faces. At these temperatures, heat transfer by radiation is negligible. The contribution of conduction to the heat transfer within the mat structure becomes larger as the mat is compressed to the target thickness due to increased contact area between adjacent strands. The conductive heat flow is described by Fourier's first law: dz where: qz = heat flux in the thickness direction (J/s/m2) Tc = temperature (°C) kz - thermal conductivity of the mat in the thickness direction (J/s/°C/m) z = distance in Cartesian coordinates (m) The rate of the conductive heat transfer for a given temperature differential is determined by the thermal conductivity of the mat, kz, which is a function of the mat structure, i.e. local flake density, flake orientation, and void fraction and the internal environment, i.e., local partial pressures of all gases (Zombori, 2001). Mat thermal conductivity in the in-plane directions can be obtained by using the same approach. Hubert and Dai (1999) described their heat transfer model with a discretized mat containing wood strands as a 1-D linear model of elements through the mat thickness (z -direction). It is assumed that the gradients in the in-plane directions (the x-y plane) compared to z -direction are small and negligible. It is also assumed that a significant amount of steam can escape from the mat edges due to the presence of a pressure differential between the center of the panel and the edges. In their model, the mat was subdivided into a.series of layers with the initial conditions, material properties, and assigned strands dimensions. 15 The temperature distribution through the thickness of the mat during hot-pressing can be obtained by solving the heat transfer equation for conduction and convection in the mat. Applying energy conservation to a control volume produces the following expression for heat transfer: Es,=Ein-E0Ut+Egen (2.12) where: Est = energy accumulation within in the control volume (J/m3/s) Ein = energy input into the control volume (conductive heat) (J/m Is) Eout = energy output from the control volume (convective heat) (J/m3/s) Eg = generated energy in control volume (J/m3/s) Equation 2.13 is the 1-D heat transfer model (Hubert and Dai, 1999) and can be obtained by substituting the equivalent terms into Equation 2.12; dT , d2T dT Pmcm-z- = kz—T-pvcvv— + q (2.13) dt dz oz where: pm = mat density (kg/m3), cm = mat specific heat (J/kg/°C), pv = steam density (kg/m3), cv = steam specific heat (J/kg/°C), v z = velocity of steam (m/s), and q = heat fluxes (J/s/m3) 2.4.2 Mass transfer mechanism The three general forms of water that can exist in a composite mat were stated by Zombori (2001) as follows: 1. The presence of free water within the cell lumens and in the void spaces between the flakes. 2. The water vapor that occupies the remaining portion of the cell lumens and the void spaces between flakes. 3. The bound water in the cell walls due to hydrogen. 16 For wood composites where the furnish has a moisture content much below the fiber saturation point (FSP), it is very unlikely that free water can be found anywhere in the mat. Typically, the FSP of wood is within the range of 26-33% moisture content depending on wood species, and as wood temperature increases the FSP decreases (Siau, 1984). Because of the low initial moisture content of the strands (a maximum of 12%) free water would not be present in the voids at the beginning of the pressing process. However, as steam reaches the cool center of the mat, it may condense, creating liquid water. Another source of water is the resin which is added to the furnish. If it is a liquid PF resin, they usually contains 50-55% water, and thus the resin adds a small amount of liquid water to the mat (Zombori, 2001). During pressing, water vapor and air are transported by bulk flow and diffusion within the mat and through the edges to the surrounding environment. As described previously, bulk flow is based on total pressure gradient, whereas diffusion of gas molecules occurs due to the partial pressure gradient of that gas. Zombori (2001) derived the following equation that describes the transport mechanism for the vapor in the gaseous phase given by the combination of bulk flow (Darcy's law) and diffusion (Fick's law). dP „ d(pv^ ox ax where: Jv = vapor mass flux (kg/m2/s), km = , mat apparent gas permeability (s), Dm = mat gas diffusivity (kg/m/s/Pa), P = P a + Pv> total pressure (Pa), pa = air pressure (Pa), and pv = partial pressure of steam (Pa) pv = density of steam (kg/m3) vd = void fraction in the mat (-) pg = viscosity of the gas phase, air and vapor (Pa • s) (2.14) 17 2.5 Mat deformation during pressing Permeability of the mat in its principal directions is controlled by the void spaces between wood elements and their connectivity depending on the mat density and wood elements arrangement (Dai and Yu, 2004) and most likely their dimensions. Figure 2.5 shows the relationship between mat pressure and time for hot-pressed OSB boards. Mat pressure increase as mat thickness reduces and four regions have been identified during the compaction of a mat on a typical mat load pressure curve as a function of time (Wolcott et al., 1990). The first stage is press closing (A), which happens in the early stages of press closure when the mat pressure is low. The flakes are compressed to reduce a large portion of voids. As densification continues, the mat pressure rises sharply (Wolcott et al., 1990). The pressure on the mat varies through the sub-layers of the mat structure from faces to the core layers depending on the layer stress-strain relationships and creep properties as well as the temperature and moisture content of each layer (Dai et al., 2000). The second stage (B) starts when the mat reaches maximum pressure whereby the platens are at target thickness and remain there until the resin is cured. During this time the stresses within the mat decrease rapidly and there is considerable stress relaxation within the deformed flakes in the mat through their elastic deflection. The third stage (C) begins when mat pressure suddenly drops and then steadily stabilized for few seconds. As can be seen in the figure this can take 300 s. A small amount of plastic deformation may occur during this period. At the end of the pressing cycle the remaining stresses in the mat are locked in different areas of the board as the resin cures. At the last stage (D), the press gradually opens to vent and relieve the internal vapor pressure. More over, the mat deformation is usually not uniform across the mat thickness due to the variations in heat and moisture content from the surface to the core layer. Increased temperature and causes the resin to polymerize and form permanent bonds between wood elements (Dai, 2001). Thus, the pressing history of the board is expected to affect the permeability of the wood composites which is the subject of this thesis. 18 0 100 200 300 400 500 time (s) Figure 2.4: A typical mat pressure as a function of time during hot-pressing (adapted from Dai et al., 2000). Vertical density profile (VDP) of wood composite panels: V D P is an important characteristic as it affects almost all physical and mechanical properties of panels (Wolcott et al., 1990; Dai et al., 2000). The variation in density through the thickness of the board is established during hot pressing. According to Maloney (1993), as sufficient heat and steam reach the core of particleboard for plasticization, the faces of the panel have already increased in density and resin has cured which holds these layers at the higher density and as a result, the core deforms less since most of the densification has occurred in the surfaces. It can be noted that local densification and the way it develops is the main reason for the formation of the vertical density profile during pressing time (Maloney, 1993). Smith (1982) concluded that the typical M-shape of the density profile might be due to steam migrating from the surface into the interior layers. This causes the furnish to plasticize and reduces the mat's resistance to compression. 2.6 Factors affecting permeability of solid wood Wood is an anisotropic, hygroscopic, porous, and non-homogeneous material. As a result its permeability is affected by many factors including principal directions, moisture 19 content, anatomical structure, wood macrostructure, and sample length. These factors are discussed in more detail in the following paragraphs. Principal directions: The principal directions in wood are the longitudinal, tangential, and radial directions. For most softwoods the permeability in the longitudinal direction is about 10,000 times larger than that in the transverse direction (Dinwoodie, 1989). Tangential permeability is higher than radial permeability (Resch and Ecklund, 1964). Thus, the measurement direction is an important factor. Moisture content: Changes in the moisture content of the wood below the fiber saturation point cause shrinking or swelling of the cell walls (Siau, 1995). For measuring permeability, the moisture content of the sample will ideally be between zero to FSP (Tesoro et al. 1974). Within this range, gas permeability increases as moisture content decreases. Anatomical structure: Large variation in permeability can be observed between hardwood and softwood depending on their anatomical structures. In softwoods the tracheid lumens and bordered pits are the two principal components that have the most significant effect on flow properties. Whereas for hardwoods the principal flow paths is through the vessels and perforated plates (Siau, 1995). Macrostructure of wood: Sapwood is generally much more permeable than the heartwood. Wiedenbeck et al. (1990) measured the permeability of heartwood and sapwood of lodgepole pine (Pinus contorta) and found that the gas permeability of the sapwood was ten times higher than the heartwood due to the presence of extractives in the heartwood. In softwoods, the permeability of the heartwood is also reduced compared with sapwood because of the presence of high levels of resin in the heartwood (Panshin and de Zeeuw, 1980). Avramidis and Mansfield (2005) found that the longitudinal permeability of sapwood in six different aspen clones (Populus tremuloides) was about 12 times higher than that of heartwood. 20 Sample length: Bramhall (1971) found that permeability of Douglas-fir (Pseudotsuga menziesii) increased with decreasing specimen length from 35 mm to 5 mm. Perng (1980) reported permeability values of birch (Betula sp.) and maple (Acer sp.) decreased with increasing specimen length. Wood treatments: Normally, aspiration of pits occurs as wood is dried. It reduces the permeability of the dried wood and increases the difficulty of liquid impregnation (Comstock and Cote, 1968; Siau, 1995). Earlywood has more and larger pits than latewood, thus in the dried condition (due to the pit aspiration) it usually has much lower permeability (Siau, 1995). Permeability variation between samples of the same species: D'Onofrio (1994) measured permeability of white-pine, red-spruce, and balsam-fir and found there can be significant differences in permeability within the same wood species. He reported that permeability of white-pine (Pinus strobus) varies between 18.1xl0" 1 4 and 31.3xl0~1 4 m 3/m, for red-spruce (Picea rubens) between l.OxlO" 1 4 and2.5xl0~1 4 m 3/m, and for balsam fir (Abies balsamed) between 2.0xl0" 1 4 and 4.4xl0" 1 4 m3/m. 2.7 Factors affecting permeability of wood composites A wood-based composite mat is a complex heterogeneous structure. The permeability of the mat during and after hot-pressing is affected by many factors including mat density, size of wood elements, and voids structure and their connectivity. These are discussed below: Mat Density: Density is the main factor controlling the permeability of a composite mat (Humphrey and Bolton, 1989; Hata, 1993; Bolton and Humphrey, 1994; Haas et al., 1998). Kelley (1977) noted that the final average density of a composite mat is controlled by the density of the raw material and the mat compaction ratio during hot-pressing. Work by Hata (1993) on particleboard indicates that board permeability is inversely proportional to density, as shown in Figure 2.5. Hata attributed this to the reduction of the void spaces between particles which increases the resistance to steam flow in the mat. 21 Inspection of the figure also shows that permeability in the longitudinal and transverse directions decreases with increasing density, but that they later converge towards the same value at a board density to 0.6 g/cm3. 25000 - • •—\ ia 20000 - -L: longitudinal 0.3 0.4 0.5 0.6 board density (g/cm3) Figure 2.5: Effect of board density on superficial gas permeability of boards (adapted from Hata, 1993) Garcia and Cloutier (2005) measured the permeability of UF (Urea-Formaldehyde) bonded MDF boards of different densities. The mat was cold pressed until it reached to its target thickness and then heated for an additional 2.5 minutes after the core of the mat reached to 120°C. Total pressing time was 55 minutes and permeability was measured at four density levels: 400, 650, 900, and 1150 kg/m 3. Their results indicated a decrease in gas permeability with an increase in board density due to a reduction in the fraction of voids in the mat, Figure 2.6. 22 1E-11 1E-12 If 1E-13 4 1E-14 a. o 1E-15 a. 1E-16 Log K=6x10-6^m2+0.0048/om-12.93 350 550 750 950 1150 density (kg/m) Figure 2.6: Specific permeability as a function of MDF panel density, adapted from Garcia and Cloutier (2005). Empirical correlation between permeability and density: For modeling purpose, it is required to describe the permeability as a function of known variables. Several researchers have considered permeability to be a function of density only. The permeability data of Sokunbi's work (1978) on particleboard were reported by Humphrey and Bolton (1989) and an exponential curve was fit to those data by Carvalho and Costa (1998). The following equation was generated to describe the specific transverse permeability, Kt, as a function of mat density, p (Carvalho and Costa, 1998); K, = 17.4xl0~"e 13 -0.00806p (2.15) Humphrey and Bolton (1989) established a ratio of 59:1 to describe the magnitude of longitudinal permeability to transverse permeability of particleboard. The relationship between longitudinal permeability as a function of mat density is shown in Figure 2.7. 23 1E-10 ? 1E-11 i 0 200 400 600 800 1000 density (kg/m) Figure 2.7: Permeability of longitudinal direction, K,, and transverse direction, K,, as a function of mat density (adapted from Carvalho and Costa, 1998). Haas et al. (1998) measured the permeability of MDF, particleboard, and OSB of different densities and resin contents, and fitted the results to a non-linear equation of form, f 1 N\ K = exp l (2.16) where, K is the permeability in transverse or in-plane directions and aj, bj, and cj are regression coefficients which vary with permeability direction, board type and resin content. Size of wood elements: Hata (1993) performed a series of experiments that measured the effect of particle size on the gas permeability of boards made by the steam-injection process. The effect of particle length and width on air permeability is shown in Figure 2.8. Both longitudinal and transverse permeabilities decrease with increasing particle length and width for constant particle thickness. It can be seen that permeability in the longitudinal direction is much higher than permeability in the transverse direction. 24 I t" • T " • -r - -1 - - -I 10 20 30 40 50 60 70 80 particle length (mm) Figure 2.8: Effect of particle length on superficial gas permeability of boards (adapted from Hata, 1993). The effect of particle thickness on the air permeability is shown in Figure 2.9 from Hata's work (1993). Air permeability values in the longitudinal and transverse directions both increase with increasing particle thickness. 30000 j L: longitudinal T: transverse | 25000 E o "fj 20000 + Particle: length width o 20 mm, 2 mm A 80 mm, 10 mm 0.3 0.4 0.5 0.6 0.7 0.8 0.9 particle thickness (mm) Figure 2.9: Effect of particle thickness on superficial gas permeability of boards (adapted from Hata, 1993). 25 Hood (2004) examined the transverse and in-plane permeability values of a series of OSB mats made of yellow-poplar (Liriodendron tulipiferd) strands of identical length and width (102 mm x 25.4 mm) but having three different thicknesses (0.5, 0.76, and 1 mm) and various compaction ratios. Boards were cold pressed without resin. He found that permeability increased with increasing strand thickness. D'Onofrio (1994) measured the permeability of OSB mats made from aspen flakes with an average length of 3.5 inches. He found that panels composed of wider flakes form longer and more tortuous pathways for the gas as it traveled from the faces to core of the mat and resulted in low permeability. Voids structure: The gas pathways and connectivity of the void spaces between the wood elements initially affect the composite mat permeability (Bolton and Humphrey, 1994). The flow of the gasses in the mat is primarily around the wood elements rather than through them (Zombori, 2001). Figure 2.10 illustrates a hypothetical transverse pathway around flakes in laboratory made OSB. a hypothet ical g a s pathway Figure 2.10: Cross-sectional area of a laboratory made OSB panel with an average density of 520 kg/m3, at Wood Composites lab, The University of British Columbia. 26 Bolton and Humphrey (1994) discussed their visual observations of the different void structures as affected by particle size and shape and mat compression ratio during pressing of the particles. They examined void structure in low and high density panels made from three types of wood particles (small cubes, splinters, and flakes) to represent the range of the particle size and shapes that can exist in the surface layers of particleboard and core layers of both particleboard and waferboard as shown in Figure 2.11. The boards were made by steam injection-pressing to minimize the density gradient across the thickness. At low densities, 300 kg/m 3, small cubes (measuring 0.5 mm to 0.8 mm a side) produced a relatively uniform porous mat with a homogenous void structure, Figure 2.11a. The void structure of panels made with chips or splinters (0.5 mm x 0.8 mm x 13 mm) consisting of rather longer and tortuous pathways and was more homogeneous than that of flakes, Figure 2.11b. Flakes, 0.5 mm x 3 mm x 25 mm, overlap one another to form a highly heterogeneous void structure, Figure 2.11c. These voids must be linked by low, long and wide slit voids between the flat faces of flakes. At high mat density, 900 kg/m , boards composed of small cubes tend to form a homogeneous and denser mat, Figure 2.1 Id. In this case, it is expected that the mat will have low permeability due to fewer pores in the board structure. A board made with splinters at higher densification, Figure 2.1 le, may be more permeable as a result of some large pores remaining in the mat structure. Flakes form a polygonal pores system in the mat, and the void structure changes during compression in two different ways. First, the thickness of the voids is reduced, shortening the gas pathways and thus increasing permeability. Second, the compact slit-like pores, which exist between the flake faces indicate more resistance to flow (Figure 2.1 If). 27 cubes chip & splinters flakes Figure 2.11: Surfaces of boards made by steam-injection pressing at low and high density from aspen (from Bolton and Humphrey, 1994). 2.8 Summary The importance of permeability of a composite mat for the modeling of the heat and mass transfer during hot-pressing was discussed and the main factors affecting permeability of wood and wood-based composites were identified and described. Since industry needs to manufacture OSB panels efficiently, the effect of the furnish size on permeability and the research reports in this regard were presented. However, no published work has addressed permeability of a mat consisting of mixture of fines and strands in the core of OSB during the mat consolidation. The next chapter describes a pilot scale experiment to establish the appropriate methodology for manufacturing OSB with different core fines contents and densities in the laboratory. The experiment is also designed to test whether meaningful and comparable permeability measurements can be made on OSB samples made at lower than normal density. This work forms the foundation for a larger scale experiment to examine the effects of density and fines on transverse permeability of OSB panels. 28 CHAPTER 3 Pilot Study of Manufacturing OSB Panels with Various Fines Content 3.1 Introduction For a preliminary experiment, the transverse permeability of commercial panels, e.g., OSB, MDF, and PB, was measured in order to explore whether wood element size significantly affects permeability. The complete description of this work can be found in Appendix A. The results from the preliminary experiment suggested that the effect of density on the transverse permeability of wood composite panels is strongly confounded by the size of the wood elements in the composite. It was hypothesized that increasing the fines content in the core layer may increase its permeability. This chapter (pilot study) and Chapter 4 examine the effect of fines content in the core layer on the transverse permeability of full thickness samples, and samples of face and core layers of different densities. The purpose of this chapter is to establish the methodology for board manufacture and determine whether meaningful permeability measurements can be obtained from the samples. Three different permeability apparatus, all of them similar to those described by Siau (1995), were evaluated in order to determine which was most compatible with the samples. Objectives: 1. To determine the feasibility of fabricating OSB boards pressed to various densities in the laboratory where the core layer is composed of a mixture of fines and strands. 2. To determine the minimum possible mat density needed to maintain adequate consolidation of the board from which an intact sample can be prepared. 3. To determine an appropriate test methodology and the most convenient apparatus for the measurement of permeability. 29 3.2 Experimental design The density of commercial OSB varies from 500-800 kg/m 3 (Suchsland and Woodson, 1986). Local variation of the horizontal density distribution (HDD) of the panel can be significant. For example, Kruse et al. (2000) examined commercial panels and found that density ranged from 500 to 850 kg/m for a panel with mean density of approximately 650 kg/m3. In this work three target board densities were used for the experiment to represent the changes of the mat density during pressing. The low density (LD) corresponding to the early stages of the pressing cycle was 450-550 kg/m3. The medium density (MD) that approximately simulated the half-way point during pressing was taken to be 550-650 kg/m3. The high density (HD) was considered the mat at its target thickness with a mean density of 650-750 kg/m 3. In this study 3-layer boards were fabricated with top and bottom face layers consisting of unidirectionally aligned strands with no fines, and a randomly oriented core layer consisting of a mixture of strands and fines. Fines content, Mf, is the mass fraction of fines in the core layer which is computed as follows; Mf= f— (3.1) m{ +ms where mf is the mass of the fines and ms the mass of strands in the core layer. The variable factors and measured responses for the pilot experiment are given in Table 3.1. 30 Table 3.1: Factors and responses of the pilot experiment. Factors: 1. Density levels (target density) 2. Fines content: 0, 50, 100% Responses: 1. Transverse permeability of the full thickness board sample 2. Transverse permeability of each face layer 3. Transverse permeability of core layer 4. Vertical density profile (VDP) Number of replicates per treatment: 1 3.3 Materials and methods 3.3.1 Materials The furnish for this study was provided by Ainsworth Lumber Co. Ltd. (100 Mile House, BC). It was composed of approximately 70% aspen (Populus tremuloides) and 30% of other species such as spruce (Piceae), pine (Pinaceae), and birch (Betula sp.). The length, width, and thickness of the strands were measured using digital caliper on about 100 flakes randomly selected from the furnish. The frequency distribution of strand length is shown in Table 3.2 and Figure 3.1. In a similar manner, the mass portion of fines used in the pilot study is shown in Figure 3.2. The moisture content of the strands was measured to be about 9% (oven dry basis). The adhesive used was a liquid phenol formaldehyde resin (Borden GPF-59M). No wax was added to the furnish. Table 3.2: Frequency, mass, mass fraction, and mean lengths and widths of the strands classified by length. strand length number mass mass average average (mm) of flakes (g) fraction length width (-) (%) (mm) (mm) strands>l 14.3 4 4.89 11.59 118.72 17.20 88.9-114.3 - .-51. 2'2.4: -53.09 -107.50 14.80 63.5-88.9 21 5.1 12.09 75.55 9.10 38.1-63.5 15 4.9 11.61 51.56 8.20 12.7-38.1 6 1.8 4.27 25.98 5.80 strands<12.7 10 3.1 7.35 9.20 2.90 total 107 42.19 100 80.05 11.23 : L D (450 kg/m3) MD(550 kg/m3) HD (650 kg/m3) 31 60 50 40 a? g 30 1 f> 20 10 0 strands M i l <12.7 12.7- 38.1- 63.5- 88.9- > 114.3 38.1 63.5 88.9 114.3 length class (mm) Figure 3.1: Frequency distribution of strands used in pilot study. <0.125 0.125- 0.25- 0.5-1 1.0-2 2.0-4 >4 0.25 0.5 (b) length class (mm) Figure 3.2: Fines used for the pilot study: (a) appearance of a bulk sample of fines, and (b) the weight portion of fines falling in a length classification. 3.3.2 Board preparation Nine boards of different densities and fines content combinations were manufactured as shown in Table 3.3. The mat layers (two faces and the core) were equally weighted with 32 1/3 of the furnish mass for each. Of the nine boards made, seven were used to determine the cutting procedure of the samples, their permeabilities and VDPs. Table 3.3: The treatment combination for density and fines contents of the nine boards made in the pilot study. Samples were only cut from those boards whose label appears in bold. board target density board thickness fines content (%) (kg/m3) (mm) 0 50 100 low: 450 27 LO L5 LI medium: 550 22 MO M 5 M l • high: 650 19 HO H5 H I 3.3.3 Resination The required mass of resin was sprayed onto the furnish using a spray nozzle to achieve an even distribution of resin throughout the strands and fines. The resin content (the weight ratio of solid resin to oven dried furnish, was 3% for the entire mat furnish (both face and core layers). The required mass of strands and fines were blended separately inside a laboratory drum blender, shown in Figure 3.3, according to the manufacturing parameters listed in Table 3.4 plus an additional 10%. The blender was loaded with the furnish to make five boards at a time. The strands were blended, removed from the blender, the blender cleaned, and the fines then blended. The adhesive was sprayed onto the furnish for approximately 10 minutes and the blender was allowed to run for a further 15 minutes to ensure good mixing. The blender was operated at 34 R P M for blending both strands and fines. Once the strands and fines were blended, the mass needed for one board was weighed out and the mat formed. 33 Figure 3.3: The 150 cm (5 ft) diameter blender used to resinate the strands and fines separately. The blender has the depth of 61 cm (2 ft), and rotated in counter-clockwise direction at 34 RPM. The front cover removed to demonstrate tumbling trajectory of strands. Table 3.4: Constituent mass required to produce one board. Description Value Unit Board Length: 30 (cm) Board Width: 30 (cm) Board Thickness: 1.9 (cm) Board Volume: 0.00171 (m3) Board Moisture Content: 1.5 (%wt ofod* board) Board Resin Content: 3 (%wt of od board) Board Wax Content: 0 (%wt of od board) Resin Solids Content: 59 (%wt) Shipping Density: 650 (kg/m3) Furnish Moisture Content: 9 (%wt of od furnish) Board Weight (wet): 1.11 (kg) Board weight (od): 1.10 (kg) Furnish weight (od) 1.06 (kg) (1) Furnish weight (wet) 1.16 (kg) Resin Weight (od) 0.032 (kg) (2) Resin (wet) 0.054 (kg) (1+2) Resinated Furnish Weight 1.21 (kg) * oven dried 3.3.4 Mat formation Each mat consisted of 3 layers as shown schematically in Figure 3.4. Strands in the top and bottom layers were oriented unidirectionally along the board longitudinal direction 34 using an orienter as shown in Figure 3.5a. The core layer was randomly oriented with the mixture of strands and fines for the different fines content, Figure 3.5b. The total weight of the boards was kept constant to within ± lg. Wax paper was used as a release film. The fully formed mat, shown in Figure 3.5c, was laid-up on a sheet of wax paper and another sheet placed on top of the furnished mat to prevent sticking to the aluminum caul plates. ' core j- -^~r — fines j - strands - bottom - -strands-Figure 3.4: Schematic of mat formation in 3 layers. Core layer contains the mixture of fines and strands, and face layers are aligned unidirectional. L is the thickness of the board (dimensions in mm). Figure 3.5: Steps of mat formation: (a) unidirectional strands were laid up in bottom layer by an orienter, (b) randomly oriented the combined strands and fines in core layer, and (c) the entire mat after being formed. 3.3.5 Hot-pressing The mats were hot pressed to the appropriate densities using a 54 ton, 300 mm x 300 mm, electrically heated Wabash press (Figure 3.6a). The mats were pressed between 6.35 mm aluminum caul plates. The controls were limited to platen position, platen pressure, and 35 time. The press closed from daylight to final thickness in nine steps over 76 seconds. The press was then held at the target thickness for 30 minutes. A long pressing time was chosen to ensure that the resin was completely cured. The opening phase was 60 seconds at zero pressure (degassing time). Platen temperatures were set to 185 °C. The press was programmed using PressMan control system. The press cycles for the 19, 22, and 27 mm thickness boards are given in Appendix B. Once all the boards were made, approximately 0.5 mm of material was sanded off each face to remove the wax paper. fi * 1 • '••'11 » f (b) 91 Figure 3.6: The (a) Wabash press was used to press the mats in the Wood Composite Lab of Wood Science Department at UBC and (b) board just after being pressed and laid on the floor of the lab. 3.3.6 Sample preparation The low density zone around the perimeter of the 300 mm x 300 mm board was removed by trimming 50 mm from each edge. The remaining 200 mm x 200 mm portion of the board was cut into 9 square pieces measuring 65 mm x 65 mm as shown schematically in Figures 3.7 using vertical band saw and each labeled with the appropriate number as indicated in Figure 3.8a. Of the 9 pieces, 6 were used for permeability measurements and 2 for vertical density profile test. The test specimens and number of measurements per panel were as follows: 1. Transverse permeability of full thickness of the board 2 observations. 2. Transverse permeability of top layer 1 observation. 3. Transverse permeability of bottom layer 1 observation. 4. Transverse permeability of core layer 2 observations. 5. Vertical density profile (VDP) 2 observations. 36 To avoid any location bias, the properties measured at a given sample position randomly assigned. 300 were Figure 3.7: Cutting pattern of the samples for measuring permeability and VDP tests which were selected randomly (dimensions in mm). The top, bottom, and core permeability specimens were sanded on a wide belt sander (Figure 3.8b) to a final thickness of 5 to 6 mm. The cylindrical permeability specimens, 51 mm in diameter, were cut out of the 65 mm x 65 mm pieces for top, bottom and core layers as well as the full thickness samples using a plug cutter (Figure 3.8c). The remaining square specimens for measuring VDP were cut to 51 mm x 51 mm. A l l test specimens were conditioned at 20°C and 65% relative humidity for over two weeks before being tested. Figure 3.8: Preparation of test specimens: (a) initial board specimens, the nine pieces of 65mm x 65mm, (b) sanding specimens to required thickness, and (c) cutting cylindrical permeability samples. 37 3.3.7 Measurement of permeability The gas permeability measurements were carried out using the falling-water displacement method. The apparatus used to measure the transverse permeability is shown schematically in Figure 3.9. The gas for measuring the permeability is air as described by Siau (1995). Air is drawn through the specimen under vacuum. The available apparatus is best suited for measuring specimens with permeability values up to 1.5x10"" m3/m. Each specimen was placed under high vacuum; thick wall rubber tubing that was clamped tightly to the sample using hose clamps to prevent air leakage between the tubing wall and specimen. The applied vacuum draws water above point 1, shown in Figure 3.9. When the vacuum is turned off, air will flow through the specimen allowing the water to drop. The time required for the level to drop through the distance Az was recorded with a stopwatch. For each specimen, the experiment was repeated three times. Before starting the experiment and between the runs, the specimen was replaced with a rubber plug to test whether the system has remained airtight. It is important that Az be less than 20% of z to reduce errors due to the unsteady-state nature of the method. Vacuum A z ' v r v d Test specimen Figure 3.9: Schematic diagram of gas permeability apparatus by the falling water displacement method (adapted from Siau, 1995). The apparatus was set up in the Wood Physics Laboratory in the Wood Science Department at the University of British Columbia! -38 The superficial gas permeability was calculated from the following equation adapted for use with the apparatus of Figure3.9: Vd-C-L(Palm-0.074z) 0J60mHg (3.2) t -A(0.074z)x(Palm -0.037J) 1.013 x l O 5 Pa where: = superficial gas permeability (m3/m/Pa/s) vd = nr2 Az (m3) Az = change in height of water during the period of measurement (m) r = radius of measuring tube (m) L = length of specimen (m) t = time (s) A = cross section area of specimen (m2) P aim = atmospheric pressure (mHg) c = correction factor for expansion of gas due to change of static head and viscosity of water vr = total volume of system above point 1, including the volume of the hoses (m3) z = average height of water over surface of reservoir during period of measurement (m) The average pressure difference across the specimen is 0.074z (mHg), because the entire pressure differential is due to the static head of water. The average pressure then becomes Patm - 0.037z . The correction factor was calculated using the following equation (Siau, 1995): The superficial gas permeability in the transverse direction was calculated for each specimen as the average of the three runs. The average superficial gas permeability values were multiplied by the viscosity of air C = l + F r(0.074Az) (3.3) Vd(Patm-0.074z) (K = kg • /ja, /i a=1.846xl0" 5 P a s ) for conversion to specific permeability, in unit of m /m, which is independent of the measuring fluid and is solely a function of the material itself. 39 In this chapter there was no attempt to perform any statistical analysis on the measured densities and permeability data except to evaluate the ranges of their mean values. This was done for two reasons. First, the objective of this chapter only focused on the feasibility of the experimental design. Second, in some cases, specimens were broken during sample preparation and the numbers of samples were unbalanced, making the statistical analysis difficult to perform. 3.4 Results and discussion 3.4.1 Transverse permeability Permeability data for full thickness board samples as well as the top, bottom, and core layers were plotted against specimen density in Figure 3.10. Overall, permeability is inversely proportional to density. It was expected that permeability of the top and bottom layers would be similar based on the uniform distribution and spatial arrangement of the furnish in each layer. Figure 3.10b clearly shows that permeability values of the top layers were lower than those of bottom layers, which was not expected. This result may be explained by possible migration of fines from core layer into the bottom layer during the mat formation and/or pressing process. Note also that the top layers were slightly higher in density than bottom layers. l.E-10 l.E-11 I.E-12 l.E-13 l.E-14 l.E-15 bottom top layer f ° 400 500 600 700 800 900 1000 400 500 600 700 800 900 1000400 500 600 700 800 900 1000 (a) density (kg/m') (b) density (kg/m3) (c) density (kg/m3) Figure 3.10: Relationship between permeability and density: (a) .Ml thickness board samples, (b) face, and (c) core layers. 40 The mean permeability and density results of the core layer for each density range are compared in Figure 3.11. Figure 3.11a shows the densities of the core layer for each fines content, and Figure 3.11b indicates increasing permeability with increasing fine content in each density level. Because the connection lines of the means are not parallel, there may be an interaction between fines content and board density. This will be examined more detail in the next chapter. 0 50 100 0 50 100 (a) fines content (%) . (b) fines content (%) Figure 3.11: Comparison of the effect of fines on: (a) mean density versus fines content for different density levels in core layer and (b) relationship between specific permeability and levels of fines content for different density levels in core layer. During cutting of permeability samples, it was noticed that it was difficult to obtain a smooth edge around the perimeter of the specimens, especially for the core layer as fines content increased. This was due to poor consolidation caused by either low resin content or uneven distribution of the resin on strands, and resulted in weak bonds between wood elements. During preparation, some of specimens broke into many smaller fragments and made it impossible to measure their permeability. To overcome this issue, it was decided that the resin content of future boards should be increased from 3% to 5%. It was also found that 450 kg/m was the lowest density of board that could be made and cut into permeability samples without excessive damage to the specimens. 41 3.4.2 Vertical density profile VDPs through entire thickness for each board density level are shown in Figure 3.12 and have the usually M-shape profile. There was considerable density fluctuation through the thickness of the board, similar to those of commercial OSB panels (Section A.4.1). Densities of the face layers for all density levels varied between 800-1000 kg/m 3 and this variation for the core layers was about 420-550 kg/m 3, which are in agreement with the results of the commercial OSB panel measured in the preliminary experiment (Appendix A). Figure 3.12 shows that as the density level or fines content increased, the minimum density of the core layer was increased slightly, except for the board sample HD-100%, Figure 3.12g. This may be caused by an uneven distribution of fines in the core for that particular zone of the board, lack of sufficient furnish, or migration of the fines from core into the bottom layer or a combination of these. These effects will be examined in the next chapter. 42 900 800 -"£700 *600 & i 5 0 0 400 300 200 MD-0% (8) K MD-50% (7) A '(c). , , , , 1 1 ! ; "(d) • 5 10 15 20 thickness (mm) 10 15 20 thickness (mm) 30 0 10 15 20 thickness (mm) 25 30 Figure 3.12: Vertical density profile, low density: (a) 0% and (b) 50%, medium density: (c) 0%, (d) 50%, and (e) 100%, high density: (f) 50% and (g) 100%. The position of each sample on the board is shown in parentheses. 3.5 Conclusions 1. The resin content should be increased from 3% tc 5%; to ensure adequate board consolidation at the lower densities. 2. The lower limit of board density required for cutting intact permeability specimens was 450 kg/m . 43 3. The falling-water displacement method was found to be suitable for measuring the transverse permeability of OSB in the density range examined in this study. 4. The results show that it should be possible to apply the methodology developed in this chapter to a larger scale of experiment. 44 CHAPTER 4 Effect of Fines Content on Transverse Permeability of OSB Panel and Internal Bond Strength 4.1 Introduction The results from the pilot study in Chapter 3 indicated that the permeability of OSB boards of various densities where the core layer is composed of a mixture of fines and strands can be measured using the apparatus based on the falling water method. The minimum target density of the boards was kept at 450 kg/m 3. This forms the lower limit of density to which boards can be pressed and produce samples that remain intact during cutting. In this chapter a larger experiment, similar to the pilot study, examined the effects of density and core fines content on transverse permeability of OSB. Boards were made with five fines content levels (0, 25, 50, 75, and 100%) and three density levels (450-550 kg/m3, 550-650 kg/m 3, and 650-750 kg/m3). Relationships between permeability, density and fines content of the full thickness and core layer sample were examined. The IB of these panels was also measured to determine i f there was a significant reduction of IB strength with increasing core fines content. Prior to IB tests, the vertical density profiles of the samples were measured for comparison with that of commercial OSB. The results in this chapter form the basis for the development of a permeability model of the core layer of OSB at any given core fines content as a function of density (Chapter 5). The main objectives of this chapter are: 1. To evaluate the effect of increasing fines content in the core layer on transverse permeability of partially consolidated OSB mats at three different target densities. 45 2. To examine the impact of increasing fines content and board density on IB strength and whether the V D P of these boards are comparable to commercial OSB. 4.2 Experimental design The experimental design listing the factor levels, experimental responses, and the number of replicates per treatment is given in Table 4.1. Table 4.1: Fixed factors and experimental responses for the laboratory made OSB. Factors: 1. Density levels (target density): L D (low density) 450 kg/m M D (medium density) 550 kg/m 3 HD (high density) 650 kg/m 3 2. Fines content*: 0, 25, 50, 75, 100% Responses: 5. Transverse permeability of the full thickness board sample 6. Transverse permeability of each face layer 7. Transverse permeability of core layer 8. Vertical density profile (VDP) 9. Internal bond test (IB) Number of replicates per treatment: 3 * The portion of fines in the core layer by weight (Equation 3.1) 4.3 Materials and methods Unless otherwise stated, the experimental techniques used in this chapter were identical to those described in Chapter 3 and only those that are different will be described in detail. 4.3.1 Furnish Source: Commercial OSB furnish was obtained from the Ainsworth Lumber Co. Ltd. (100 Mile House, British Columbia). The flakes were collected after the drier and before the rotary screener, so as to obtain flake samples composed of strands and fines. At this plant the furnish is screened into fines and strands; dust and fines are separated and removed using a rotary screener. Fines are defined as any flakes that pass through a screen with 4.8 mm 46 (3/16 inch) square openings. A schematic diagram of the strand flow in the OSB plant from the strander to the mat formers is shown in Figure 4.1. A photograph of the furnish falling from the pneumatic pipe from which the strands were collected is shown in Figure 4.2. The species mix in the furnish was approximately 60% aspen, 30% pine, and 10% birch. Ten batches of flakes were collected in plastic bags at 30-minute intervals between each collection batch to account for possible variation in species composition and flake size. The moisture content of each batch was measured and found to be between 3 to 4% (oven dry basis). strander wet flakes dryer furnish collection point rotary screener dust collector fines fines storage fines -X blender pMDI face layer strands core layer strands core blender pMDI resinated fines & strands T to core layer former face blender PF resinated strands T to top & bottom layers former Figure 4.1: Schematic diagram of the strands flow from the strander to the mat former at Ainsworth OSB plant. Note that strands were collected between the dryer and rotary screener. 47 Figure 4.2: Mixture of strands and fines falling from silo. Screening: To separate the furnish into strands, fines, dust, and unwanted strand sizes, batches 1 to 6 were screened using a mechanical screener as shown in Figure 4.3. Flakes that were retained on the 14.3 mm (9/16 inch) square opening screen were collected and designated as "strands" while those flakes that passed through the 4.8 mm (3/16 inch) square screen, but remained on the 3.2 mm (1/8 inch) round screen, were collected and designated as "fines". Figure 4.4 illustrates the size classification used for the furnish material. Figure 4.3: Mechanical shaker table for screening furnish into 4 size classes in the Composite Lab of Wood Science Department at U B C . To provide sufficient furnish to manufacture all 15 boards for one replicate, the two consecutive batches of screened strands were combined in the following way: batches 1 and 2 for replicate 1, batches 3 and 4 for replicate 2, and batches 5 and 6 for replicate 3. 48 Figure 4.4: Schematic diagram of size classes of screened furnish. 4.3.2 Furnish characteristics To determine the uniformity of furnish across the replicates, the screened strands and fines were analyzed for frequency distribution of length, width, and thickness. The dimensions of the furnish wood elements were measured using a digital caliper on samples randomly selected from each replicate. The length, width, and thickness of strands and fines were measured by taking the longest dimension of each element as its length; the width and thickness were measured at the approximate mid-point of each strand. Only the results for strand length are discussed here; additional information on strand width and thickness can be found in Appendix C. Strands: The frequency of strands in each length interval is shown in Figure 4.5a and the corresponding mass fraction of the total mass of all strands falling within each length interval is shown in Figure 4.5b. The frequency and average length of strands shown in Figure 4.5a was nearly identical for all 3 replicates. Comparison of the mass fraction of each length category, Figure 4.5b, shows essentially the same results. Based on this, it was concluded that the furnish from all 3 replicates was the same. The average length, width, and thickness of the strands used for the three replicates are summarized in Table 4.2. Average piece length of material designated as strands was 85 mm, average width was 12 mm, and average thickness was 0; 7 mm. 49 <12.7 12.7- 38.1- 63.5- 88.9- >114.3 <12.7 12.7- 38.1- 63.5- 88.9- >114.3 (a) 38.1 63.5 88.9 114.3 (b) 38.1 63.5 88.9 114.3 length class (mm) length class (mm) Figure 4.5: Comparison of: (a) frequency distribution and (b) mass fraction of strands in each length class for the 3 replicates. Table 4.2: Average length, width, and thickness of strands for each replication. replicate average average average (-) length width thickness (mm) (mm) (mm) 1 87.26 12.80 0.69 2 84.60 11.49 0.74 3 83.89 11.54 0.74 overall 85.25 11.94 0.72 average Fines: As can be seen from the photograph of a sample of fines, Figure 4.6, material designated as "fines" can be slightly larger than 4.8 mm in width and of nearly any length. The length and mass distribution of the fines from replicate 1 are compared with the results for strands in Figure 4.7. The average length, width, and thickness of the fines used for the experiment were measured as 23.12 mm, 2.85 mm, and 0.62 mm, respectively. Dimensions and weight specification for the six classes of the fines are given in Appendix C. 50 Figure 4.6: Appearance of a fines class showing: (a) a bulk sample of fines and (b) the length range of the fines. Note that the magnification of both images is the same. <12.7 12.7- 38.1- 63.5- 88.9- >114.3 <12.7 12.7- 38.1- 63.5- 88.9- >114.3 (a) 38.1 63.5 88.9 114.3 (b) 38.1 63.5 88.9 114.3 length class (mm) length class (mm) Figure 4.7: Comparison of the averages of: (a) frequency distribution and (b) mass fraction in each length class for fines and strands. 4.3.3 Resination and Mat formation The adhesive type and blending procedure for manufacturing boards were as described in Chapter 3, but the resin content was increased from 3 to 5%, to ensure better mat consolidation and permit the cutting of viable permeability samples. 51 The moisture content of the strands for each replicate was measured just before blending and found to be approximately 4%. The mass of the constituents for each board is given in Appendix D. Mats were formed as described in Chapter 3, except the core furnish content was increased from 33 to 45% (by weight), which is closer in quantity to that used in commercial OSB panels. 4.3.4 Sample preparation Each board was cut into 9 pieces measuring 65 mm x 65 mm, and specimens were prepared for the following tests with the locations of the specimens for each test randomized. The test specimens and number of measurements per panel were as follows: 1. Transverse permeability of the board, full thickness 2. Transverse permeability of top layers 3. Transverse permeability of bottom layers 4. Transverse permeability of core layer 5. Vertical density profile (VDP) 1 observation. 2 observations. 2 observations. 4 observations. 2 observations. 6. Internal bond (the same specimens as used for VDP test) 2 observations. A l l specimens were conditioned at 20°C and 65% relative humidity prior to testing for over two weeks. 5.3.5 Appearance of boards and their core layers The top face of the boards is shown in Figure 4.8. For comparison, the upper face of the core layers that were obtained by sanding down full thickness samples is shown in Figure 4.9. 52 Figure 4.8: The 15 boards of replicate 2 (dimensions in mm). fo}} 1 X *« 0% 25% •00% 50% Figure 4.9: Comparison of typical appearance of the core layers of high density boards with increasing fines contents. 4.3.6 Measurement of internal bond (IB) strength IB strength was measured in accordance with A S T M D1037 (2000). A total of 90 specimens (measuring 51 mm x 51 mm) were cut, i.e. two specimens from each board. Both faces of the specimens were bonded to 51 mm square aluminum alloy plates with hot melt glue. The specimens were conditioned for a further week to constant weight and 53 moisture content in a conditioning chamber maintained at 20°C and 65% (RH). The IB of the boards was evaluated using a universal testing machine (Sintech 30D), shown in Figure 4.10 Tensile load was applied continuously and uniformly at a rate of 1.27 mm/min until failure of specimens occurred. Figure 4.10: Testing IB samples on the Sintech 30D load frame: (a) test set-up and (b) a close-up of a sample just after failure. 4.3.7 Statistical analysis The permeability and IB measurements were made on the samples from all 15 boards for each replicate consisting of a combination of different fines contents (5 levels) and densities (3 levels). Statistical analyses were conducted using JMP statistical software (version 4.0.3). Two-way analysis of variance (ANOVA) was used to test for significant effects of fines content, density, and any interactions on permeability and IB. The complete description of each analysis can be found in Appendix E. The significant results are plotted graphically and error bars presenting the least significant difference (LSD) are included on graphs to determine significant differences between means at the 5% significant level (p< 0.05). Prior to the A N O V A , the distribution of the permeability and IB values was checked using the Shapiro-Wilk test. Permeability values were found to be not normal. The data were transformed (logio), rechecked, and found to be normally distributed. The IB values were also found to be normally distributed. 54 4.4 Results and discussion 4.4.1 Vertical density profile In general, there was considerable density variation through the thickness of the samples. Examination of density variation of the faces and core layers of the made OSB boards showed that the core density was approximately 18% lower than the mean density of the board and the surface layer was 15% higher. This is in approximate agreement with Bolton and Humphrey's (1994) report the density of hot-pressed particleboard which showed that the board surface layer was 25% denser and the core layer 20% less dense than the overall average for the board. Two typical VDPs that came from different locations within the same board for each fines content and density combination are shown in Figure 4.11. It can be seen that there were large density fluctuations through the thickness of the boards similar those observed in commercial OSB (Appendix A , Figure A.8). This is a characteristic of OSB panels, and reflects the non homogeneity of the internal structure of the composite (Wang and Dai, 2004). It might be expected that fines content in OSB affects the shape of the VDP. From Figure 4.11, it can be seen that the density fluctuations in the core are smallest for higher density boards with the higher fines contents of 75 and 100%, possibly due to the more homogeneous structure of the core furnish at these fines contents. 55 LD MD HD KJ IM 100(4) ^ ^ ^ ^ 1M75(7) i i i i i 1 1 1 1 1 1 -1 1M25(5) t i i i i V i 1M(H)(9) K A ^"A 1H5(1(2) . y 1H50(4) 1H25(3) 1H(KI(7) ' ' ' 1 1 ' ' ' 1 1 ' 1 ' 1 1 1 ' 5 10 15 20 25 30 0 5 10 15 20 25 30 0 • 5 10 15 20 25 30 • thickness (mm) thickness (mm) • thickness (mm) Figure 4.11: Comparison of vertical density profiles from two different locations within a board for each fines content (y-axis) and density level (x-axis). VDPs are for the boards of first replicate and the position of each sample on the board is shown in parentheses. 56 4.4.2 Internal bond strength The results obtained from two-way A N O V A showed that among fines content, density and their interaction, only density significantly affected the IB of the boards. The IB values as a function of core density of the samples are shown in Figure 4.12. In general there was only a slight positive correlation between IB strength and sample density, which is in agreement with previous work on European OSB by Sackey (2001), which showed that IB had no strong relationship with density. Similarity, Semple et al. (2005a and b) found little relation between IB and core density of particleboard. Closer examination of Figure 4.12 shows that as density level increased (low to high) the LB strength became more scattered. In general, variation of LB within a product may be due to several underlying factors. It may reflect variability in resin distribution through the furnish, inappropriate pressing schedule, very high platen temperatures and low degassing time, or a combination of these factors. The mean EB strength for each density range are given in Table 4.3 and shows that the average LB strength increases by approximately 35% from low to medium density, while the mean core density increased 13% from 470 to 530 kg/m 3. This result agrees with those of earlier researchers (Maloney, 1993 and Wong et al., 1999) who found that panel strength properties are improved with increasing compression, since there is a better inter-particle contact of wood elements during pressing. Table 4.3 shows that the IB strength for the high density boards was 13% less than medium density boards. It is possible that the reduction in IB strength in the high density boards was caused by the fracture of wood elements. The reduction of LB in highly compacted boards may also have been caused by reduced permeability, causing a build up of trapped steam (Smith, 1982). Trapped gas may have had insufficient time to dissipate and i f the press was opened too early then expansion of the gas may have ruptured bonds leading to reduced LB strength of the cured boards (Smith, 1982). 57 0.7 0.6 «£ 0-5 c °- 4 o X I 13 0.3 E u •S 0.2 0.1 0 LD N=90 MD HD O • * o k A o • I . . . . I 300 400 500 600 700 core density (kg/m ) Figure 4.12: IB as a function of density for all density levels. 800 Table 4.3: Overall core density and IB strength for low, medium, and high density. Items: overall means L D M D HD core density (kg/m3) 464.71 527.32 611.72 IB strength (MPa) 0.27 0.36 0.31 4.4.3 Comparison of measured permeability values with previous research Although there is no previous work identical to this experiment in the open literature, there are several other published studies that are sufficiently similar to enable some comparisons to be made. The permeability of the core layer of commercial OSB panel measured in preliminary experiment (Appendix A) was similar to that of OSB boards at the medium density level (MD) containing 25% fines content in the core. From this it was concluded that the permeability of samples made here were comparable to that of commercially made boards. Hood (2004) gave permeability values for cold pressed mats at 560 and 640 kg/m3 made of identical strands made of yellow-poplar (Liriodendron tulipiferd) with a density of approximately 433 kg/m . His results may be comparable with the core layer of OSB 58 boards composed only of strands at high density level board. Comparison of Hood's permeability values with those of boards here shows that at densities of 550 and 650 kg/m3 of Hood's boards were approximately 5 to 12 times less permeable, possibly due to differences in species type and flake size. D'Onofrio (1994) reported the permeability measurements of an OSB mat consisting only aspen flakes an average length of 89 mm and average thickness of 0.64 mm, with their width varying between 6 and 20 mm at 100°C and the density of 720 kg/m 3. These permeability values were compared with the permeability of the full thickness OSB boards made here at the same range density. The strands in D'Onofrio's board were roughly of similar size to those used here, and thus the fact that his boards had permeabilities approximately 80 times higher than the boards measured here to be is likely due to pressing temperature and species differences and an absence of resin in his boards. One would expect resin to clog the gas flow paths between furnish particles and decease permeability. Sokunbi (1978) reported a decrease in gas permeability from 64 x l O " 1 5 to 2 x 10"15 m3/m as particleboard density increased from 425 to 875 kg/m 3. Measured permeability for the full thickness samples varied from 1000 x lO" 1 5 to 10 x l O " 1 5 m 3/m for the density range from 477 to 870 kg/m 3 indicating that the OSB boards made in this study were 15 to 5 times more permeable than the particleboard measured by Sokunbi, probably due to differences in wood element size and resin content. 4.4.4 Comparison of permeability of full thickness, faces, and core layers of OSB The results from all replicates were pooled and since the number of observations for each sample category was different (unbalanced data), only the means of the permeability data for the faces, core, and full thickness samples are shown in the Figure 4.13. Comparison of the mean permeability of the core samples with the other permeability values for all three density levels shows that the permeability of the core was much higher than the two face layers or the full thickness samples. A t-test at the 0.05 level on the permeability of top and bottom layers showed that there was no significant difference between the 59 permeability values for either face layer (ANOVA results can be found in Appendix E). The mean permeability of the top, bottom, and core layers at each density level decreased with increasing density as shown in Figure 4.13. For comparison the mean density and permeability values are given in Tables 4.4 and 4.5. full top core bottom sample category Figure 4.13: The permeability of full thickness, top, core, and bottom samples. The total number of the samples for each category is shown on the figure. Table 4.4: Mean density of full thickness samples and top, bottom, and core layers. sample category mean density (kg/m3) LD M D HD full thickness 526.82 636.95 732.64 top layer 610.95 741.41 802.01 bottom layer 608.91 747.29 810.17 core layer 460.71 503.10 586.41 Table 4.5: Mean permeability of full thickness samples and top, bottom, and core layers. sample category mean permeabilityxlO13 (m3/m) L D M D HD full thickness 5.06 1.16 0.38 top layer 2.96 0.57 0.58 bottom layer 3.32 0.42 0.48 core layer 65.1 44.2 33.3 60 4.4.5 Effect of density and fines content on permeability Summary of the effects on permeability: A summary of the two-way A N O V A on the effect of fines content and density on permeability is shown in Table 4.6. The results show that fines content, density and their interaction affected the permeability of the core layer. Table 4.6: p-values for density, fines content, and their interaction on permeability of the top, bottom, core layers, and full thickness samples. significance of permeability through the source of layers of lab made OSB variation top core bottom full thickness density l x l O " u 8 x l 0 " l b 4x10"" 6x10"" fines content — 6*10" 3 2 — 0.0428 density* fines — 6xl0" 5 -— 0.53 Full thickness: From Table 4.6 can be seen that density had a significant effect on permeability of full thickness samples for various fines contents, while the effect on the density-fines content interaction was border line significant. The mean permeability as a function of density level for each fines content is shown in Figure 4.14a. The figure shows that for each of the 0, 25, and 50% fines contents, the effect of density on permeability was greatest between low and medium density levels. The changes in permeability of the full thickness board are significant in the early stages of mat compaction where the mat is loosely integrated to medium density level. It can be implied that the largest reduction in void spaces and their connectivity in entire board and mainly in face layers happened in these stages, since the core layer was not quite densified. 61 -10.0 5=5-10.5 E i t - n . o 3 -11.5 •8 | -12.0 5-12.5 i a E-13.0 5 -13.5 -14.0 (a) LSD=0.89 0% 100% 25% 75%' 50% L D full thickness N=45 M D density level H D -10.0 5=5-10.5 E H-11.0 3 -11.5 •9 -12.0 12.5 £ - 1 3 . 0 -S -13.5 -14.0 (b) LSD=0.48 esBst core layer #=180 L D M D density level H D Figure 4.14: Comparison of permeability as a function of density for all fines contents: (a) full thickness sample and (b) core layer. Core layer: The permeability of the core layers as a function of density level for each fines content is shown in Figure 4.14b. It can be seen that in contrast to the full thickness samples, the changes in permeability of the core layers for 0, 25, and 50% fines content was significant in the late stages of pressing (medium to high density level). Core permeability as a function of fines content for each density level is also compared in Figure 4.15. The figure shows that fines content had a significant effect on permeability of the core layer at all density levels. In general, as fines content increased permeability increased for each level of density. The effect of fines content was more pronounced for the denser boards. There was no significant difference between the permeability of the low and medium density samples, however both showed a significant increase in permeability as fines content increased from 0 to 100%. 62 0 25 50 75 100 0 25 50 75 100 fines content (%) W fines content (%) Figure 4.15: Comparison of the permeability of the core layer as a function of fines content for all density levels: (a) log scale and (b) linear scale; note that the LSD bar is asymmetrical on a linear scale and is therefore not shown. At high densities, the permeability of the core layer composed of only strands (0% fines) is much lower, suggesting that the voids are much smaller and the length of the pathways is longer. As fines are added to the core, the permeability values increase steadily until they converge with the permeability values for the lower density boards. The permeability of the core layer becomes independent of density level for 75% and 100% fines contents as can be clearly seen in Figure 4.14b. Since permeability is largely controlled by the connectivity of the void space surrounding furnish particles, one would expect the permeability to decrease at higher density levels, however, this is not the case and suggests that the much smaller flake size of the fines offsets the expected decrease in permeability. More over, there is similarity between Bolton and Humphrey's (1994) model description for the arrangement of flakes in particleboard (see Chapter 2) and the OSB boards in this study consisting mixture of fines and strands in the core layer. Permeability of the core layer depends on two different pore systems. One is the pores built up among the flakes that form the polygonal shape of voids as shown in Figure 4.16a, and the second is due to the very thin slit-like pores between adjacent flake faces, Figure 4.16b. As the density of 63 the board increases there would be no change in the area of the polygonal shape of the voids surrounded by flakes, but rather mat compression shortens the height of polygonal and closes off the very thin slit-like pores and increases the tortuosity of the pathways. However, as the fines content in the core increases (over 75%), this increases the interconnectivity of the voids, and therefore the flow of the gas through these voids becomes easier. Figure 4.16: Schematic structure of the flakes in an OSB board: (a) random position of flakes and the void spaces in between (b) layers of the OSB mat consisting of strands and fines (adapted from Dai et al., 1997). 4.5 Conclusions The major findings of this study are as follows: 1. At each density level, core permeability increased significantly with increasing fines content and the rate of permeability increase with fines content was steeper for board of higher density. 64 2. The transverse permeability of the core layers containing of 0 to 75% fines content, faces layers, and full thickness samples decreased with increasing density. 3. Core permeability was mainly affected by its fines content, when the volume of fines increased to 75% and over in the core layer. At this stage density had little effect on the core permeability. 4. The permeability behavior of the core layer was heavily masked by the highly dense surface layers in full thickness samples. 5. The VDPs of the laboratory made boards were similar to those of commercial OSB panels. 6. Board IB strength increased with increasing board density between low to medium density levels, with no further increase. 7. There was a trend of decreasing IB with increasing core fines content. This was more exaggerated for the high density boards and although not statistically significant. This should be investigated further before attempting such high fines contents in an industrial setting. 65 CHAPTER 5 Permeability Model 5.1 Introduction At the present time, there is no model for the permeability of wood composites composed of a mixture of fines and strands available in the literature. The purpose of this chapter is to develop a model for the permeability of such a composite, based on the similarity in the geometry of the work was previously done on thermal conductivity of fiber and matrix in composite material (Loos and Springer, 1983; Smith, 1992). The permeability of the composite is considered to consist of two adjacent layers, one composed of 100% fines and the other of 100% strands. These layers can be arranged in parallel where the gas flows through both layers simultaneously or in series where gas flows through each layer sequentially. In the parallel model the total gas flow is the sum of gas flow through both layers, whereas in the series model the same amount of gas flows through one layer then the other. As was described in Chapter 2, Darcy's law is used to express the rate of steady-state flow of fluid, Q j due to a pressure gradient, A P , across a porous medium of thickness L. Specific permeability, K, was defined in Chapter 2, Equation 2.10, and can be rearranged to produce an expression for gas flow as: (5..) where, A is cross-sectional area of the specimen and ju the dynamic viscosity of the fluid. 5.2 Model development In order to develop a model for permeability, it. is assumed that the OSB mat, can be modeled as mixture of fines and strands as described above. The motivation for modeling 66 the OSB mat in this way comes from observation of the cross-section of OSB itself, a sample of which is shown in Figure 5.1a. (a) (b) Figure 5.1: Cross-section of OSB panel: (a) mixture of strands and fines in the core layer and (b) concentration of fines and strands oriented in different positions. Examination of the figure shows that some regions consist of layers of strands oriented horizontally with a layer of fines above or below the strands, e.g. position 1, the location of which is shown in Figure 5.1a and schematically in Figure 5.1b. In other regions, the fines are packed to the left or right of the strands, e.g., position 2. In practice, the most common arrangement of strands and fines is a layer of strands oriented at an angle between 0 and 90° to the horizontal with fines located on either one or both sides, as shown in position 3 in the figure where the strands are oriented approximately 30° to the horizontal. It is further assumed that gas flow through these regions can be decomposed into parallel and series flows. The permeability of the system, KsysU!m, shown schematically in Figure 5.2a, can be described by the permeability contributions of a parallel model, Kparallel, shown in Figure 5.2b, and a series model, Kserjes, shown in Figure 5.2c, as a simple rule of mixtures, i.e., K system = <* - Kseries + 0 ~ a)K parallel (5 -2) where a , series coefficient, is the fraction of the permeability associated with the series model. Expressions for Kseries and Kparallel as a function of density are developed in the following two sections. 67 • »* t* fines & strands ^V$* 100% $ fines j« Ml 00% fines (a) (b) (c) Figure 5.2: Schematic of the models: (a) mixture of strands and fines in the core layer of OBS panel, (b) a complex of 100% fines and 100% strands is a parallel model, and (c) a complex of 100% fines and 100%) strands is a series model. 5.2.1 Parallel model development The geometry of the parallel model is shown in Figure 5.3. It is assumed that there is no horizontal flow between the two layers. For convenience, the cell is considered to be of unit depth in the z-direction (into the plane of the page). The total flow through the system, Q,, is the sum of the individual flows through the strands, Qs, and the fines, Qf, as: Q,=Qr+Qs (5.3) Qf a .- - W H i i S E W H ;^ioo% sag i l l f i n e s m Pi - 100% : ~ strands Figure 5.3: Schematic diagram of the parallel model consisting two layers; one of 100% fines and the other of 100% strands. 68 Applying Darcy's law to Qt, Qf, and Qs, and substituting into Equation 5.3 produces, Kparallel ' At ' ^  ^ Kf • Af • A P < K , • A$ • A P where, Kparallel is the predicted permeability of the parallel model, Kf is the permeability of the layer composed of 100% fines, and Ks is the permeability of the layer composed of 100% strands. Recalling that L is the thickness of the layer, A the cross-sectional area of the specimen, and that the subscripts t, f and s refer to the total, fines, and strands layers, respectively. By canceling like terms, this equation may be written as: Kparallel(Af+As) = Kf-Af+Ks-As where, Xf and \ are the width of fines and strands layers, respectively. The above equation can be solved for K llel to produce: Kparallel ~ * * (5-4) Af +/is The density, p, is the ratio of its mass, m, to its volume, which is equal to the product of its cross-sectional area A and its thickness L . Recalling that the cell is considered to be of unit depth in the z-direction, the mass of fines, mf, and strands, ms, can be expressed as, mf = pfAfL = Pf/lfL (5.5) and ms = PsAsL = PsKL (5.6).. As was described in Chapter 3 (Equation 3.1), the fines content, Mf, was expressed as the ratio of the mass of the fines to the total mass of the core layer, i.e., 69 m r Mf= '— (3.1) mf +ms Substituting Equations 5.5 and 5.6 into Equation 3.1, and solving for Xs gives, v LlLLi (5.7) MfPs f Substituting Equation 5.7 into Equation 5.4 and rearranging terms, Kparnllel can be expressed as follows: MfKf +(\-M f)Ks^-Ps K parallel = ~ ~ (5-8) M f + ( 1 - M , ) — From the results in Chapter 5, the ratio of ps I pf for the low, medium, and high density boards was 1.039, 1.045, and 1.152, respectively. The difference in densities of ps and pf for the low and medium density boards can probably be ignored, but the effect of such as assumption for the high density boards is uncertain. For convenience in the derivation of the model it is assumed that pf = ps. The impact of this assumption will be discussed later in Section 5.3.4. Thus, Equation 5.8 reduces to: Kpamllel=MrKf+{\-Mf)-K5 (5.9) 5.2.2 Series model development The geometry of the series model is shown in Figure 5.4. Let P{ and P2 be the pressures of the gas flow through the upper and lower faces of the system, respectively. Let •Pi be the pressure of the interface between the two layers of fines and strands. Since the cross sectional-area A perpendicular to the flow direction is the same through both layers, Qf =QS. Applying Darcy's law, Equation 5.1, into Qf and Qs: 70 Q p, 100% fines Ijjjijjjjjijjjjji' 100%. strands ./L Figure 5.4: Schematic diagram of the series model, consisting two layers; one of 100% fines and the other of 100% strands. f L, L, Setting these two equations equal to each other and solving for Pj, gives: P = P2 +<pP \ + <p where <p = KfLs KLf (5.10) In the same manner, the gas flow through the entire system is equal to the gas flow that passes through the fines layer, i.e., Q - Qf. Substituting Darcy's law for the terms Q and Qf into this equation, an expression for permeability of the series model, Kseries, can be obtained as, ' * - • -KsemsA(Pi-P2) = KfA(P]-Pi) where, Lf and Ls are the thicknesses of fines and strands layers, respectively. 71 The expression for the Pn Equation 5.10, can be substituted into the above expression, gives, KfA K series W-P2) ' which simplifies to, L series \ 1 •* 2. K. 1 v \ + <j> KfKs(Lf+Ls) K5Lf+KfLs The mass expression for each layer is: (5.11) mf = pfLfA and ms = psLsA Substituting the terms for these masses into the fines content expression, Equation 3.1, produces: PfLf pfLf+psLs An expression for Ls in terms of Lf can be obtained by rearranging the above expression to produce: (5.12) By substituting Equation 5.12 in Equation 5.11 and rearranging, gives: KfKs K. Mf+{\-Mf) Pf '* J Kf (\-M f)— + M fK Ps (5.13) Assumingpf = ps, reduces Equation 5.13 to: KfKs Kf{\-Mf) + MfKs (5.14) 72 The three Equations 5.2, 5.9, and 5.14 define the permeability of a mixture of fines and strands. One only needs to find values for the parameters in the model. These can be obtained using the results from Chapter 4, where the permeability of OSB with different fines contents and densities were measured. 5.3 Results and discussion 5.3.1 Description of Kf and Ks as functions of density, p The expressions for the permeability of the parallel and series models given by Equations 5.9 and 5.14 require knowledge of permeability of the core layer composed of only fines, Kf, or only strands, Ks, as functions of density, i.e., Ks=f(p) and Kf=f(p) These equations can be obtained from the permeability results of Chapter 4 for the specimens in which the core layer was composed of only strands or fines. These data are shown in Figure 5.5a for the strands layer and Figure 5.5b for the fines layer. The method of least squares was used to fit an exponential equation of the form of K - aebp to the permeability data, where a and b are constants and p is the density of the core layer. This produced the following equations for the 100% strands layer: K, =45937 x l O - 1 3 * ? - 0 0 ' 6 5 ' (5.15) and for the 100% fines layer, Kf =1405.2 x l O - 1 3 e - ° 0 0 5 V (5.16) These equations have the same form as used by Carvalho and Costa (1998) to fit curves to the particleboard data presented by Humphrey and Bolton (1989), as noted in Chapter 2. 73 Figure 5.5: Permeability as a function of density and best fit curves for: (a) 100% strands layer and (b) 100% fines layer. The predicted permeability of the parallel and series models for 25, 50, and 75% fines contents can be obtained by substituting Equations 5.15 and 5.16 into Equations 5.9 and 5.14. The permeabilities of each model as a function of density for 0, 25, 50, 75, and 100% fines contents are compared in Figure 5.6. Note that the predictions for the 0 and 100% fines contents are identical to the curve fits for Ks and Kf, shown in Figure 5.5. Comparison of model prediction in Figure 5.6a with those in Figure 5.6b shows that the parallel and series models are quite different. In Figure 5.6a the permeability of parallel model increases steadily at a constant rate with fines content for a given density, reflecting the proportionality between K llel and Mf implicit in Equation 5.9. For the series model, the increase in permeability with fines content at constant density also increases, but each successive increase is larger than the one before, due to the inverse proportionality between AT r e, / e j'and <Mf implicit in Equation 5.14. 74 Figure 5.6: Model results for permeability as a function of density for 0, 25, 50, 75, and 100% fine contents: (a) parallel model and (b) series model. The predictions of both models are more easily quantified by comparing the predicted permeabilities at two chosen densities, i.e., 500 and 600 kg/m 3. These density values were chosen since most of the permeability data fall within this density range. The permeability values of the series and parallel models and their ratios are shown in Table 5.1. It can be seen that increasing density reduces the permeability faster for the series model compared with the parallel model, i.e., the permeability of 50% fines content board at a density of 500 kg/m3, K5m, is 4.9 times larger than that of 50% fines content board at a density of 600 kg/m , Km , whereas for the series model, permeability increases only 1.9 times for parallel model. The differences between the parallel and series model as a fraction of fines content at these two densities are compared in Figure 5.7. The comparison between Figure 5.7a and b for the parallel model also shows that the rate .of predicted permeability "at each density is constant for the entire fine content range. From this figure one can see that the range of fines contents corresponding to the greatest difference between the two models is from approximately 60 to 90% fines contents, a range that is well above the fines contents used in commercial OSB. At a fines content similar to commercial OSB e.g. 30%, the 75 difference between the models is still significant as can be seen from the table, where the ratio of A^ 5 0 0 / Km is essentially the same as 50%. Table 5.1: Comparison of the permeabilities of the two models at the densities of 500 and 600 cg/m and their ratios at the indicated fines content. model permeabilityxlO (m3/m) at 500 kg/m 3, K500 permeabilityxlO (m3/m) at 600 kg/m3, K600 -^500 -^600 25% 30% 50% 75% 25% 30% 50% 75% 25% 30% 50% 75% series parallel 15.4 16.3 21.3 34.8 32.7 36.7 53.3 73.9 3.0 3.2 4.4 8.1 15.4 18.1 28.6 41.7 5.1 5.0 4.9 4.3 2.1 2.0 1.9 1.8 100 =to 80 * 1 40 °- 20 500 kg/m3 / / / f / I parallel s j s 1 / f series x ' / / s / • • • • / _—-~ • — " , , 1 1 1 (a) 20 40 60 80 fines content (%) 100 100 80 : -* -60 ---1 40 : -<u -w Q. 20 :-o + (b) 0 20 40 60 80 fines content (%) 100 Figure 5.7: Comparison: of the predicted permeability of parallel and^series models for 0-100% fines contents at the densities of: (a) 500 and (b) 600 kg/m3. 5.3.2 Contribution of series and parallel models to system permeability The predicted permeabilities of the parallel and series models for 25, 50, and 75% fines contents are compared with the permeability data in Figure 5.8. The parallel and series models approximately correspond to the upper and lower bounds of the data, respectively. The predicted permeability of the parallel model is higher than the series model for any given fines content and density. Examination of Figure 5.8 shows that the predicted permeabilities of the series and parallel models are closer together at the low fines content and then diverge as the fines content and density increase and later 76 converge. This difference is smallest at either end of the density range (400 and 700 kg/m ) examined. The largest difference between the series and parallel models occurs over the density range 440-510 kg/m3 for the 25 and 50% fines contents and between 500-530 kg/m3 for the 75% fines contents. 400 (a) 500 600 density (kg/m ) 700400 (b) 500 - V 75% fines - - \ • - - \ \« #=36 \ • • • • \ * \ * •S * • • " V / parallel L • > • 1 'series 600 density (kg/m) 700400 (C) 500 600 700 density (kg/m ) Figure 5.8: Upper and lower boundaries produced by parallel and series model for: (a) 25, (b) 50, and (c) 75% fines contents in the core layer. 5.3.3 Determination of the series coefficient, a The final parameter required for a complete description of the permeability of the system is an estimate of the value of the series coefficient, a . In Chapter 4, the permeability of boards with different fines contents were measured as a function of density at 25, 50, and 75%. These results will be used to determine the value of a . Rule of mixtures approach: To fit the combined parallel and series models (system model, Ksystem) to the measured permeability data for 25, 50, and 75% fines contents, the method of least squares is used. The series coefficient, a, determines the weighting given to the series model and can have values ranging from 0 to 1. It is thought that oc may correspond to the volume fraction of the furnish arrangement in a series model configuration, but further work is required to precisely identify the interpretation of a . 11 To obtain Ksystem, Equation 5.2, the value for a that minimizes the sum of square errors between the measured permeability data and the models must be determined. This was done manually using a spread sheet and the value of a adjusted until the minimum sum of square errors between the data and the model was reached. In practice, computing a to two significant figures was sufficient for our purposes. Table 5.2 compares the sum of the square errors for different a values. As can be seen from the table, the value of a that minimizes this sum is 0.47 for the 25 and 50% fines contents and 0.49 for the 75% fines content. Since these values are so close, it was concluded that a is independent of fines content within the range 25-75% for this furnish. The average value of 0.48 was chosen as the value for a to be used over the whole range of fines contents. Figure 5.8 compares the predicted permeability curves for boards with 25, 50, and 75% fines content generated from Equation 5.2, using a equal to 0.48 with the permeability data collected in Chapter 4. From the figure it can be seen that the model is a reasonably good descriptor of permeability over the range of densities and fines contents examined. Table 5.2: Comparison of the sum of square errors of the predicted permeabilities of the system for different a values. series sum of square errorsxlO2 3 coefficient for each fines content a 25% 50% 75% 0.40 8.936 9.445 1.731 0.46 8.900 9.301 1.698 0.47 8.900 9.299 1.696 0.48 8.902 9.304 1.695 0.49 8.905 9.315 1.694 0.50 8.910 9.332 1.695 •"* 0.60 9.048, 9.853 i:749 78 system (25% fines) system (50%fmes) W=36 system (75%frnes) density (kg/m) 700400 (b) 500 600 density (kg/m) 700400 (C) 500 600 700 density (kg/m ) Figure 5.9: Curve fit to the permeability data using a rule of mixtures approach, combined parallel and series models using an a of 0.48 for: (a) 25, (b) 50, and (c) 75% fines contents. The fact that a = 0.48 minimize the error between the model and the results is interesting because it is so close to 0.5. Since a= 0.48 for these results, it suggests that the arrangement of the strands and fines is very close to being random. If the arrangement of strands and fines were completely random the frequency of the series and parallel configuration will be the same, i.e., a = 0.5. If true, Equation 5.2 can be simplified to: ^system « series ^ parallel > 5.3.4 Sensitivity analysis of the Models Predictions of Ksystem, Kwries, and Kparalkl can be computed from Equations 5.2, 5.8, 5.13, 5.15 and 5.16. These equations are the functions of six independent parameters, i.e., Mf, Ks, Kf, ps, pf, and a . Since the parameters ps and pf are presented in the form of pf Ips, Equations 5.8 and 5..13 can be reduced to only five parameters by considering their ratio, i.e., pf I ps. In order to identify which parameters are likely have the most influence on the models, the effect of small changes in each input parameter, i.e. ±15%, were computed and 79 compared with the permeability at the default case. The value of 15% was chosen since the ratio of ps I pf was measured to be 1.152:1 for high density boards and so using this value should permit the error incurred by the assumption P f = P s to be determined. The default value for fines content was set to 30%, which is the approximate fines content used by the Ainsworth plant in the core layer of their OSB panels and is thought to be representative of other OSB plants. The model predictions for ±15% changes in each parameter of M f , Ks, Kf, pf I ps, and a are compared in Figure 5.10. The solid line is the default case and the dashed lines represent the permeability of the models resulting from ±15% change in the corresponding parameter. Inspection of Figure 5.10 shows that the changes in predicted permeability values are relatively small for all parameters. With the exception of Ks, which affects the series model more than the parallel model. The other parameters affect the parallel model the most. Also of note is that the change in the predicted permeability caused by a 15% increase in a given parameter is of approximately the same magnitude, but opposite to that caused by a 15% decrease in that parameter. This symmetry is more easily seen in Figure 5.11 where the variation between the ±15% changes in each parameter is composed with the permeability value computed for the default case as, — —--1 x 100% " - V K system^ P) ' J where <j> is the designated value for Mf, Ks, Kf, P f I ps, or a . From Figure 5.11, one can see that the response of the system Ksystem, is dominated by the parallel model Kparallel, for Mf, Kf, and P f I P s whereas it is the series model Kseries, that dominates the system in the case of Ks. 80 Figure 5.10: Comparison of the predicted permeability of the system model (left column), series model (center column), parallel model (right column) for ±15% changes in, by row from top to bottom, Mf, Ks, Kf, pf I ps, and a, respectively. 81 system model Mr 15 & 10 I 5 •£ ° 1 -5 & -10 -15 15 & 10 c S 5 1 * ° 1 -5 i °--io -15 15 £] io c o '8 5 •5 > * o P. 500 600 density (kg/m3) series model +15% \ - _ -15% ~ ~ - - - ^ +15% -15% -15% • ' +15% - -15% . 1 . . i i parallel model -15%-500 600 density (kg/m3) 700400 500 600 density (kg/m3) Figure 5.11: Comparison of the predicted permeability variation of the system model (left column), series model (center column), parallel model (right column) for ±15% changes in, by row from top to bottom, Mf, Ks, Kf, pf I ps, and a, respectively. 82 To facilitate the identification of which parameters should be quantified more precisely, a comparison of the effect of each parameter on permeability at 520 kg/m 3 is given in Table 5.3. The value of 520 kg/m3 was chosen because it corresponds to the largest difference between the series and parallel models in Figure 5.8 and is also close to the average core density of the commercial OSB panels measured to be 540 kg/m3 in preliminary experiment (Appendix A). From the table, the effect of the ±15% change in each parameter on the system permeability Ksyslem, from highest to lowest variation is Mf, Kf, pf Ips, a and finally Ks. The effect of the fines content Mf, and the permeability of 100% fines layer Kf, are about 9.5%. It is thought that both of these parameters can be controlled by OSB plants. The fines fraction of the core layer can be adjusted easily just by changing the proportion of fines and strands mixed together for the core furnish. The permeability of the fines layer Kf, is more subtle, but should be able to be controlled by the plant as well. Table 5.3: Comparison of the variation on permeability models caused by ±15% changes in each parameter at 30% fines content and density of 520 kg/m 3. permeability variation (%) parameters system series K parallel -15% +15% -15% +15% -15% +15% M, -9.5 9.7 -5.2 5.9 - -10.9 10.9 Ks -5.9 5.8 -14.5 14.3 -2.9 2.9 Kf -9.2 9.1 -0.7 0.5 -12.1 12.1 Pf!p, 7.5 -6.0 4.5 -3.4 . 8.5 -6.9 a 6.4 -6.4 na na na na na= not applicable From the works of Hata (1993), D'Onofrio (1994), and Hood (2004), it is known that the size of the flakes determines their permeability. If the fines in a plant are produced only by screening the dried furnish, one would have to make adjustments in the strander configuration or in the conditioning of the logs in order to change their size. However, i f the plant is able to reduce the size of the fines further, then they could control Kf by 83 changing particle size. In order to choose the correct particle size, knowledge of Kf as a function of particle size would be required. 5.5 Conclusions 1. Permeability of the core layer of an OSB panel consisting of a mixture of fines and strands can be modeled as a combination of parallel and series models consisting of two layers composed of 100% fines and 100% strands only. 2. The upper bound of the permeability results is approximately coincident with the parallel model and the lower bound corresponding with the series model. The permeability data were well described by a rule of mixtures combination of parallel and series models. 3. The value of the series coefficient a for the system model is independent of fines content from 25-75% fines content and was found to be equal to 0.48. 4. The predicted permeabilities of parallel and series models are closer at low density and diverge as density increases to about 520 kg/m3 and then converge at higher densities. 5. The permeability, of the system ismdst sensitive to the fines content, • M / , and the permeability of the fines, Kf, with system permeability, Ksyslem, changing by 9% for a 15% change in Mf or Kf. OSB plants may be able to control the permeability of the core layer by adjusting the fines content of the layer and the size of the fines in it. 6. The model is general and can be applied to multi-layered boards and for furnish of other conformations. 84 CHAPTER 6 Summary, Recommendations and Future Work 6.1 Summary The literature review (Chapter 2) on the permeability of wood composites and the research undertaken on measuring permeability of wood-based composite panels indicated that there was a need for further research on wood element size and its effect on the permeability of OSB mats. In order to accomplish the objectives of this research, a preliminary experiment on commercial panels (MDF, PB, and OSB) was conducted to examine the feasibility of measuring the transverse permeability of surface and core layers of on these panels (Appendix A). The results showed that in spite of the lower density of the surfaces and core of OSB, their permeability were lower than that of the other composites, contrary to expectation. This finding highlighted that density is not the only factor controlling permeability of wood composite panels. This likely reflects the more tortuous network of flow paths around strands compared with the shorter path between particles or fibers present in PB and M D F . The proportion of large and small wood elements (strands and fines) in the core of OSB is therefore expected to affect the permeability of the mat. This formed the basis of the experiments conducted in Chapters 3 and 4. A series of laboratory scale OSB boards were manufactured with different core fines contents and densities and the transverse permeability of different layers measured (Chapter 3). In this chapter (pilot study) the feasibility of manufacturing OSB in the laboratory with different core fines content and the measurement of transverse permeability was tested. A set of 9 boards were manufactured to determine appropriate resin content and the minimum density to which boards could be compacted but still enable permeability samples to be cut and tested. The pilot experiment indicated that the falling water displacement method was best suited to measurement of transverse 85 permeability of OSB samples. The minimum board density for adequate consolidation was 450 kg/m and the resin content had to be increased to 5%. A comprehensive experiment to measure the effect of core fines content and board density on transverse permeability of OSB was conducted (in Chapter 4). Three replicate boards were made for each of 3 different density levels and 5 levels of fines contents. It was found that fines content in the core layer had a statistically significant effect on the permeability of the core layer at all three density levels. Permeability of the core layer increased with increasing fines content at all density levels and the rate of increase in permeability was higher in the denser boards. It was observed that permeability of the core layer is affected mainly by its fines content when the volume of fines increased to 75% (by weight) in the core layer and density has little influence on the permeability. A model for the transverse permeability of a core layer consisting of mixtures of fines and strands was proposed (in Chapter 5). The model was assumed to contain two adjacent layers, one composed of only fines and the other only strands. The two layers were arranged in either parallel or series. The measured permeability data in Chapter 4 were used to test the model. It was found that the parallel and series models provide the upper and lower bounds for permeability, respectively. The predicted permeabilities of parallel and series models are closer at low density, and diverge as density increases to about 520 kg/m3 and then converge at higher densities. The permeability data were quite well described by a Rule of Mixtures approach for parallel and series models. The proposed model predicts the core permeability of OSB board as a function of fines content and density. Practical implication of this work.-'It was found that the model is sensitive to mass fraction of fines and the permeability of the core layer consisting of 100% fines Kf and 100% strands Ks. Practically in the plant one would have to make adjustments in the configuration of the mixture of fines and strands and fines size in order to obtain the optimized core permeability and therefore reducing press cycle times. 86 6.2 Recommendations and Future Work Since mat permeability is a key factor affecting heat and mass transfer during hot-pressing, accurate knowledge of mat permeability values in the three principal directions (transverse and in-plane) is needed for comprehensive modeling of hot-pressing processes. Research on the influences of fines content on permeability in both in-plane directions should be undertaken. For the purpose of modeling the permeability of core layer, as described in Chapter 5, the series coefficient, a, needs to be explored more deeply. A range of other factors including the shape and size of strands and fines, their distributions and alignment, and void system are likely to influence a . Since void volume, distribution, and their inter-connectivity determine permeability, further research should be conducted to relate the permeability of OSB boards to their volume and the spatial arrangement of voids in the mat. Modeling the effect of void spaces in the core of OSB during mat consolidation as a function of fines content is a useful area for further research that should permit the optimum fines size and geometry to be determined. 87 References American Society for Testing and Materials. 2000. Annual book of A S T M standards; Section 4 A S T M D-1037-Construction (Wood). Vol . 04.10. Avramidis, S., S.D., Mansfield. 2005. On some physical properties of six aspen clones. Holzforschung 59(1): 54-58. Barnes, D. 2001. A model of the effect of strand length and strand thickness on the strength properties of oriented wood composites. Forest Products Journal 51(2): 36-46. Bolton, A.J. , P.E. Humphrey. 1988. The hot pressing of dry-formed wood based composites. Part I. A review of the literature, identifying the primary physical processes and the nature of their interaction. Holzforschung 42(6): 403-406. Bolton, A.J. , P.E. Humphrey. 1994. The permeability of wood-based composite materials, Part I: A review of the literature and some unpublished work. Holzforschung 48(Suppl): 95-100. Bramhall, G. 1971. The validity of Darcy's law in the axial penetration of wood. Wood Science and Technology 5(2): 121-134. Brumbaugh, J. 1960. Effect of flake dimensions on properties of particle boards. Forest Products Journal 10(5): 243-246. Cafferata, A . 2003. The effect of fines on oriented strand board bending properties. MSc. Thesis, Faculty of Forestry, The University of British Columbia. 75 p. Carvalho, L . M . H . , C .A .V. Costa. 1998. Modeling and simulation of the hot-pressing process in the production of medium density fiberboard (MDF). Chem. Eng. Comm.'1-70: 1-21. 1 " Comstock, G.L., W.A. Cote. 1968. Factors affecting permeability and pit aspiration in coniferous sapwood. Wood Science and Technology 2(4): 279-291. Dai, C , P.R. Steiner. 1993. Compression behavior of randomly-formed wood flake mats. Wood and Fiber Science 25(4): 349-358. Dai, C , P. Hubert, S. Chen. 1997. Advances in modeling mat formation and consolidation for wood composite panels. Proceedings First European Panel Products Symposium, Liandudno, Wales, 9 and 10 October: 21-27. Dai, C , C. Yu , P. Hubert. 2000. Modeling vertical density profile in wood composites during hot pressing. Proceedings 5 t h Pacific Rim Bio-Based Composites Symposium, Canberra, Australia, December 10 t h - 13 t h: 220-226. 88 Dai, C. 2001. Viscoelastic behavior of wood composite mats during consolidation. Wood and Fiber Science 33(3): 353-363. Dai, C , C. Yu . 2004. Heat and mass transfer in wood composite panels during hot-pressing. Wood and Fiber Science 36(34): 585-597. Dinwoodie, J .M. 1989. WOOD: Nature's Cellular, Polymeric Fiber-Composite. The Institute of Metals. D'Onofrio, M . 1994. The superficial gas permeability of north-eastern conifers and aspen flake composites. MSc. Thesis, Dept. of Wood Science and Technology, University of Maine. 98p. Garcia, P., S. Avramidis, F. Lam. 1999. Oriented strand board hot pressing and flake orientation. Proc. 4 t h Int. Conf. on the Development of Wood Science, 14-16 July 1999, Missenden Abbey, U K . ed. M . hie, B. Matthews, A . Evans, A . Pitman and A. Comben: 365-374. Garcia, P.J. 2002. Three-dimensional heat and mass transfer during OSB hot-pressing. PhD dissertation, Faculty of Forestry, University of British Columbia, Canada. 254p. Garcia, R.A., A . Cloutier. 2005. Characterization of heat and mass transfer in the mat during the hot-pressing of M D F panels. Wood and Fiber Science, 37 (1): 23-41. Haas, G.V., A . Steffen, A . Fruhwald. 1998. Permeability of fiber, particle and strand mats for gas. Holz als Roh und Werkstoff (German) 56(6): 386-392. Harless, T.E., F.G. Wagner, P.H. Short, R.D. Seale, P.H. Mitchell, D.S. Ladd. 1987. A model to predict the density profile to particleboard. Wood and Fiber Science 19(1): 81-92. Hartley, I.D., V . Mloch. 1998. Permeability apparatus designed and built at Forintek Canada Corp. Personal communication, Hata, T. 1993. Heat flow in particle mat and properties of particleboard under steam-injection pressing. Wood Research, Bulletin of the Wood Research Institute, Kyoto University-80: 1-47. Hood, J.P. 2004. Changes in oriented strand board permeability during hot-pressing. MSc. Thesis, Virginia Polytechnic Institute and State University, Blacksburg, Virginia. 85 p. Hubert, P., C. Dai. 1998. Process modeling for wood-based composites. Part 2: A modular approach to modeling mat consolidation during hot-pressing. Proceeding of Engineered System Using Structural Panels: 37-42. 89 Hubert, P., C. Dai. 1999. A n objective-oriented finite element processing model for oriented strand board wood composites. Proceedings of the 12 t h International Conference on Composite Materials (ICCM-12), Paris, France. Humphrey, P.E., A J . Bolton. 1989. The hot pressing of dry-formed wood based composites. Part II. A simulation model for heat and moisture transfer, and typical results. Holzforschung 43(3): 199-206. Kamke, F.A., L.J . Cassey. 1988. Gas pressure and temperature in the mat during flakeboard manufacture. Forest Products Journal 38(3): 41-43. Kamke, F. A . 2004. Physics of hot pressing. In: Proceedings-Fundamentals of Composite Processing, Nov. 5-6, 2003 Madison, WI, Gen. Tech. Rep. FPL-GTR-149. J.E. Winandy and F.A. Kamke (eds.): 3-10. Kelley, M.W. 1977. Critical literature review of relationships between processing parameters and physical properties of particleboard. U S D A General Technical Report FPL-10. U S D A Forest Service, Forest Products Laboratory, Madison, WI: 57-64 Kruse, K. , C. Dai, A . Pielasch. 2000. An analysis of strand and horizontal density distributions in oriented strand board. Holz als Roh-und Werkstoff 58(4): 270-277. Lang, M.E. , M.P. Wolcott. 1996. A model for viscoelastic consolidation of wood strand mats. Part II: Static stress-strain behavior of the mat. Wood and Fiber Science 28(3): 369-379. Loos, A.C. , G.S. Springer. 1983. Curing of epoxy matrix composites. Journal of Composite Materials 17(2): 135-169 Maloney, T .M. 1993. Modern particleboard and dry-process fiberboard manufacturing. Miller Freeman Inc., San Francisco. Panshin, A.J . , C. de Zeeuw. 1980. Textbook of wood technology. McGraw Hi l l Book Co., New York. 4 t h edition, 722 p. Perng, W.R. 1980. Studies on flow in wood, I: Permeability and axial variation. Forest Products Journal 26(3): 132-138. Resch, H. , B. Ecklund. 1964. Permeability of wood exemplified by measurements on redwood. Forest Products Journal 14(4): 199-206. Petty, J.A., R.D. Preston. 1969. The dimensions and number of pit pores in conifer wood. Proc. R. Soc. London B 172: 137-151. 90 RISI. 2002. Wood products yearbook. Resource Information Systems Inc. Bedford, M A . Sackey, E.K. 2001. Influence of density on mechanical strength of OSB (German). Diplom (MSc.) Thesis, Faculty of Biology, Institute of Wood Technology, University of Hamburg. 124p. Semple, K.E . , E.K. Sackey, H.J. Park, G.D. Smith. 2005a. Properties variation study of furniture grade M2 particleboard manufactured in Canada. Forest Products Journal (In press). Semple, K.E . , E .K. Sackey, H.J. Park, G.D. Smith. 2005b. Properties variation study of furniture grade MS and M2 particleboard manufactured in Canada. Forest Products Journal (In press). Siau, J.F. 1984. Transport processes in wood. Springer-Verlag, Berlin Heidelberg New York. Siau, J.F. 1995. WOOD: Influence of moisture on physical properties. Department of Wood Science and Forest Products, Virginia Polytechnic Institute and State University. Skaar, C. 1988. Wood-water relations. Springer-Verlag, Berlin Heidelberg New York. Smith, D.C. 1982. Waferboard closing strategies. Forest Products Journal 32(3): 40-45. Smith, G.D. 1992. Modeling and experimental issues in the processing of composite laminates. MSc. Thesis, Department of Metal and Materials Engineering, The University of British Columbia. 249p. Sokunbi, O. K . 1987. Aspects of particleboard permeability. MSc. Thesis, University of Wales, U .K. Spelter, H. , R. Wang, P. Ince. 1996. Economic feasibility of products from inland west small-diameter timber. General Technical Report FPL-GTR-92. Madison, WI: U.S. Department of Agriculture, Forest Service, Forest Products Laboratory. 17p. Spelter, H. , D. Mckeever, I. Durbak. 1997. Review of wood-based panel sector in United States and Canada. General Technical Report FPL-GTR-99: Madison;.WI: U.S. Department of Agriculture, Forest Service, Forest Products Laboratory. 45p. Suchsland, O., G.E. Woodson. 1986. Fiberboard manufacturing practices in the United States. U S D A Forest Service. Agriculture Handbook No. 640. 263 p. -Tesoro, F.O., E.T. Choong, O.K. Kimbler. 1974. Relative permeability and the gross pore structure of wood. Wood and Fiber 6(3): 226-236. 91 Turner, D.H. 1954. Effect of particle size and shape on strength and dimensional stability of resin-bonded wood-particle panels. Forest Products Journal 4(5):210-223. Wang, S., Dai, C P . 2004. Press control for optimized wood composite processing and properties Part 2. Properties and press control strategies. In Proceedings-Fundamentals of Composite Processing, Nov. 5-6, 2003 Madison, Wis., J.E. Windandy and F.A. Kamke: 65-82. Wiedenbeck, J.K., K . Hoffman, P. Peralta, C. Skaar, P. Koch. 1990. Air permeability, shrinkage and moisture sorption of lodgepole pine stemwood. Wood and Fiber Science 22(3): 229-245. Wolcott, M.P., F.A. Kamke, D.A. Dillard. 1990. Fundamentals of flakeboard manufacture: Viscoelastic behavior of the wood component. Wood and Fiber Science 22(4): 345-361. Wong, E.D., M . Zhang, Q. Wang, S. Kawai. 1999. Formation of the density profile and its effects on the properties of particleboard. Wood Science and Technology 33(4): 327-340. Wood Markets. 2002. The solid wood products outlook. International WOOD Markets Research Inc. Vancouver, British Columbia. 585p. Zombori, B.G. 2001. Modeling the transient effects during the hot-pressing of wood-based composites. PhD dissertation, Department of Wood Science and Forest Products, Virginia Polytechnic Institute and State University. 92 Appendix A Preliminary Observations of Transverse Permeability of Faces and Core layers of Commercial Panels (OSB, PB, and MDF) A.1 Introduction In the literature review, it was noted that the geometry of wood elements (size and shape) used in wood composite panels has a critical affect on permeability and hence on heat and mass transfer during hot-pressing of composite mats. In this chapter the transverse permeability of three commercial types of wood composite panels, PB, OSB, and MDF, are measured and compared as a function of density. The question of whether density alone controls permeability of wood composite and how fines content in OSB may influence the permeability of the panel is discussed. Objectives: 1. To measure and compare the transverse permeability of three types of commercial wood composite panels and determine the effect of density on permeability. 2. To consider the effect of wood element size on transverse permeability of wood composite panels. A.2 Materials and methods A.2.1 Materials and sample preparation One commercial panel of each, PB, OSB and MDF, 19 mm in thickness, was conditioned to the equilibrium moisture content (EMC) at 20°C and 65% R H for over a period of six weeks. Three 200 mm x 200 mm square sub-panels (SI, S2, and S3) were cut from each panel (Figure A . l ) . Each sub-panel was reduced in thickness to 5 mm by sanding (Figure A.2). SI was sanded to leave the top 5 mm portion, S2 was sanded on both sides 93 to leave the middle 5 mm portion, and S3 was sanded from the top to leave the top 5 mm portion. Dust remaining on the sanded surfaces was blown-off using compressed air. 2440 S1 S2 S3 lllliillWi^ iii^ ^^ i v X 1 —fr 610 | — 1 2 2 0 Figure A. 1: Location of the three sub-panels in each commercial panel. Dimensions are in mm. I h ^ zuu SI i i i l f t ^ j j ''''••'•^llNpi^^ ? 20 1 S2 5 20 t 2 0 S3 -•5-,„ • iM -Figure A.2: Sanding of sub-panels to leave top (SI), core (S2), and bottom (S3) layers. The gray areas were sanded away with a wide belt sander, and all dimensions are in mm. 94 From each layer nine samples were cut as shown in Figures A.3 and A.4: four cylindrical samples, 59 mm in diameter, for permeability measurements and five 51 mm x 51 mm square samples for vertical density profile (VDP) tests. Samples cut from the sub-panel were conditioned at 20°C and 65% RH for another two more weeks prior to being tested. 610 200 (1) Top layer (2) Core layer (3) Bottom layer Figure A.3: Cutting templates for permeability samples (circles), and V D P samples (square); dimensions in mm. Figure A.4: Preparation of permeability specimens showing: (a) sub-panels after sanding, (b) specimen cutting using a hole saw, and (c) permeability specimens. 95 A.2.2 Measurement of transverse permeability Superficial gas permeability, kg, was measured using an apparatus similar to that developed for wood specimens by Petty and Preston (1969) and shown in Figures A.5 and A.6. The specimen was placed in the cylindrical holder shown in Figure A.5b. Figure A.5: Measurement of permeability: (a) apparatus for measuring transverse permeability, (b) cylindrical holder for placing samples. Apparatus was designed and built by I. D. Hartley and V. Mloch (1998) at Forintek Canada Corp. to env i ronment R5 mercury manomete r open to env i ronment mercury manomete r p ressu re regulator PRV p ressu r i zed air Figure A.6: Schematic diagram of permeability apparatus; O M , open mercury manometer to environment for measuring the inlet pressure upstream; M M , differential mercury manometer; F M , air flow rate meter (rotometer); TS, test specimen; D, desiccant; PR, pressure regulator; N V , needle valve; TWV, three-way valve, and R1-R4 are the readings from manometers in mmHg, and R5 is the reading from rotometer. 96 As described in Chapter 2, superficial permeability of the air flow, kg, through the sample was then calculated from the measurement of steady-state flow rate, Q; inlet pressure upstream of sample, P\, outlet pressure downstream of sample, P2, and pressure drop across the sample, P i - Pi or AP, as shown in Figure A.6, from Darcy's Law for gases, (Equation 2.8), as follows: k ^ QL-P_ 8 AAPP The calculated superficial gas permeability values were multiplied by the viscosity of air, K = kg • jia, jua =1.846xl0"5 Pa • s, for the calculation of specific permeability in units of m3/m, which is independent of measuring fluid and solely a function of the material itself. In the remainder of this thesis, the term permeability refers to specific permeability. A.2.3 Measurement of vertical density profile (VDP) The densities of the commercial panels depend on species and product, but generally the density of OSB ranges from 500 to 800 kg/m 3, while densities for PB and M D F range from 600 to 800 kg/m 3 (Suchsland and Woodson, 1986). The continuous density variation through the panel thickness was measured using a non-destructive test for vertical density profiles from the two specimens cut from each sub-panel layer. A n X-ray density profiler, Quintek Measurement Systems, model: QDP-01X, shown in Figure A.7, was used to measure the density of the board material at intervals of 0.06 mm through the cross section of the specimen. The density profiler operates on the principle of measuring the absorption of X-rays that are transmitted through the material. The density of the material at each location throughout its cross section is proportional to the quantity of absorbed X-rays. 97 Figure A.7: The QMS vertical density profiler. A.3 Results and discussion A.3.1 Vertical density profile Typical VDPs for the three types of composite are shown in Figure A.8. Composite types differed markedly in the density of their face and core layers. Average surface density for OSB was around 700 kg/m 3, particleboard around 830-890 kg/m 3, and M D F around 850-900 kg/m3. Density profile through the core layer of each panel type was less variable, with average core density of 540 kg/m3 for OSB, 590 kg/m 3 for particleboard and 600 kg/m3 for MDF. 1 2 0 0 1 . , 400 1 200 -I ' ' ' 1 i • • • • i • • • . l r . 1 0 5 10 15 20 thickness (mm) Figure A.8: Typical Vertical density profiles of tested 19 mm (VA") commercial panels. 98 A.3.2 Relationship between permeability and layer density The permeability of surface and core layers for each composite is plotted as a function of its average density in Figure A.9. As expected, the transverse permeability was lower for the denser surface layers for all board types. In wood composite mats, the transverse and in-plane gas permeability decreases with increasing density (Humphrey and Bolton, 1989; Hass, 1998; and Garcia et al., 1999). In the case of OSB, Figure A.9a, the permeability values of the surface and core layers were much more variable than MDF or PB, Figures A.9b and c. The permeability of OSB for the faces ranged from 4x10"'5 to 8xl0" 1 4 m3/m. This variation in permeability is approximately 10 times larger that of PB (1.5xl0" 1 3 to 6.4xl0" 1 4 m3/m) and M D F (5.2xl0" 1 4 to 1.07xl0"13 m3/m). This was thought to be due to the more heterogeneous nature of OSB which usually consists of a mixture of strands and fines. 1E-11 ~ IE-12 + .E % = 1E-13 1E-14 O S B k • A • (a) • • • top face A core Layer • bottom face PB • *. tf • top face (b) * core layer • bottom Jace MDF (c) • top fecer A core layer • bottom face 1E-15 500 600 700 800 900 1000 500 600 700 800 900 1000 500 600 700 800 900 1000 density (kg/m3) density (kg/m3) density (kg/m3) Figure A.9: Relationship between transverse permeability and density for top, core, and bottom layers of commercial panels: (a) OSB, (b) PB, and (c) MDF. A.3.3 Effects of board type and layer position on permeability and density While the permeability of the three board types generally decreased with increasing layer density, the relationship was not consistent between composite types, as shown in Figure A. 10. The average density values for the surface and core layers of OSB, PB and MDF are compared in Figure A. 10a. The average density for surface and core of PB and MDF were very similar; with surface density between 850 and 900 kg/m 3 and core density around 600 kg/m 3. The density of the surface layers of OSB was about 20% lower than that of PB and MDF, while the core density of OSB was approximately 10% lower than that of the other two composites. 99 The average permeability of each board type for each layer is shown in Figure A. 10b. Despite their lower density, permeability of the OSB samples was lower than the other board types, especially for the face layers. This suggests that density only partly explains the differences in permeability of the different composite types. The lower permeability of the OSB surface layers is expected since it is comprised primarily of strands with some fines. Strands are much larger in length and width than those in PB or MDF, thus the network of flow paths around strands (i.e., the path of least resistance) is much longer and more tortuous than in PB or MDF. Therefore, the variation in wood element dimensions is an important factor affecting the permeability of composite panels. These findings are in agreement with those of Hata (1993) who also found permeability of particleboard decreases as particle length or width increased. 1000 Core layer position 80 -.-70 v * 60 :-ity (m3 /; 50 v permeabil 40 :-permeabil 30 :-permeabil O 20 :-sped 10 :-o :-Bottom (b) Top Core Bottom layer position Figure A. 10: Interaction between board type and layer position for (a) density, and (b) permeability. Void volume and the connectivity of flow paths through the composite determine its permeability (Bolton and Humphrey 1994). Based on these and previous findings (Hata 1993), one might speculate that incorporating a larger number of shorter wood elements into a composite mat such as OSB may increase the void space and number of flow paths 100 between wood elements, thereby increasing the ease with which vapor flows into and through the core during pressing. A .4 Conclusions These preliminary observations suggest that: 1. The VDP of OSB was much more variable than those of particleboard and MDF. The surface layers of OSB were lower in density (around 700 kg/m3) than the surfaces of PB (830 to 890 kg/m3) or M D F (850 to 900 kg/m3), whereas their core zone densities were about 540, 590, and 600 kg/m 3, respectively. 2. The permeability of OSB was significantly lower than that of the other composites despite its lower surface density. The surface layers of OSB, which contain few fines, had the lowest permeability values of any layer from all board types. This likely reflects a more tortuous network of flow paths around strands compared with flow path around smaller discrete particles or fibers present in PB and M D F . 3. The effect of density on wood composite permeability appears to be strongly confounded by wood element size and arrangement. The proportion of small and large wood elements in the core of OSB is therefore expected to strongly affect the permeability of the mat and therefore the efficiency of heat transfer during hot pressing. 101 Appendix B Press Cycles and PressMan Output for Each Board Table B . l : Press schedule for high density board (target thickness 19 mm). SEG CONTROL SET POINT (mm) SEG. TIME (Second) END CONDITION 1 FASTPOSN -12.7mm/SEC 40 POSITION<=51mm 2 PRESSURE 50% 1 3 POSITION 40 5 4 POSITION 35 5 6 POSITION 29 5 8 POSITION 26 5 9 POSITION 23 5 10 POSITION 21 5 11 POSITION 19 5 target thickness 12 POSITION 19 1800 cooking 13 PRESSURE 0 60 degassing 14 POSITION 170 20 opening Table B.2: Press schedule for medium density board (target thickness 22 mm). SEG CONTROL SET POINT (mm) SEG. TIME (Second) END CONDITION 1 FASTPOSN -12.7 mm/SEC 40 POSITION<=51mm 2 PRESSURE 50% 1 3 POSITION 40 5 4 POSITION 35 • 5 • . . . . . . 6 POSITION 32 5 8 POSITION 29 5 9 POSITION 26 5 . 10 POSITION 24 , 5 11 POSITION 22 5 target thickness 12 POSITION 22 1800 cooking 13 PRESSURE 0 60 degassing 14 POSITION 170 20 opening 102 Table B.3: Press schedule for low density board (target thickness 27 mm). SEG CONTROL SET POINT (mm) SEG. TIME (Second) END CONDITION 1 FASTPOSN -12.7mm/SEC 40 SEC POSITION<=51mm 2 PRESSURE 50% 1 3 POSITION 40 5 4 POSITION 35 5 6 POSITION 34 5 8 POSITION 32 5 9 POSITION 30 5 10 POSITION 28 5 11 POSITION 27 5 target thickness 12 POSITION 27 1800 cooking 13 PRESSURE 0 60 degassing 14 POSITION 170 20 opening The press cycle figures for all 15 boards in each replicate can be found as follows: 103 Replicate 1 Figure B . l : Mat pressure, temperature, and gas pressure as a function of time for replicate 1. Columns show boards with different density levels (left to right) and rows indicate fines content (top to bottom). 104 Replicate2 1000 IJ00 time (sec) 14JG 1000 1500 lime (sec) ;ure B.2: Mat pressure, temperature, and gas pressure as a function of time for replicate 2. Columns show boards with different density levels (left to right) and rows indicate fines content (top to bottom). 105 Replicate 3 1 0 0 0 1 5 0 0 time (sec) 1 0 0 0 I J 0 0 time (sec) 1 0 0 0 time (sec) L Aspen comp 0 T J " W ^...7 " 1 4 0 £ , a 1 2 0 ^ 1 0 0 — 8 0 §1 1 0 0 0 time (sec) 1 0 0 0 1 5 0 0 time (sec) 1 0 0 0 time (sec) 1 0 0 0 1 5 0 0 time (sec) Figure B.3: Mat pressure, temperature, and gas pressure as a function of time for replicate . 3. Columns show boards with different density levels (left to right) and rows indicate fines content (top to bottom). 106 Appendix C Length, Width, and thickness Classifications of Strands and Fines Table C . l : Weight, average length, average width, and average thickness of the strands of the first replicate among approximately 390 strands. group length number weight weight average average average of strands of flakes (g) (%) length width thickness (mm) (-) (mm) (mm) (mm) strands> 114.3 28 17.10 13.62 117.77 20.08 0.76 88.9-114.3 194 85.31 67.93 108.18 14.79 0.73 63.5-88.9 77 15.41 12.27 74.29 11.16 0.68 38.1-63.5 61 7.26 5.78 54.19 9.34 0.64 12.7-38.1 26 0.47 0.38 23.64 4.23 0.52 strands<12.7 3 0.02 0.02 6.35 2.50 0.48 Table C.2: Weight, average length, average width, and average thickness of the strands of the second replicate among approximately 880 strands. group length number weight weight average average average of strands of flakes (g) (%) length width thickness (mm) (-) (mm) (mm) (mm) strands>114.3 46 29.62 11.67 117.92 18.46 0.87 88.9-114.3 424 180.70 71.16 108.21 14.41 0.78 63.5-88.9 188 27.43 10.80 74.62 8.41 0.69 38.1-63.5 153 14.83 5.84 52.27 8.43 0.70 12.7-38.1 50 1.25 0.49 24.96 5.38 0.56 strands<12.7 23 0.09 0.04 8.78 2.64 0.59 Table C.3: Weight, average.length, average width, and average thickness of the strands of the third replicate among approximately 1100 strands. group length number weight weight average average average of strands of flakes (g) (%) length , width thickness (mm) (-) (mm) (mm) (mm) strands>114.3 32 17.55 5.80 118.06 15.06 0.81 88.9-114.3 536 217.87 71.97 107.33 13.63 0.79 63.5-88.9 252 44.72 14.77 75.58 11.13 0.69 38.1-63.5 187 20.65 6.82 52.81 9.07 0.72 12.7-38.1 80 1.92 0.63 25.80 4.85 0.58 strands<12.7 15 0.04 0.01 10.09 2.25 0.50 107 Table C.4: Weight, average length, average width, and average thickness of the fines for the first replication among over 5000 fines. group length number weight weight average average average of fines of flakes (g) (%) length width thickness (mm) (-) (mm) (mm) (mm) fmes<12.7 1376 7.06 10.51 8.19 3.28 0.76 12.7-38.1 2943 34.66 51.60 22.79 2.75 0.61 38.1-63.5 576 16.30 24.27 47.04 2.88 0.60 63.5-88.9 99 5.39 8.02 72.13 3.21 0.65 88.9-114.3 36 3.55 5.29 101.83 3.64 0.70 fines>114.3 1 0.21 0.31 115.94 5.07 0.36 108 Appendix D Board Constituents Table D . l : Constituent mass required to produce one board of the comprehensive experiment (Chapter 5). Description Value Unit Board Length: 30 (cm) Board Width: 30 (cm) Board Thickness: 1.9 (cm) Board Volume: 0.00171 (m3) Board Moisture Content: 1.5 (%wt of od* board) Board Resin Content: 5 (%wt of od board) Board Wax Content: 0 (%wt of od board) Resin Solids Content: 59 (%wt) Wax Solids Content: 0 (%wt) Shipping Density: 650 (kg/m3) Furnish Moisture Content: 4 (%wt of od furnish) Board Weight (wet): 1.11 (kg) Board weight (o.d): 1.09 (kg) Furnish weight (o.d) 1.04 (kg) (1) Furnish weight (wet) 1.08 (kg) Resin Weight (o.d) 0.052 (kg) (2) Resin (wet) 0.088 (kg) Wax Solid Weight (o.d) 0 (kg) (3) Wax (wet) 0 (kg) (1+2+3) Resinated Furnish Weight': 1.17 (kg) -* oven dried 109 Appendix E Summary of Analysis of Variance and Raw Data Table E . l : Summary of two-way A N O V A for the permeability of full thickness samples Summary of group averages for each factor Two-way ANOVA Factor Group Sum (density (fines Source of . of Mean F-level) content%) Count Average Variance variance square DF square F-ratio p-value critic 0 12 -12.045 0.001 25 12 -12.395 0.089 Density level 9.968 2 4.984 56.390 6.87E-11 3.315 LD 50 12 -12.664 0.130 75 12 -12.499 0.041 100 12 -12.301 0.070 Fines content 0.994 4 0.248 2.813 0.0428 2.689 0 12 -13.143 0.016 25 12 -13.193 0.058 MD 50 12 -13.462 0.177 Interaction 0.631 8 0.078 0.893 0.53 2.266 75 12 -13.068 0.009 100 12 -12.798 0.203 0 12 -13.637 0.078 . Error 2.651 30 0.088 25 12 -13.459 0.241 HD 50 12 -13.595 0.023 75 12 -13.619 0.145 Total 14.246 44 100 12 -13.258 0.037 Table E.2: Summary of two-way A N O V A for the permeability of core layers. Summary of group averages for each factor Two-way ANOVA Factor Group Sum (density (fines Source of of Mean F-level) content%) Count Average Variance variance square DF square F-ratio p-value critic 0 12 -11.78 0.144 25 12 -11.41 0.134 Density level 10.149 2 5.074 43.164 8.37E-16 3.051 LD 50 75 12 12 -11.32 -11.15 0.060 0.032 100 12 -10..93 0.012 Fines content 29.204 4 7.301 62.099 6.25E-32 2.426 0 12 -11.98 0.121 25 12 -11.82 0.165. MD 50 75 100 12 12 12 -11.45 -11.41 " -11.04 0.046 0.185 ' 0.010 Interaction 4.214 8 0.526 4.480 '6.13E:05 1.995 0 12 -12.75 0.284 Error 19.399 165 0.117 25 12 -12.35 0.218 HD 50 12 -11.85 0.228 75 12 -11.44 0.087 Total 62.967 179 100 12 -11.07 0.028 110 Table E.3: Summary of two-way A N O V A for internal bond (IB). Summary of group averages for each factor Two-way A N O V A Factor Group Sum (density (fines Source of of Mean F-level) content%) Count Average Variance variance square DF square F-ratio p-value critic 0 6 0.318 0.021 25 6 0.318 0.002 Density level 0.117 2 0.058 4.254 0.017 3.118 LD 50 75 6 6 0.219 0.255 0.005 0.003 100 6 0.255 0.005 Fines content 0.109 4 0.027 1.983 0.105 2.493 0 6 0.390 0.012 25 6 0.292 0.003 MD 50 75 100 6 6 6 0.382 0.376 0.365 0.003 0.004 0.009 Interaction 0.149 8 0.018 1.357 0.228 2.064 0 6 0.429 0.021 Error 1.032 75 0.013 25 6 0.333 0.041 HD 50 6 0.336 0.042 75 6 0.234 0.021 Total 1.408 89 100 6 0.220 0.007 Table E.4: Probability values of the t-test at the 0.05 level, applied to permeability and density of top and bottom layers for each density level. probability LD 0.5003 MD 0.3436 HD 0.4429 Table E.5: Raw data for full thickness board specimen test mass atm. press. length diameter ave. time permeability density ID position (s) (mmHg) (mm) (mm) (s) (nrVm) (kg/m3) 1018 W 30.57 752.60 19.86 50.92 87.13 4.83216E-14 755.87 1028 W 32.51 752.60 18.6 50.55 409.26 9.77675E-15 870.91 1031 w 26.99 752.60 19.63 50.64 137.42 3.06208E-14 682.66 1043 w 25.81 752.60 20.02 50.86 131.74 6.48923E-14 634.57 1055 w 27.59 752.60 20.09 50.96 94.45 9.04752E-14 673.32 1063 w 30.88 752.60 22.67 50:74 88.98 5.43958E-14 673.65 1073 w 27.24 752.60 22.69 50.87 85.15 1.13746E-13 590.69 1081 w 28.86 752.60 22.77 50.58 210.25 2.32702E-14 630.79 1098 w 32.13 752.60 23.1 50.74 147.73 6.70877E-14 687.87 1103 w 28.06 752.70 23.05 50.78 167.55 5.89254E-14 601.09 1116 w 28.87 752.70 28.05 50.72 74.20 8.4359E-13 509.41 1122 w 29.5 752.70 28.13 50.62 123.43 5.10619E-13 521.10 1131 w 31.46 753.80 28.05 50.2 194.37 2.26425E-13 566.67 1146 w 31.43 753.80 28.22 50.93 97.76 1.84973E-13 546.70 1159 w 27.28 753.80 27.98 50.45 86.26 5.03909E-13 487.74 2013 w 30.08 753.60 19.84 50.66 288.25 1.46534E-14 752.17 2021 w 27.07 755.70 19.69 50.2 150.31 8.57728E-14 694.62 2036 w 27.47 755.70 20.17 50.84 134.23 3.16766E-14 670.89 2041 w 31.2 755.70 19.62 50.27 347.49 1.21737E-14 801.21 2053 w 29.17 755.70 19.94 50.83 173.12 4.88097E-14 720.91 2062 w 28.88 754.50 22.82 50.88 148.43 9.81526E-14 622.44 2079 w 32.35 754.50 22.79 50.57 261.20 3.74143E-14 706.73 2082 w 31.12 754.50 23.34 50.8 143.52 1.04153E-13 657.84 2097 w 27.91 754.50 22.65 50.21 143.60 1.03403E-13 622.33 2105 w 24.41 754.50 23.2 50.76 75.60 4.68211E-13 519.93 2112 w- 28.05 754.50 27.83 50.8 69.26 8.87311E-13 497.28 2123 w 31.18 754.50 27.85 50.57 .89.92 6.90203E-13 557.41 2131 w 30.25 754.50 27.82 50.52 127.79 4.86089E-13 542.44 2146 w 26.57 754.50 27.6 50.67 143.14 4.27996E-13 477.41 2152 w 27.69 754.50 28.1 50.61 68.17 9.17181E-13 489.84 3017 w 32.14 753.60 19.71 50.74 242.65 1.72386E-14 806.43 3021 w 30.1 753.60 20.01 50.46 172.53 5.00105E-14 752.20 3033 w 30.23 753.50 19.76 50.19 254.65 1.68333E-14 773.26 3041 w 30.94 753.50 19.88 50.81 239.16 1.75948E-14 767.57 3052 w 25.61 753.50 20.11 50.61 113.29 3.78706E-14 633.05 3062 w 29.98 753.50 23.23 50.75 70.89 6.95292E-14 638.00 3075 w 31.56 753.50 22.96 50.71 79.18 6.16203E-14 680.59 3089 w 33.35 752.70 23.13 50.97 287.04 1.69677E-14 706.64 3093 w 28.16 752.70 23.13 50.78 165.08 9.00183E-14 601.15 3106 w 28.76 752.70 23.12 50.77 102.21 1.45391E-13 614.46 3112 w 30.18 752.70 28.34 50.22 66.28 9.73246E-13 537.62 3124 w 32.62 752.70 28.02 50.51 98.97 1.84723E-13 580.99 3138 w 32.01 752.70 27.74 50.42 130.50 9.23536E-14 577.94 3144 w 28.24 752.70 27.73 50.41 107.81 4.00814E-13 510.26 3153 w 28.05 752.60 27.93 50.6 160.11 2.69825E-13 ' 499.43 112 Table E.6: Raw data for top layer specimen ID test position mass (S) atm. press. (mmHg) length (mm) diameter (mm) ave. time (s) permeability (m 3/m) density (kg/m 3) 1013 T 9.84 753.8 5.44 50.93 112.22 1.0257E-14 887.89 1016 T 8.94 753.8 5.34 50.91 113.67 9.94728E-15 822.43 1022 T 9.21 753.8 5.74 50.96 115.63 1.04908E-14 786.68 1027 T 9.36 754.3 5.77 50.96 96.91 1.25744E-14 795.34 1035 T 8.62 754.3 5.36 51.22 76.99 4.40815E-14 780.50 1036 T 8.37 754.3 5.63 50.76 69.65 1.72053E-14 734.66 1041 T 8.76 754.5 5.4 51.04 93.65 1.21363E-14 792.86 1049 T 8.83 754.5 5.53 51.04 130.89 8.89265E-15 780.41 1053 T 8.93 754.5 5.36 51.16 158.95 7.06425E-15 810.47 1057 T 9.27 754.5 5.51 51.19 100.70 1.14492E-14 817.46 1062 T 8.68 754.5 5.65 51.02 97.06 3.71349E-14 751.45 1067 T 8.32 754.5 5.8 50.9 69.99 1.2636E-13 704.97 1078 T 8.88 754.5 5.39 50.97 125.06 9.09628E-15 807.43 1079 T 8.81 754.5 5.69 51.03 84.07 4.31609E-14 757.05 1083 T 8.84 754.5 5.62 50.88 71.43 5.04698E-14 773.63 1087 T 7.81 754.5 5.33 51.01 67.49 5.0402E-14 717.01 1093 T 7.7 754.5 5.65 50.93 98.03 2.44815E-14 668.97 1094 T 6.7 754.5 5.24 51 64.65 1.23105E-13 625.91 1108 T 8.15 754.5 5.61 50.94 145.16 5.88355E-14 712.83 1109 T 8 754.5 5.64 50.99 66.61 5.40804E-14 694.63 1112 T 7.72 754.5 5.74 50.91 87.34 4.21062E-14 660.71 1118 T 6.88 754.5 5.72 50.67 70.74 1.80364E-13 596.49 1126 T 6.65 754.5 •5.69 50.51 64.41 1.9831E-13 583.26 1128 T 6.22 754.8 5.62 50.6 16.25 1.1317E-12 550.38 1134 T 6.71 754.8 5.56 50.95 80.35 1.52605E-13 591.93 1136 T 6.91 754.8 5.81 50.27 40.29 3.26658E-13 599.23 1142 T 7.36 754.8 5.89 51.04 82.93 7.61519E-14 610.73 1147 T 6.75 754.8 5.67 50.95 79.53 1.57222E-13 583.91 1155 T 6.75 754.8 5.72 50.74 27.78 4.57803E-13 583.60 1156 T 6.56 754.8 5.73 50.69 10.04 1.272E-12 567.30 2011 T 10r37 754.8 • -.5,37 50.93 346.49 3.27484E-15 947.91 2017 T 8.24 754.4 '• . '5.56 50.94 113.45 2.08117E-14 727.18 2024 T 10.07 754.4 5.62 50.98 102.58 1.15595E-14 877.82 2029 T 8.81 754.4 5.68 50.96 16.32 7.67815E-13 760.46 2033 T 9.12 754.4 5.5 50.94 169.27 6.86652E-15 813.63 2037 T 7.81 754.4 5.48 50.93 23.07 5.24547E-13 699.57 2044 T 9.05 757 5.62 51.17 98.09 2.40297E-14 783.05 2046 T 8.75 757 5.33 51.08 125.05 1.794E-14 801.10 2052 T 8.21 757 5.37 51.08 76.10 1.47804E-14 746.07 2058 T 9.09 757 5.4 51.04 90.16 1.25648E-14 822.73 2067 T 8.29 757 5.66 51.16 81.89 4.37072E-14 712.50 2069 T 9.12 757 5.63 51.05 64.72 9.29712E-14 791.42 2073 T 9.02 757 5.36 50.92 85.70 1.31817E-14 826.37 2078 T 8.67 757 5.68 50.92 61.42 5.9029 IE-14 749.56 2083 T 8.77 757 5.61 50.9 66.82 3.55886E-14 768.27 113 Raw data for top layer (continued) specimen ID test position mass (g) atm. press. (mmHg) length (mm) diameter (mm) ave. time (s) permeability (m 3/m) density (kg/m 3) 2089 T 8.78 757 5.69 50.9 107.07 1.12096E-14 758.33 2092 T 8.7 757 5.71 50.92 84.18 7.28615E-14 748.20 2093 T 9.06 757 5.65 51.01 55.06 1.09837E-13 784.66 2104 T 7.41 757 5.6 50.81 71.30 1.19776E-13 652.59 2107 T 8.06 757 5.6 51.05 93.80 2.51581E-14 703.18 2113 T 6.87 757 5.54 51.01 63.51 1.31985E-13 606.80 2114 T 7.91 757 5.93 50.95 75.82 1.18619E-13 654.25 2122 T 7.63 757 5.7 50.91 58.01 1.49256E-13 657.59 2125 T 7.34 757 5.74 50.94 37.69 3.35019E-13 627.45 2132 T 7.82 757 5.88 50.93 69.44 7.23846E-14 652.82 2135 T 7.65 757 5.94 50.83 74.63 8.58025E-14 634.67 2144 T 6.38 756.7 5.7 50.98 40.37 3.1026E-13 548.35 2147 T 7.84 756.7 . 5.79 50.87 90.94 1.40494E-13 666.23 2153 T 8.11 756.7 5.88 50.95 114.23 2.17846E-14 676.50 2155 T 7.27 756.7 5.74 50.81 79.17 7.82447E-14 624.65 3014 T 8.92 754 5.36 50.93 55.43 2.04543E-14 816.89 3018 T 9.22 754 5.4 50.91 175.90 6.49881E-15 838.77 3024 T 9.78 754 5.35 50.99 142.71 7.91118E-15 895.21 3025 T 9.03 754 5.37 50.99 66.39 1.70692E-14 823.48 3035 T 9.65 ' 754 5.36 51.12 69.30 3.26324E-14 877.18 3038 T 8.73 754 5.76 51.17 ' 108.31 2.23942E-14 737.01 3042 T 9.4 754 5.42 51.1 214.18 5.31729E-15 845.66 3045 T 8.33 754 5.37 50.86 85.01 2.69247E-14 763.53 3054 T 8.36 754 5.38 51.11 121.96 1.86183E-14 757.39 3055 T 7.9 754 5:39 51.02 54.06 4.2231E-14 716.91 3061 T 8.95 752 5.76 51.09 80.22 1.09025E-13 757.95 3068 T 8.52 752 5.61 50.82 96.32 2.49308E-14 748.72 3074 T 9.32 752 5.73 51.12 75.13 3.22654E-14 792.48 3077 T 8.65 752 5.61 51 96.16 1.234E-14 754.78 3081 T 8.78 752 5.61 51.04 54.19 6.62166E-14 764.93 3082 T 9.08 752 5.64 51.01 39.55 9.13068E-14 787.78 3094 T 7.98 752 5.52 51.06 60.22 3.88661E-14 706.01 3098 T 8.48 752 5.65 51.03 49.52 4.844E-14 733.85 3105 T 8 752 "5.55- 50.98 . 52.18 4.52455E-14.. 706.17 3109 T 8,97 • 752 5.63 ' - 50.97 • 59.80 8.0898E-14 780.85 3111 T 7.35 752 5.78 50.95 51.43 1.71593E-13 623.71 3117 T 6.97 752 5.74 50.94 28.15 3.11453E-13 595.82 3122 T 7.42 752 5.86 50.84 45.15 1.990 IE-13 623.74 3125 T 6.59 752 5.77 " 50.91 7.19 1.77921E-12 561.07 3132 T 7.53 752 5.97 50.92 46.41 5.48459E-14 619.38 3133 T 6.89 752 5.65 51.06 45.88 5.22182E-14 595.55 3146 T 6.67 752 5.71 50.02 30.06 4.36212E-13 594.45 3149 T 7.62 752 5.88 51.08 49.95 1.78833E-13 632.39 3151 T 7.02 752 5.85 51.12 43.29 2.04958E-13 584.67 3156 T 7.11 752 5.75 50.35 58.63 6.44528E-14 621.03 114 Table E.7: Raw data for bottom layer specimen ID test position mass (s) atm. press. (mmHg) length (mm) diameter (mm) ave. time (s) permeability (nvVm) density (kg/m 3) 1013 B 9.73 754.5 5.49 50.95 120.80 9.59957E-15 869.29 1016 B 9.2 754.5 5.28 50.99 198.73 5.60294E-15 853.28 1022 B 9.04 754.5 5.61 50.9 60.39 1.96601E-14 791.92 1027 B 9.41 754.5 5.7 50.98 125.82 9.55734E-15 808.77 1035 B 8.76 754.5 5.72 51.1 54.26 6.70458E-14 746.75 1036 B 8.4 754.5 5.64 51.06 70.52 3.38015E-14 727.36 1041 B 9.01 754.5 5.46 51.31 76.48 2.9878E-14 798.06 1049 B 8.97 754.5 5.4 50.91 79.49 2.8878E-14 816.02 1053 B 8.69 754.5 5.53 51.01 129.58 8.99312E-15 768.94 1057 B 9.44 754.5 5.54 51.26 80.37 1.43835E-14 825.69 1062 B 8.34 754.5 5.59 50.98 46.73 5.07175E-14 730.91 1067 B 7.83 754.5 5.6 50.97 52.43 4.52962E-14 685.26 1078 B 9.25 754.5 5.4 51 56.04 2.03144E-14 838.53 1079 B 8.64 754.5 5.63 50.96 125.15 9.49821E-15 752.41 1083 B 8.9 752.5 5.65 50.95 80.39 1.4884E-14 772.62 1087 B 7.17 752.5 5.31 51.05 53.73 8.45978E-14 659.70 1093 B 7.62 752.5 5.57 51.08 82.81 2.84811E-14 667.59 1094 B 7.22 752.5 5.34 51.03 62.40 3.63037E-14 661.08 1108 B 8.03 752.5 5.71 51 58.57 4.14061E-14 688.41 1109 B 7.82 752.5 5.44 50.97 51.85 2.22011E-14 704.51 1112 B 7.68 752.5 5.77 50.97 56.16 6.58448E-14 652.33 1118 B 6.72 752.5 5.72 51.03 14.47 8.71912E-13 574.42 1126 B 7.29 752.5 5.74 50.98 54.94 4.44085E-14 622.19 1128 B 5.92 752.5 5.73 50.23 19.16 6.80777E-13 521.38 1134 B 7.07 752.5 5.84 51 48.33 1.84011E-13 592.62 1136 B 7.09 752.5 5.78 50.47 55.24 2.35873E-13 613.14 1142 B 7.23 752.5 5.79 50.19 32.18 1.18941E-13 631.15 1147 B 6.62 752.5 5.94 50.4 33.73 3.98112E-13 558.63 1155 B 7.24 752.5 5.71 51.09 24.15 5.2024E-13 618.50 1156 B 6.55 752.5 5.65 50.77 • 23.22 5.18983E-14 572.65 2011 B 10.53 751 5.31 51.11 249.75 4.48362E-15 966.57 2017 B 8.69 751 5.45 51.01 99.58 1.15872E-14 780.23 2024 B 10 751 5.49 51 164.77 7.05662E-15 891.66 2029 B 8.9 751 5.5 50.88 88.15 1.32769E-14 795.87 2033 B 9.64 751 5.46 50.9 87.80 1.3222E-14 867.68 2037 B 7.93- 751 5.5; 50.94 45.53 7.76743E-14 707.46 2044 B 8.92 751 5.4 51 42.86 8.08164E-14 808.61 2046 B 8.53 751 5.32 50.09 191.09 6.11246E-15 813.66 2052 B 8.09 751 5.47 50.87 41.26 5.67118E-14 727.69 2058 B 8.98 " 751 5.38 50.95 80.63 1.4159E-14 818.68 2067 B 8.88 751.5 5.78 50.95 43.75 2.80171E-14 753.54 2069 B 9.2 751.5 5.72 51.05 42.94 2.81387E-14 785.80 2073 B 8.86 751.5 5.4 50.98 56.07 2.0401E-14 803.80 2078 B 9.57 751.5 5.74 50.99 63.52 1.91324E-14 816.47 2083 B 9.15 751.5 5.7 51.08 51.52 4.69109E-14 783.35 115 Raw data for bottom layer (continued) specimen ID test position mass (g) atm. press. (mmHg) length (mm) diameter (mm) ave. time (s) permeability (nrVm) density (kg/m 3) 2089 B 8.68 751.5 5.73 50.91 165.17 7.36864E-15 744.16 2092 B 8.47 751.5 5.79 51.06 57.82 4.24925E-14 714.42 2093 B 9.05 751.5 5.55 50.85 45.69 2.58614E-14 802.94 2104 B 7.61 751.5 5.74 50.7 45.77 1.93515E-13 656.70 2107 B 8.29 751.5 5.63 50.84 42.65 2.81129E-14 725.35 2113 B 6.98 751.4 5.83 51.03 70.14 1.83563E-13 585.39 2114 B 8.03 751.4 5.77 50.93 69.92 1.82953E-13 683.13 2122 B 7.89 751.4 5.84 51.05 61.45 8.14717E-14 660.06 2125 B 7.2 751.4 5.72 50.89 35.03 1.76919E-13 618.84 2132 B 7.38 751.4 5.81 50.72 42.25 2.9429E-14 628.68 2135 B 7.46 751.4 5.78 51.01 30.35 2.0535E-13 631.55 2144 B 6.14 751.4 5.64 50.88 6.76 1.85341E-12 535.43 2147 B 7.64 751.4 5.75 50.92 37.62 1.65422E-13 652.47 2153 B 8.24 751.4 5.94 51.1 56.40 2.22081E-14 676.41 2155 B 6.72 751.4 5.78 51.04 33.65 3.79226E-13 568.24 3014 B 8.97 751.4 5.2 51.01 73.84 1.49018E-14 844.09 3018 B 8.96 751.4 5.47 51.14 75.95 1.51614E-14 797.46 3024 B 10.05 752.7 5.19 50.99 79.42 1.38138E-14 948.28 3025 B 8.9 751.4 5.36 51.15 59.73 1.88853E-14 808.06 3035 B 9.74 752.7 5.26 51.14 53.22 2.07722E-14 901.49 3038 B 8.74 752.7 5.53 51.11 35.03 3.32121E-14 770.34 3042 B 8.99 752.7 5.43 51.08 66.33 1.72446E-14 807.92 3045 B 8.4 752.7 5.35 50.89 47.32 4.822E-14 771.92 3054 B 8.22 752.7 5.46 51.13 16.39 7.31697E-13 733.23 3055 B 8.25 752.7 5.46 51.05 67.77 1.69914E-14 738.21 3061 B 9 752.7 5.77 51.01 ; 40.58 3.00343E-14 763.25 3068 B 8.24 . 752.7 • 5:37. 51.11 28.49 3.9663E-14 747.91 3074 B 9.3 752.7 5.65 51.05 127.54 9.34252E-15 804.18 3077 B 9.22 752.7 5.66 51.14 56.55 6.37089E-14 793.05 3081 B 9.05 752.7 5.65 51.14 36.64 9.81485E-14 779.81 3082 B 9.36 752.7 5.74 51.02' 67.18 1.80399E-14 797.61 3094 B 8.05 752.7 5.66 50.91 45.94 5.25053E-14 698.69 3098 B 8.5 752.7 5.29 51.02 95.53 1.16927E-14 785.95 3105 B 7.67 752.7 5.43 51.19 40.81 1.42242E-13 686.33 3109 B 9.47 752.7 5.69 51.01 109.02 1.10245E-14 814.40 3111 B 7.21 752.4 5.84 50.73 56.37 1.59461E-13 610.81 3117 B 7.06 752.4 5.74 50.94 44.86 8.21085E-14 603.51 3122 B 7.58 752.4 6.02 50.99 43.65 5.86051E-14 616.61 3125 B 6.77 . 752.4 5.82 50.93 33.49 3.84782E-13 570.99 3132 B 7.1 752.4 5.9 51 8.86 2.26256E-12 589.08 3133 B 6.91 752.4 5.75 50.95 40.87 9.02528E-14 589.43 3146 B 7.29 752.4 5.55 51.03 44.93 1.0583E-13 642.23 3149 B 7.01 752.4 5.75 50.93 36.93 1.68223E-13 598.43 3151 B 7.17 752.4 5.73 50.87 38.94 1.59345E-13 615.67 3156 B 7.2 752.4 5.79 50 52.37 2.43162E-14 633.32 116 Table E.8: Raw data for core layer specimen ID test position mass (s) atm. press. (mmHg) length (mm) diameter (mm) ave. time (s) permeability (nvVm) density (kg/m3) 1011 c 8.51 752 6.27 50.73 44.17 9.60108E-14 671.49 1012 c 6.53 752 6.11 50.74 26.23 1.01358E-12 528.54 1014 c 8.49 752 6.33 50.89 47.75 5.91336E-14 659.40 1019 c 7.53 752 6.26 50.73 42.20 1.68742E-13 595.12 1021 c 7.3 752 6.22 50.73 8.56 3.16342E-12 580.65 1024 c 8.58 752 6.27 50.81 48.83 5.74621E-14 674.89 1025 c 7.9 752 6.32 50.84 31.26 2.29027E-13 615.76 1026 c 8.47 752 6.12 50.85 36.08 1.92073E-13 681.49 1033 c 7.3 752 6.27 50.83 10.02 2.7126E-12 573.75 1037 c 6.9 752 6.42 50.73 21.38 1.30728E-12 531.73 1038 c 6.82 752 6.49 50.63 6.66 4.25917E-12 521.96 1039 c 7.22 752 6.29 50.82 11.52 2.36864E-12 565.88 1044 c 6.39 752 6.15 50.87 3.57 7.45854E-12 511.23 1045 c 6.98 752 6.13 50.81 5.58 4.76474E-12 561.57 1047 c 6.91 752 6.1 50.81 21.16 1.25089E-12 558.68 1048 c 7 752 6.07 50.78 11.51 2.29139E-12 569.42 1051 c 6.61 754.5 6.12 50.86 4.13 6.39974E-12 531.63 1052 c 6.69 754.5 6.03 50.72 3.68 7.12226E-12 549.11 1054 c 6.55 754.5 6.21 50.74 3.29 8.19046E-12 521.63 1056 c 6.09 754.5 6.21 50.8 2.52 1.06538E-11 483.85 1064 c 6.04 754.5 5.92 50.8 27.77 4.93473E-13 503.38 1065 c 6.12 754.5 5.97 50.83 18.85 1.36942E-12 505.18 1066 c 6.13 754.5 5.88 50.74 36.05 7.07692E-13 515.58 1068 c 6.24 754.5 5.95 50.82 10.92 2.35689E-12 517.02 1071 c 6.32 754.5 5.99 50.84 19.06 1.35833E-12 519.74 1074 c 5.34 754.5 5.89 50.78 8.85 2.88337E-12 447.66 1076 c 6.16 754.5 5.96 50.84 43.26 5.95473E-13 509.14 1077 c 6.66 754.5 6.05 50.83 25.30 2.70147E-13 542.49 1082 c 5.8 754.5 5.97 50.82 4.85 5.32082E-12 478.95 1084 c 5.66 754.5 . 5.89 50.83 4.44 5.74026E-12 473.56 1085 c 6.43 ' 754.5 6.02 ' 50.73 5.97 ' 4.3773E-12 528.44 1088 c 6.47 754.5 5.95 50.73 22.04 1.1719E-12 537.98 1091 c 6.56 754.5 6.04 50.62 17.15 1.53547E-12 539.68 1092 c 6.77 754.5 6.02 50.82 20.51 3.31706E-13 554.41 1096 c 5.22 754.5 5.91 50.61 2.51 1.0256E-11 439.06 1099 c 6.43 754.5 6.02 50.82 6.95 3.74856E-12 526.57 1101 c 6.44 754.5 5.93 50.82 4.21 6.0928E-12 535.39 1102 c 6.02 754.5 5.91 50.58 3.04 8.48926E-12 506.95 1106 c 5.37 754.5 5.88 50.73 2.28 1.12114E-11 451.83 1107 c 6.14 754.5 5.88 50.86 3.37 7.52799E-12 513.98 1111 c 5.99 751.8 6.09 50.81 23.94 1.1044E-12 485.09 1113 c 5.53 751.8 5.73 50.75 32.98 7.55891E-13 477.10 1115 c 5.77 751.8 6.11 50.82 20.36 1.30238E-12 465.56 1117 c 5.97 751.8 6.07 50.76 19.89 1.32756E-12 486.02 1121 c 5.86 751.8 6.14 50.82 30.37 8.77255E-13 470.51 117 Raw data for core layer (continued) specimen test mass atm. press. length diameter ave. time permeability density (kg/m 3) ID position (R) (mmHg) (mm) (mm) (s) (m 3/m) 1123 C v 6.39 , 751.8 6.12 50.72 21.80 1.22313E-12 516.77 1124 c 5.62 751.8 6.08 50.87 6.71 3.924E-12 454.80 1125 c 4.95 751.8 6.06 50.81 2.38 1.10682E-11 402.85 1132 c 5.99 751.8 6.08 50.77 6.93 3.81441E-12 486.65 1135 c 5.53 751.8 6.1 50.79 3.59 7.37475E-12 447.45 1137 c 6.02 751.8 6.14 50.49 17.98 1.5012E-12 489.70 1139 c 5.3 751.8 6.06 50.6 6.94 3.82377E-12 434.92 1143 c 5.37 751.8 6.06 50.79 4.24 6.21389E-12 437.38 1144 c 5.12 751.8 6.03 50.8 2.07 1.26704E-11 418.92 1148 c 5.43 751.8 6.1 50.72 6.81 3.90017E-12 440.58 1149 c 5.91 751.8 6.07 50.82 4.85 5.43062E-12 480.00 1151 c 5.35 751.8 6.06 50.77 2.01 1.31079E-11 436.09 1152 c 5.35 751.8 6.03 50.6 2.05 1.28746E-11 441.21 1157 c 5.5 751.8 6.13 50.56 2.20 1.21966E-11 446.89 1158 c 6.13 751.8 6.11 50.65 2.27 1.17578E-11 497.93 2012 c 7.72 754.5 6.09 50.8 34.23 3.95279E-14 625.44 2015 c 8.23 754.5 6.14 50.81 20.41 6.68284E-14 661.06 2016 c 6.93 754.5 6.12 50.79 23.54 1.74956E-13 558.90 2018 c 7.12 754.5 6.18 50.84 20.74 1.28769E-12 567.53 2022 c 7.14 754.5 6.12 50.81 35.80 3.95573E-13 575.38 2025 c 8.16 754.5 6.1 50.86 33.89 4.15763E-13 658.44 2026 c 8.56 754.5 6.1 50.84 28.33 9.30539E-13 691.26 2028 c 7.86 754.5 6.19 50.82 34.57 7.7445E-13 626.00 2032 c 7.35 754.5 6.08 50.91 31.89 8.21867E-13 593.86 2034 c 6.96 754.5 6.11 50.8 14.01 1.88749E-12 562.02 2035 c 7.17 754.5 6.04 50.84 32.58 4.28521E-13 584.76 2039 c 7.58 754.5 6.11 50.87 25.61 2.691E-13 610.40 2042 c 7 754.5 6.24 50.74 4.87 5.55991E-12 554.78 2043 c 6.6 754.5 6.33 50.85 4.01 6.82578E-12 513.41 2047 c 6.81 754.5 6.19 50.78 3.53 7.60421E-12 543.23 2048 c 7.39 754.5 6.23 50.81 6.23 4.32497E-12 585.02 2051 c 6.62 754.5 6.13 50.52 2.52 1.06475E-11 538.74 2055 c 6.49 754.5 6.27 50.49 2.21 1.24331E-11 516.98 2056 c 6.22 754.5 6.22 50.94 . 2.09 1.27923E-11 490.67 2057 c 6.48 754.5 6.18 50.79 ' 2.40 1.1167E-11 "517.54 2061 c 6.19 754.5 6.11 50.8 24.39 1.0846 IE-12 499.84 2064 c 6 754.5 6.06 50.82 27.38 9.57261E-13 488.11 2065 c 5.41 754.5 5.99 50.83 . 4.26 6.08458E-12 . 445.08 2066 c 6.37 754.5 6.03 50.77 ' 25.76 1.01467E-12 "521.82 2071 c 6.14 754.5 6.02 50.79 27.70 9.41182E-13 503.41 2074 c 5.76 754.5 6 50.77 11.87 2.1914E-12 474.21 2075 c 6.01 754.5 5.97 50.88 11.52 2.23571E-12 495.13 2077 c 6.68 754.5 6.03 50.87 20.99 1.2402 IE-12 545.06 2084 c 5.88 754.5 5.96 50.82 6.45 3.99903E-12 486.38 2085 c 6.73 754.5 6.03 50.79 15.00 1.74094E-12 550.87 118 Raw data for core layer(continued) specimen ID test position mass 00 atm. press. (mmHg) length (mm) diameter (mm) ave. time (s) permeability (nvVm) density (kg/m3) 2087 c 5.86 754.5 5.99 50.72 4.25 6.12058E-12 484.20 2091 c 6.5 756.8 6.05 50.8 5.18 5.0459 IE-12 530.08 2095 c 5.49 756.8 5.94 50.68 3.19 8.08608E-12 458.17 2096 c 5.84 756.8 6.02 50.83 2.93 8.86033E-12 478.06 2098 c 5.52 756.8 5.96 50.66 2.59 9.97743E-12 459.49 2103 c 6.12 756.8 6.02 50.66 2.80 9.33404E-12 504.35 2106 c 5.78 756.8 5.91 50.64 2.46 1.04524E-11 485.58 2108 c 5 756.8 5.88 50.84 2.01 1.26105E-11 418.88 2109 c 6.42 756.8 5.89 50.82 4.00 6.34725E-12 537.35 2111 c 6.71 756.8 6.19 50.84 17.58 1.51783E-12 533.99 2115 c 5.42 756.8 6.12 50.71 8.18 3.24302E-12 438.50 2117 c 6.13 756.8 6.13 50.76 13.99 1.89479E-12 494.16 2118 c 5.13 756.8 6.08 50.71 3.52 7.47694E-12 417.77 2121 c 5.26 756.8 6 50.84 5.48 4.71977E-12 431.85 2126 c 5.03 756.8 6.06 50.73 2.69 9.74124E-12 410.65 2127 c 6.29 756.8 6.17 50.83 10.63 2.50229E-12 502.38 2129 c 6.02 756.8 6.08 50.81 7.20 3.64616E-12 488.32 2133 c 5.82 756.6 6.13 50.7 4.52 5.87553E-12 470.28 2134 c 5.03 756.6 6.05 50.66 2.08 1.26306E-11 412.47 2138 c 5.73 756.6 6.08 50.79 4.82 5.45333E-12 465.16 2139 c 5.51 756.6 6.07 50.63 3.81 6.92037E-12 450.88 2141 c 5.5 756.6 6.07 50.79 6.51 4.03028E-12 447.23 2142 c 5.41 756.6 6.08 50.82 2.49 1.05365E-11 438.67 2145 c 4.89 756.6 6.06 50.82 2.31 1.13039E-11 397.81 2149 c 5.56 756.6 6.1 50.5 2.24 1.18827E-11 455.06 2154 c 5.2 756.6 5.97 50.76 1.80 1.43456E-11 430.42 2157 c 5.28 756.6 6.07 50.8 2.00 1.31286E-11 429.17 2158 c 5.29 756.6 6.03 50.74 1.93 1.35012E-11 433.86 2159 c 6.45 756.6 6.08 50.77 4.20 6.26392E-12 524.02 3011 c 7.58 756 6.14 50.82 46.73 5.67368E-13 608.61 3013 c 7.59 756 6.22 50.86 38.92 3.68444E-13 600.63 3015 c 8.31 756 6.28 50.82 32.82 2.15903E-13 652.35 3016 c 8.82 756 6.29 50.82 43.43 3.2102E-14 691.29 3022 c 7.31 756 6.11 50.78 43.67 ' 4.96976E-13 590.75 3023 c 7.94 756 6.17 50.82 16.70 1.59569E-12. 634.42 3027 c 6.24 ' 756 ' 6.04 50.79 49.79 4.30788E-13 509.92 3029 c 8.77 756 6.21 50.84 44.29 1.58094E-13 695.68 3031 c 7.75 756 6.23 50.84 56.27 2.55487E-13 612.79 3036 c 8.12 .756 6.22 50.89 ' 25.52 1.04956E-12 '641.82 3037 c 7.13 756 6.24 50.84 6.45 4.17422E-12 562.86 3039 c 6.87 756 6.32 50.77 3.75 7.29177E-12 536.95 3043 c 8.07 756 6.23 50.82 19.77 1.36074E-12 638.60 3044 c 8 756 6.21 50.82 19.65 1.36442E-12 635.10 3046 c 7.66 756 6.23 50.82 5.42 4.96037E-12 606.15 3047 c 7.72 756 6.22 50.82 6.91 3.88504E-12 611.88 119 Raw data for core layer(continued) specimen ID test position mass ( B ) atm. press. (mmHg) length (mm) diameter (mm) ave. time (s) permeability (m 3/m) density (kg/m 3) 3056 c 6.96 756 6.25 50.74 2.98 9.08499E-12 550.73 3058 c 7.33 756 6.11 50.81 5.65 4.6715E-12 591.66 3059 c 8.38 756 6.26 50.86 6.98 3.86843E-12 658.91 3063 c 7.38 755.7 6.28 50.83 33.26 4.36049E-13 579.12 3066 c 5.82 755.7 5.96 50.91 11.57 2.21791E-12 479.71 3067 c 6.82 755.7 6 50.77 47.53 5.46324E-13 561.47 3069 c 7.08 755.7 6.15 50.8 32.95 4.3155E-13 567.99 3071 c 6.73 755.7 6.03 50.78 27.63 5.05008E-13 551.09 3072 c 5.76 755.7 6.05 50.77 3.17 8.25158E-12 470.29 3073 c 6.3 755.7 6.09 50.63 9.85 2.69168E-12 513.83 3078 c 5.83 755.7 5.95 50.74 8.83 2.921E-12 484.57 3083 c 5.68 755.7 5.95 50.81 7.45 3.45123E-12 470.81 3085 c 6.13 755.7 6.04 50.86 7.44- 3.50282E-12 499.55 3087 c 6.13 755.7 6.03 50.65 7.81 3.35896E-12 504.54 3088 c 6.39 755.7 6.04 50.8 9.38 2.78269E-12 521.97 3091 c 6.39 750.8 6.01 50.71 7.26 3.61182E-12 526.44 3092 c 5.55 750.8 5.85 50.72 3.54 7.20048E-12 469.56 3097 c 6.79 750.8 6.05 50.82 12.53 2.0981E-12 553.29 3099 c 6.03 750.8 6.02 50.74 8.00 3.28067E-12 495.37 3101 c 6.32 750.8 6 50.73 3.22 8.13189E-12 521.13 2102 c 5.53 750.8 6.01 50.71 2.49 1.0545E-11 455.59 3107 c 6.23 750.8 5.96 50.84 3.14 8.23913E-12 514.92 3108 c 5.67 750.8 5.92 50.78 2.16 1.19066E-11 472.92 3114 c 4.98 750.8 6.07 50.85 3.55 7.42613E-12 403.99 3115 c 5.61 750.8 6.06 50.84 18.61 1.41323E-12. 456.03 3118 c 6.5 750.8 6.13 .50.8 • 22.53 1.18307E-12: 523.16 3119 c 6.33 750.8 6.11 50.84 40.60 3.49385E-13 510.34 3123 c 5.93 750.8 6.11 50.79 6.40 4.15006E-12 479.03 3126 c 5.26 750.8 5.99 50.82 2.84 9.15178E-12 432.91 3127 c 4.79 750.8 6.03 50.83 2.87 9.13431E-12 391.46 3129 c 5.92 750.8 6.08 50.82 15.38 1.7177E-12 480.02 3131 c 6.23 750.8 6.09 50.77 9.98 2.65612E-12 505.32 3134 c 5.65 750.8 6.05 50.77 3.29 8.01237E-12 461.31 3136 c 5.68 750.8 6.09 50.82 7.26 3.6424E-12 459.80 3139 c 6.26 750.8 6.02 50.59 7.56 3.4923 IE-12 517.32 3141 c 5.54 755.2 6.09 50.69 3.30 8.00908E-12 450.77 3142 c 6.13 755.2 6.06 50.85 4.06 6.44358E-12 498.10 3143 c 5.91 755.2 6.04 50.74 4.64 5.6439IE-12 483.90 3148 c 5.73 755.2 6.1 50.71 4.85 5.45588E-12 465.10 3154 c 5.06 755.2 6.04 50.63 1.96 1.33964E-11 416.11 3157 c 5.62 755.2 6.1 50.31 1.97 1.3679E-11 463.46 3158 c 5.61 755.2 6.05 50.64 2.38 1.1065E-11 460.39 3159 c 6.09 755.2 6.09 50.72 3.37 7.83361E-12 494.94 3056 c 6.96 756 6.25 50.74 2.98 9.08499E-12 550.73 3058 c 7.33 756 6.11 50.81 5.65 4.6715E-12 591.66 120 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

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}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0075068/manifest

Comment

Related Items