A MODEL TO PREDICT TENSILE MECHANICAL PROPERTIES OF ROBOT FORMED WOOD FLAKEBOARD By GUANGQICHEN B.Eng., Nanjing Forestry University, 1985 M.S., Nanjing Forestry University, 1988 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF T H E REQUIREMENTS FOR T H E DEGREE OF MASTER OF APPLIED SCIENCE in FACULTY OF GRADUATE STUDIES FACULTY OF FORESTRY (Department of Wood Science) We accept this thesis as conforming to jjp required^tandard f n T H E UNIVERSITY OF BRITISH COLUMBIA A p r i l , © G u a n g q i 2 0 0 2 C h e n , 2 0 0 2 In presenting degree freely at the available copying of department publication this University for this or of thesis this of reference thesis by in for his thesis partial fulfilment of the British Columbia, I agree and scholarly or for her Department The University of British C o l u m b i a Vancouver, Canada I further purposes gain shall that agree may representatives. financial permission. DE-6 (2/88) study. requirements be It not is be that the for Library an shall permission for granted by understood allowed the advanced make extensive head that without it of copying my my or written 11 Abstract Compared with solid wood, one of the advantages of wood-based composites is the great potential for the design of material properties through manipulation of manufacturing variables. Large strides are presently being made in the design of non-veneer structural panels such as oriented strand board (OSB) by using material science and engineering principles. Scientists and engineers have been more successful in designing synthetic fiberreinforced composites than wood-based composites, mainly because of the complexity of the microstructures and the inherent variability of the wood composites. In this study, recent research in modeling and predicting the properties of flakeboards has been summarized. The relationships among the structure in terms of void volume, density distribution, and the properties of the panels are discussed. With the help of a robotic system, very thin partially oriented wood assemblies were made and tested. The relationship between flake orientation and tensile strength and tensile M O E were determined. The relationship between density and tensile strength and tensile M O E were also examined. Based on layer properties, a three-layer mathematical model was derived to predict the tensile strength and tensile M O E . Partially oriented three-layer OSB panels were made and tested to Ill verify this mathematical model. Very good agreement was found between this model and experimental results. Furthermore, a 3D finite element model was developed to simulate the probability of failure and probabilistic distribution of tensile strength of OSB. The probabilistic distributions of tensile strength and the load capacity probabilistic distributions for threelayer partially oriented OSB were predicted successfully. Good agreement between predictions and experimental data was observed. iv T a b l e o f C o n t e n t ABSTRACT ii TABLE OF CONTENTS iv LIST OF TABLES vii LIST OF FIGURES viii ABBREVIATIONS USED x ACKNOWLEDGMENTS xi CHAPTER 1 RESEARCH RATIONAL AND BACKGROUND 1 1.1 Introduction 1 1.2 Literature Review 4 1.2.1 Density variation 4 1.2.2 Vertical density distribution (VDD) 4 1.2.3 Horizontal density distribution (HDD) 5 1.2.4 Modeling flakeboard properties 8 1.2.5 Flake network model and definition in this study CHAPTER 2 EXPERIMENT MATERIALS AND METHODS 2.1 Materials 12 13 13 2.1.1 Species 13 2.1.2 Flakes 13 2.1.3 Adhesive 14 2.1.4 Single layer partially oriented flake assemblies 14 2.1.4.1 Target assembly density 15 2.1.4.2 Target assembly size 15 2.1.4.3 Actual averaged assembly size 17 2.1.4.4 General mat structure 17 2.1.4.5 Pressing conditions 17 2.1.5 Three layer partially oriented flakeboards 18 2.1.5.1 Target panel density 18 2.1.5.2 Target panel size 18 2.1.5.3 Actual averaged panel size 18 2.1.5.4 General mat structure 19 2.1.5.5 Pressing conditions 19 2.2 Experimental Design 20 2.3 Panel Testing Method 21 2.3.1 Tensile Stiffness and Strength 21 2.3.2 Density 22 C H A P T E R 3 R E S U L T A N D DISCUSSION 23 3.1. Robot-based research on tensile properties of the single layer partially oriented flake assemblies 23 3.1.1 Tensile Stiffness and Strength vs Flake Orientation 23 3.1.2 Tensile Stiffness and Strength vs Density 25 3.2 Mathematical model to predict the tensile stiffness and tensile strength of oriented strand board 34 3.2.1 Calculation of tensile strength and stiffness of layered flakeboard 3.3 Finite element analysis of the tensile strength properties of the 36 vi partially oriented flakeboard 3.4 Conclusions CHAPTER 4 SUMMARY AND CONCLUSIONS 4.1 Future Developments REFERENCE 49 57 59 61 63 vii LIST O F T A B L E S Table 2.1 Flakes dimensions used in the study 14 Table 2.2 Experimental design for single layer partial oriented flake assemblies 20 Table 2.3 Experimental design for three-layer partial oriented flakeboards 20 Table 3.1. Partially oriented flake assembly properties 27 Table 3.2. Partially oriented flake assembly properties change trend (density = 0.5/cm ) 28 3 Table 3.3. Layer density information of three-layer partially oriented flakeboards with 0.7:0.3 as the ratio of face layer to core layer 40 Table 3.4. Regression results for tensile strength and tensile M O E of panel at 45 orientation range 40 Table 3.5. Prediction of tensile strength and stiffness (average density=0.5 g/cm ) 44 3 Table 3.6. Prediction of tensile strength and stiffness (average density =0.55 g/cm ) 45 Table 3.7. Prediction of tensile strength and stiffness (average densty=0.60 g/cm ) 46 3 3 Table 3.8. Tensile strength and stiffness of three-layer partially oriented flakeboards with average density of 0.55 g/cm and 0.60 g/cm 3 Table 3.9. The comparison of the predicted and experimental results 3 47 48 viii LIST OF FIGURES Figure 2.1. 3D layout of laboratory-controlled robotics system 16 Figure 2.2. The exact displacement measured with MTS extensometer 21 Figure 3.1. Relationship between tensile strength and orientation angle range 29 Figure 3.2. Relationship between tensile stiffness and orientation angle 29 Figure 3.3. Relationship between n and tensile strength error 30 Figure 3.4. Relationship between n and tensile M O E error 30 Figure 3.5. Comparison of Hankinson's equation and tensile strength experiments 31 Figure 3.6. Comparison of Hankinson's equation and tensile M O E experiments 31 Figure 3.7. Relationship between density and tensile strength 32 Figure 3.8. Relationship between density and tensile M O E 33 Figure 3.9. A typical V D D plot reflecting the higher density face layers, lower density core layer and the nearly symmetrical density profile about the panel mid-depth. Figure 3.10. Three-layer symmetric orthotropic laminate 35 36 Figure 3.11. Three-layer symmetric orthotropic laminate under uniform axial stress in the x direction i 37 Figure 3.12. Finite element mesh of a three-layer flake assembly 50 Figure 3.13. Cumulative probability of tensile strength of flake assemblies 51 Figure 3.14. Cumulative probability of tensile M O E of flake assemblies 51 Figure 3.15. Relationship between experimental tensile M O E and tensile strength of thin assemblies 53 Figure 3.16. Cumulative probability distribution of simulation and experiment result for tensile strength of three-layer partially oriented flakeboards Figure 3.17. Cumulative probability distribution of simulation and experiment result for tensile M O E of three-layer partially oriented flakeboards Figure 3.18. Relationship between experimental tensile strength and tensile M O E of three-layer partially orientated flakeboards Figure 3.19. Relationship between simulation tensile strength and tensile M O E of three-layer partially orientated flakeboards ABBREVIATIONS USED H D D — horizontal density distribution V D D — vertical density distribution M O E — modulus of elasticity M O R — modulus or rupture OSB — oriented strand board IB — Internal bond L V L — laminated veneer lumber TS — thickness swelling F E A — finite element analysis xi ACKNOWLEDGMENT I would like to thank my supervisor Dr. Frank Lam and Dr.Stavros Avramidis, Faculty of Forestry, UBC, for their invaluable advise, supervision, help and patience throughout this research project. Also, gratitude is extended to Dr. J.D. Barrett for reviewing and providing feedback on the thesis while serving on the supervisory committee. I would like to thank Bob Myronuk and Avtar Sidhu for their laboratory assistance. Help from Dr. Liping Cai, Dr. Congjing Lu and George Lee is also readily acknowledged. Financial support from Natural Sciences and Engineering Research Council of Canada, Forestry Canada, Forintek Canada Corp. and MacMillan Boedel Co. is gratefully acknowledged. Here, I would also like to give my thanks to Ainsworth Lumber Co. for supplying the logs and to C A E Machinery Ltd. for preparing the flakes. Finally, my greatest gratitude goes to my parents and my wife, Dana Zhu for their love, care, patience, encouragement and understanding during my long educational studies. CHAPTER 1. RESEARCH RATIONAL AND BACKGROUND 1.1 Introduction Among the four basic building materials in the world (plastic, concrete, steel, and wood), only wood is renewable. Wood physical properties frequently exhibit a wide degree of variability. This variability is in part the result of the free growth conditions brought by environmental factors such as climate, soil, water supply etc. In addition, all properties of wood are in part heritable and consequently, a substantial portion of its natural variability can be attributed to genetic stock differences (Bodig et al. 1982). Although we have little or no control on the engineering properties of solid wood except through grading, the potential for the design of material properties in wood-based composites is great. Large strides are presently being made in the design of non-veneer structural panels such as Oriented Strand Board (OSB) by using material-science and engineering principles (Bodig et al, 1982; Geimer, 1981; Hunt, 1974; Triche, 1993). Because wood is the basic component of all wood-based composites, it is logical to expect that to some degree, the fundamental characteristics of wood will be retained by them. However, the properties of these composites are influenced by the manufacturing process and the properties of various added materials. It is therefore very difficult to predict the properties of those without a basic understanding of the influence of its interacting variables on product performance. Scientists and engineers have been more successful in designing synthetic fiber-reinforced composites than wood-based composites, mainly because of the complexity of the microstructures and the inherent 1 variability of the latter. However, proficiency in designing wood-based panels is necessary and mandatory if these products are to successfully compete with other materials in current and future markets. Recently, emphasis has been placed on the relationship between horizontal density distribution (HDD), vertical density distribution (VDD) and structural characteristics of wood composite mats. To achieve this, mathematical models along with computer simulation, robot mat formation and X-ray scanning techniques have been used (Lu et al, 1998; Oudjehane et al., 1998). Knowledge gained from the combination of mathematical models, simulation, predefined mat structures made by robots, and quantitative data on density distribution in commercial wood composites scanned by X-ray machines, can be used to predict and design the performance properties of existing wood composites, guide future improvements in mat forming technology, and provide insights into the development of future structural wood composites. Very few studies have been carried out on the relationship between mat structure in terms of H D D and the mechanical properties of the panel such as stiffness or modulus of elasticity (MOE) and strength or modulus of rupture (MOR). The overall objective of this project is to develop a basic understanding of how the internal structure of the flake-based composite mats in terms of varying flake position, orientation and density distribution affects the tensile strength and stiffness. It is expected that this study will contribute to the characterization of the potential structure of commercial panels and establish a database for modeling the panel's mechanical properties. Furthermore, this study will help with the development of strategies that 2 improve commercial forming practices and control the properties and qualities of the endproducts. Specifically, this project's objectives are: 1) to manufacture mats formed by a robot and develop a model to systematically describe the relationship between structure of multi-layer flake assemblies (4 mm in thickness) and tensile stiffness and strength, and to calculate the structural properties of the mat; 2) to develop a model to systematically describe the relationship between partial OSB structure and tensile MOE and tensile strength, and simulate its properties based on the above model; 3) to manufacture OSB panels formed by a robot to verify the results from the model and compare with the simulation; 4) to verify the model using finite element analysis. 3 1.2. Literature Review 1.2.1. Density variation Wood composites can be classified into two categories: veneer based products, such as plywood and laminated veneer lumber (LVL); and non-veneer products or short fiber composites such as particleboard, fiberboard, and OSB. Short fiber wood composites consist of wood particle inter-dispersed with voids in a manner that results in density variation within the volume of a panel. Within a flakeboard, the density distribution can be further subdivided into vertical and horizontal components. This type of threedimensional density distribution is correlated with the internal mat structure (Suchsland, 1967; Suchsland and X u , 1989). 1.2.2. Vertical density distribution (VDD) The density of a mat-formed and hot-pressed wood particleboard or flakeboard is not uniform in the thickness (vertical) direction. This density profile is highly dependent upon the particle configuration, moisture distribution in the mat entering the press, rate of press closing, temperature of the hot press, reactivity of the resin, and the compressive strength of the wood particle components. The panel's V D D substantially influences many strength properties. It is generally known that the steep " U " shaped gradient is beneficial to the M O E and M O R of the particleboard, and a more uniform gradient is beneficial to the internal bond (IB) of the panel. Therefore, depending on which properties are most critical in the ultimate use of the panel, modifications of the resin type and pressing step may be justified either to enhance or to restrict the formation of this 4 gradient. In the past 30 years, much emphasis was placed on the pressing parameters and characteristics of the wood element constitutents. Considerable progress has been and is still being made in understanding how the pressing parameters influence panel properties via the formation of a vertical density distribution (Engels 1980; Geimer 1975; Hansel et al, 1988; Harless, 1987; Heebink et al, 1972; Kelly 1977; Maloney 1970; Shen et al, 1970; Smith 1980; Strickler 1959; Suchsland 1962; Suo et al, 1994; Thomas et al, 1987; Winistorfer et al, 1996). 1.2.3. Horizontal density distribution (HDD) In an ideal mat, short fiber elements such as strands, flakes, or wafers would be arranged in such a manner that continuous layers are formed with no voids. However, in a commercial mat, each layer would be comprised of certain quantity of wood with an inter-dispersed void volume. If small samples were cut from different locations within a flakeboard in the horizontal plane and their densities were measured, it would be found that these density values could vary to a smaller or larger degree about their mean value, namely the overall panel density. Suchsland (1967) was the first to recognize this variation and named it the "horizontal density distribution". He theorized that HDD was binomially distributed and concluded that a particleboard mat was essentially a series of veneers, and the difficulty in achieving good glueline bonding in particleboard was then similar to the difficulty in achieving good gluelines in plywood. In a later study, Suchsland and X u (1989) tested idealized flakeboards to demonstrate the effect of H D D on panel performance. This density variation had a direct effect on both IB and the thickness swelling (TS) of idealized mats. 5 Steiner and X u (1995) used wood composites prepared with a series of precisely cut wood flakes to investigate the influence of raw material characteristics on HDD. The effect of flake size on H D D was dependent on specimen size. Two macrovoid aspects, namely, frequency and size, were identified as being responsible for this dependence. In addition, a layer concept was developed to relate flake thickness, panel thickness, wood density, and panel density to HDD. Panel uniformity was found to improve as the number of flake layers increased. Considering wood flake mats or flakeboards as a number of vertically stacked layers, Steiner and Dai (1993) presented a mathematical model for describing the structure of a randomly formed flake layer network. The structural properties of the flake network are random variables in essence characterized by the Poisson or exponential distribution. The model predicts distribution of flake centers, flake coverage, free flake length, and void size over the flake layer network. A computer program for simulating the random flake network and rapidly evaluating structural properties was developed as well. In addition, the structure of hand-formed flake layers was experimentally measured. Close agreement was found between mathematical model prediction, computer simulation and experimental measurements. The structure of a randomly formed flake-type wood composite mat was further defined and characterized by Dai and Steiner (1993, 1994). A model for the prediction of H D D in a random flake mat was presented through application of two-dimensional random field theory. This model predicts the small-scale mass density variance and the spatial correlation of flake coverage. The predictions agreed well with experimental results and computer simulations. The significance and implications of the model development towards practical manufacturing applications were discussed. In addition, equations for the calculation of general structural properties such as overall mat thickness, macrovoid volume content and maximum potential inter-flake bonded areas were presented. It is noted that no resin and additives were added to the panel and a hot pressing operation was not used. As a result, the spatial structure of the experimental panels was not representative of a commercial panel. Also the above mentioned theory dealt only with completely random orientation. Compared to V D D , only limited research has been done on how H D D influences the properties of flakeboard such as tensile strength properties. To better understand the nature of the structures of flakeboard mats, a simulation program Winmat, based on the Monte Carlo technique, has been written (Lu et al, 1998) to compute the horizontal distribution of overlap and density, free flake length and its distribution, number of flake crossings, the location and distribution of void sizes, the autocorrelation function, variance function and the degree of orientation of flakes in both simulated mats and experimental mats. This program can also determine the effect of sampling zone sizes on the density/overlap distribution. Using this program, a study of the relationships between thickness swelling and mat structures in robotformed flakeboard mats made without wax was conducted (Lu, 2000) under 95% and 90% relative humidity conditions and 24-hour water soaking tests. A model describing such relationships was established for two relative humidity conditions. With this model the thickness swelling of flakeboard mats (without wax) can be predicted, provided that the amount of moisture absorbed and the density distribution of the mat are known. 7 1.2.4. Modeling flakeboard properties In recent years, the development and commercialization of wood composite materials such as OSB and flakeboard is increasing at an accelerated rate, driven by both environment, resource and market factors (Bodig et al, 1982; Bowyer 1995; Hoover et al, 1992; Hunt 1974). However, the development of currently available structural wood composites has largely resulted from costly trial-and-error based laboratory experiments. An analytical model capable of predicting the structural properties of a wood composite material based on the properties of its ingredients would be an invaluable tool. The real value of modeling lies in its capacity to simplify description of properties and to identify the common characteristics of different composites (Bodig et al, 1982). Hunt and Suddarth (1974) developed a finite element model for random orientation flakeboards. The model consisted of a rectangular grid of rigidly connected frame elements to model the adhesive, and four noded plane stress elements placed in the rectangular openings to model the wood. The model was used to predict the tensile M O E and shear modulus of small specimens. Errors were within 13 percent for all boards tested. Hoover and Hunt (1992) used a regression equation to predict the properties of mixed species, single-layer, oriented strand panels. The first step was to produce a set of singlelayer panel for each species. The regression equations developed to predict bending M O E and M O R , edge-wise shear modulus and strength, and tensile modulus and strength, had an overall coefficient of determination of 0.63. The next step was to develop predictive equations for mixed species boards and produce a set of mixed species panels to test 8 these equations. The single-species equations were combined using the rule-ofmixtures—the mixed species panel properties predicted to be the weighted average of the properties of the single-species panel. On the average, bending M O E and M O R prediction varied less than 5 percent in the aligned and 17 percent in the cross-aligned directions from the measured mixed species panel properties. Edgewise shear modulus and strength predictions varied less than 17 percent overall from the measured data. Adequate prediction of complex properties such as IB and edgewise shear modulus may not be possible with this type of empirical approach. Triche and Hunt (1993) developed a theoretical model to predict the tensile properties (tensile stiffness and strength ) of perfectly oriented strands. Model development was aided by experimental data obtained from testing individual strands and small strand assemblies. The model, in computer form, incorporates a special finite element scheme used to model the strand and resin of the composite, and uses a substructuring routine to take advantage of the composite's repeating nature. Experimental boards were manufactured and tested to check the validity of model. Based on the their experimental results, the model successfully predicted the tensile stiffness and strength of the composites. In their study, the strand dimensions were 152.4 by 12.7 by 8.5 mm in length, width and thickness, respectively, and the specimen were 457.2 mm long, 38.1 mm wide, and only 3 layers thick. Strands were edgewise and endwise within each layer, and were staggered half of their width in adjacent layers. Stress updating and size effect adjustment procedures were not used in this model. Wang and Lam (1996) developed a three-dimensional nonlinear stochastic finite element model with size factor to model and predict the probabilistic distribution of 9 tension strength of parallel-aligned wood composites based on the experimental data of single veneer and multiple ply veneer assemblies. The relationship between the strength modification factor a and the number of layers of multiple ply veneer assemblies of different element lengths was established. Based on this relationship, the calibrated model could be used to predict the probabilistic distribution of tension strength of parallel-aligned wood composites with various sizes of finite element mesh and different number of layers. The influence of grain angle was also considered in their later model (Wang and Lam, 1997). Shaler and Blankenhorn (1990) incorporated variables into a model, which included flake geometry, flake orientation, density, resin content and species. Verification of the model was accomplished by comparing predictions with flexural M O E values measured parallel and perpendicular to the flake alignment direction of 192 specimens made from two species at 4 resin levels (3, 5, 7 and 9 percent) and 3 target densities (0.56, 0.672 and 0.832 g/cm ). Use of the longitudinal Halpin-Tsai equations in conjunction with 3 measured and estimated wood and resin properties predicted the M O E of aspen and red maple flakeboard specimens with a standard error of estimate of 0.65 MPa, a coefficient of determination of 0.9, and an average of 25 percent below experimental values. The statistical correlations were influenced by grouping of data using flake alignment direction and species. The approach was easily computed and gave reasonably accurate predictions of a single oriented flakeboard M O E manufactured over a range of resin levels, densities, flake alignment and flake geometry. Geimer (1975) realized that few attempts have been made to establish the relationship between physical and mechanical properties and layer densities. Later, Geimer (1979) established the relation between specific gravity (density) and the mechanical properties of reconstituted wood flakes. Flakeboards made with uniform densities throughout their thickness and different degrees of flake alignment were used to establish empirical relationships between bending, tension, and compression values of M O E or M O R (or stress to maximum load) and the variables of specific gravity and flake alignment. An equation using sonic velocity as an indicator of alignment was developed that describes the relations over a broad range with a high degree of confidence. Bending stiffnesses of boards having a density gradient were predicted within approximately ±20 percent using the derived relationships. In his experiment, the cold pressing method was used to obtain uniform density in the direction panel of thickness. Furthermore, he used the data as a base to predict flakeboard properties (Geimer, 1981). Regretfully, when the hot pressing method was used in his later research, the difference existed in V D D between the cold pressed boards and the hot pressed boards cannot be directly explained. In early 1990s, a large research program was initialized at the Department of Wood Science, The University of British Columbia, Canada, designed to bring a unified approach for better understanding, effective modeling and simulation of the complex relationship between the structure and properties of wood-based composites with the help of robot systems. A model for the variation of engineering properties within random composite panels was developed by Oudjehane and Lam (1998a): This model is based on a simple cost effective experimental program of uniaxial tests. It enables prediction of engineering properties of wood flakeboards using the same structure, but manufactured according to a different requirement of density or thickness. 11 A nonlinear continuum mechanics-based model has also been developed by Oudjehane and Lam (1998) to predict the physical and mechanical properties of flakeboards, taking into consideration the formation and hot-pressing processes. The developed viscoelastic model of a wood mat's consolidation during hot-pressing relates pressing strategy parameters (applied pressure, press closing time, etc.) and forming process parameters (initial mat thickness, void volume, etc.). The history of the deformation induced is then introduced to the model that predicts wood composite engineering properties. This model, developed for engineering properties of flakeboards and structural panels, allows prediction of the variation of the above properties within the principal plane of the panels. Based on the developed model, which accounts for the number of flakes deposited at any location of the mat, given that flakes are defined by their centers' coordinates, the inplane M O E is simulated for the random placement of flakes within oriented structural panels. The simulation clearly displays the strong variation of the M O E along the principal plane of the panel. 1.2.5. Flake network model and definition in this study This study assumes that: 1. the compression behavior (normal to the flat surface of the flakes) for all flakes is the same because only one wood species is used in the flake-type mat and all the flakes used are free from defects with the same physical and mechanical properties, so that; 2. all the flakes can be considered as flat before pressing; 3. there is no vertical density gradient in the multi-layer (5-7) flake assembly for the very thin thickness (4 mm in thickness). 12 CHAPTER 2. EXPERIMENT MATERIALS AND METHODS 2.1. Materials 2.1.1 Species Trembling aspen (Populus tremuloides, density 0.405 g/cm based on the over-dry wood weight) was used for this study. The trembling aspen logs were obtain from Ainsworth Lumber Co. at 100 Miles House, B C . 2.1.2 Flakes Laboratory flakes were cut with a laboratory-type-waferizer located at C A E Machinery Ltd. in Vancouver, B C . Relatively flat flakes without breaking were selected visually. The flakes were air-dried for a week before gently dried in a laboratory oven at temperature of 60°C for 3 hours to maintain the flatness of flakes. The dried flakes were stored in plastic bags to maintain the moisture content. The equilibrium moisture content of the final dried flake was 4-6 percent. Dimensions of flake used in this study are provided in Table 2.1. 13 Table 2.1 Flakes dimensions used in the study (SD = standard deviation) Average Length(mm) 102.3 Sample size 100 SD 3.1 Average Width (mm) 19.2 Sample size 100 SD 0.82 Average Thickness (mm) 0.742 Sample size 100 SD 0.05 2.1.3 Adhesive Phenol-formaldehyde powder resin C A S C O P H E N W 9IB was used throughout this study. This resin has been used both for face and core applications in commercial waferboard/OSB manufacturing. 2.1.4 Single layer partially oriented flake assemblies Basic procedure for manufacturing single layer partially oriented flake assemblies in the laboratory were as follows. Firstly, a predetermined amount of wood flakes and adhesive were mixed in a plastic bag (hand shaking very gently to minimize flake breakage). A resin content of 6 percent, based on the oven dry wood weight, was used to ensure full resin coverage on the flakes and compensate for resin lost during flake blending and mat forming stage. Pilot research was performed to verify the reliability of using this resin level. Then, a laboratory-controlled robotics CRS system was used for a 3D layout of the flakes (see Figure 2.1). The robot-controlling program enabled flakes to be laid out according to their orientation and center coordinates within the principal plane of the mat. After that, the mat was compressed and consolidated into a thin assembly under heat and pressure in the laboratory hot-press. After manufacturing, all assemblies were stored for equilibration in a conditioning room, with temperature controlled at 20±2°C and relative humidity at 50+3 percent for at least 10 days before testing. Afterwards, they were cut into the size of 190 x 20 mm. The details about assembly structures and manufacturing conditions are described in the following section. 2.1.4.1 Target assembly density Three target assembly densities of 0.50 g/cm , 0.55 g/cm and 0.60 g/cm are 3 3 3 considered in this study based on the calculation from simulation program. Assembly mat size is 250 x 250 mm and actual averaged assembly density is based on weight and volume at 7.8 percent moisture content. 2.1.4.2 Target assembly size Assembly size was targeted at 250 x 250 x 4 mm (length x width x thickness). 15 16 2.1.4.3 Actual averaged assembly size The actual averaged assembly size was achieved at 250 x 250 x 3.84 mm (length x width x thickness). 2.1.4.4 General mat structure Single layers with partial orientation range of 0°, 15°, 30°, 45°, 60°, 75° and 90° (random) are investigated in the experiment. Partial orientation range means that the average flake angle is still 0° and the flake angles are uniformly distributed between the plus (+) partial orientation range and the minus (-) partial orientation range, e.g. partial orientation range of 30 ° means that flake angles vary from -30 ° to 30 ° and the average angle is 0°. 2.1.4.5 Pressing conditions: Press equipment: Computer controlled 300 x 300 mm hot press (Wabash 27,000 kg load) Press temperature: 180 °C Press closure time: 21 seconds Pressing time: 210 seconds Press open time: 15 seconds Target thickness: 4 mm No wax was added. Preliminary experiments were performed to select the above pressing parameters to ensure proper bonding within each kind of mat. 17 2.1.5 Three layer partially oriented flakeboards Basic procedure for manufacturing three layer partially oriented flakeboards in the laboratory is nearly same as above for the assemblies except for longer hot pressing time. After manufacturing, all assemblies were stored for equilibration in a condition room, with temperature controlled at 20±2°C and relative humidity at 50+3 percent for at least 4 weeks before testing. Afterwards, they were cut into the size of 190 x 20 mm. The details about assembly structures and manufacturing conditions are as follows: 2.1.5.1 Target panel density Target panel densities are 0.55 g/cm and 0.60 g/cm (based on the calculation from 3 3 simulation program with mat size 250 x 250 mm, actual averaged panel density is based on weight and volume at 7.8 percent moisture content). 2.1.5.2 Target panel size Panel size was targeted at 250 x 250 x 12 mm (length x width x thickness). 2.1.5.3 Actual averaged panel size The actual averaged panel size was achieved at 250 x 250 x 11.8 mm (length x width x thickness). 18 2.1.5.4 General mat structure Symmetrical three layer with partial orientation Face and bottom layer orientation: 15° Core layer orientation: 75° and 90° (random) Ratio of face and bottom layer to core layer: 0.7:0.3 2.1.5.5 Pressing conditions Press equipment: Computer controlled 300 x 300 mm hot press (Wabash 27,000 kg Load) Press temperature: 180 °C Press closure time: 25 seconds Pressing time: 480 seconds Press open time: 15 seconds Target thickness: 12 mm No wax was added. Preliminary experiments were performed to select the above pressing parameters to ensure proper bonding within each kind of mat. 2.2 Experimental Design Preliminary experiment and other researchers' results (Oudjehane and Lam, 1998; Lu, 1998; Wang, 1997) have showed that the robot-controlled forming system gives much higher precision than a hand forming system, so the experimental designs shown in Tables 2.2 and 2.3 were used to minimize costs and save the time. 19 Table 2.2 Experimental design for single layer partial oriented flake assemblies (+ means this type of assembly was made) \v Angle 0° 15° + + 30° 45° 60° 75° 90° + + + + DensitjN. 0.5 g/cm 3 0.55 g/cm 0.6 g/cm + 3 + 3 Two replicates are applied to each type of structure. Thus, a total of 18 assemblies were made. Table 2.3 Experimental design for three layer partial oriented flakeboards (+ means this type of panel was made) ~^~~~-^JFace and bottom layer orientation —~-_i_Core layer orientation 15° : 90° 15° : 75° + + + + Density 0.55 g/cm 0.6 g/cm 3 3 The reason for choosing the above structure is that those types are quite similar to the commercial panel's. Two replicates are applied to each type of structure. Thus, a total of 8 panels were made. 20 2.3 Panel Testing Method 2.3.1 Tensile Stiffness and Strength The tension tested specimens were cut into the size of 190x20 mm (length x width). Then they were tested under a gauge length of 60 mm. A 222 k N MTS machine with a mechanical wedge-type grip system was used to perform tensile testing under a constant loading rate of 1.0 mm/min to obtain the tensile M O E and tension strength parallel to the grain. In order to minimize the effect of testing sample sliding within the grip, the measurement sensor — MTS Extensometer was used to measure the exact displacement as shown in Figure 2.2. The loading value and the corresponding displacement were measured. gnp /Extensometer • • — — •> a • • <=— L + A L —=> 60 mm 190 mm -—^ — ..... Figure 2.2. The exact displacement measured with MTS extensometer. A digital caliper was used to measure the cross-sectional areas of specimen with an accuracy of ±0.01 mm. A nominal tensile strength (MPa) was calculated as Tensile strength = P /Area, max 2 1 where, P m a x represents the maximum failure load (N); A represents the cross section area of specimen (mm ). 2 The calculation of nominal tensile M O E is as follows: CT P L M O E = — =— x — £ A AL (1) where, P represents testing load (N); L represents original extensomter length (mm); A L represents the corresponding increment in length; A represents the cross section area of 2 P specimen (mm ); — is the slope of the linear portion of the load deformation curve AL obtained from the clip gauge. The data from specimens that failed in the grip area were discarded. There are 180 specimens cut from single layer assemblies and 64 specimens cut from three layer partial oriented flakeboards. 2.3.2 Density After the tensile M O E and M O R tests were carried out, density samples were cut from the two ends of the tested samples into the size of 20 x 20 mm. The actually length, width and thickness were measured by a digital pair of calipers with an accuracy of +0.01 mm and the weights were determined on an electronic digital balance of +0.001 g accuracy. The densities (based on weight and volume at 7.8% moisture content) were calculated as density = Weight/(LengthxWidthxThickness). There are 180 specimens cut from single layer assemblies and 64 specimens cut from three layer partially oriented flakeboards. CHAPTER 3. Results and Discussion 3.1. Robot-based research on tensile properties of the single layer partially oriented flake assemblies 3.1.1 Tensile Stiffness and Strength vs Flake Orientation Ranges In the past years, considerable research has been carried out on how flake orientation influences the OSB's mechanical properties, but nearly all of them used the "pure orientation" method, namely, all flakes in the panel were oriented in one or two specific directions. With the help of the robot system, partially oriented flake assemblies in this study give us more details and can very closely represent commercial OSB panels. Summaries of the experimental results for tensile strength and tensile M O E are shown in Table 3.1. The tensile M O E and strength relationships with the partial orientation angles are shown in Figures 3.1 and 3.2, respectively. It is assumed that there is no vertical density distribution in these very thin assemblies. From the results, we can see that the orientation range of flakes has a direct impact on tensile strength and M O E . At a given density, the smaller of the orientation range, the higher of the tensile strength and M O E . The average tensile strength and M O E values for 0° orientation range at density of 0.5 g/cm are 30.25 MPa and 13438.56 MPa 3 respectively, while the average tensile strength and M O E values for 90° orientation range (random) are 8.01 MPa and 4364.9 MPa respectively. The ratios of average tensile strength and M O E values for 0° orientation range at density of 0.5 g/cm to those for 90° 3 23 i orientation range are 3.78 and 3.08, respectively. This further verifies the research results by Geimer (1975) and Kelly (1977). In those research papers, it was concluded that flake alignment in OSB/Waferboard served to increase panel's tensile stiffness in the direction of alignment to 2.5 to 3.5 times that of random configuration. However, the values in the current study will give us more accurate data about how flake orientations play the role in determining panel's mechanical properties. The reason is that partially orientated panel formed by robot can very closely represent commercial OSB panel's structure. It can also be observed from Figures 3.1 and 3.2 that the average tensile strength and M O E values decreased dramatically when the orientation was changed from 0° to 30° and then decreased gradually after that. The ratios of average tensile strength and M O E values for 30° orientation at density of 0.5 g/cm to those for 90° orientation are 2 and 3 1.57 respectively. Hankinson's equation (Bodig et al. 1982) as follow also shows similar results while using some of experimental results. N = _ PxQ Pxsin"t9 + <2xcos"f9 where: P = the property at 0°; Q is the property at 90°; N = the property at angle 0 . The following equation is used to determine the optimal n value to minimize O value. ^ -A Hankinson.; — Exp, , (Minimum) O = Y ( —^-f / Exp, s 1 s (3) 24 where: i is from 1 to 5 corresponding to 15°, 30°, 45°, 60° and 75°; Hankinson i represents Hankinson's equation value at a specific degree; Exp is the experimental t result at specific degree. As shown in Figures 3.3 and 3.4, tensile strength and tensile M O E errors are minimized when n is 2. The comparisons of Hankinson equation and experiments at n=2 are shown in Figures 3.5 and 3.6. When n=2, Hankinson equation values match experiments very well especially from 45° to 75°. 3.1.2 Tensile Stiffness and Strength vs Density The density of panels has a direct impact on most of the properties of the final product, particularly, mechanical properties such as stiffness and strength. Table 3.1 also summarizes the relationships between tensile strength and tensile M O E and density. We can see at a given orientation — 45°, the average tensile strength and M O E values for density of 0.5 g/cm are 14.87 MPa and 6556.38 MPa, respectively; and 3 the average tensile strength and M O E values for density of 0.55 g/cm and 0.60 g/cm are 3 3 15.57 MPa, 7003.41 MPa and 16.43 MPa, 7284.77 MPa, respectively. The ratios of average tensile strength and M O E values for density of 0.55 g/cm to 3 those for density of 0.50 g/cm are 1.06 and 1.07, respectively. Similarly the ratios of 3 average tensile strength and M O E values for density of 0.60 g/cm to those for density of 3 0.55 g/cm are 1.06 and 1.05, respectively. 3 25 From Figures 3.7 and 3.8, at a given orientation range, linear relationships between tensile stiffness and density and between tensile strength and density can be established from regression analysis. These results are close to the ones reported by Geimer (1981) and Kelly (1977). Even though the relationships are fairly weak, there exists a trend of increasing the panel's density to improve mechanical properties such as stiffness and strength. However, there is density limit and also disadvantages to do that. First, it increases product costs for more wood and adhesives are added into the panels; second, it increases transportation costs; and finally it causes production cost and difficulty for heavily compressed wood requires longer hot pressing time. 26 Table 3.1. Partially oriented flake assembly properties (SD = standard deviation, CV = coefficient of variation) Panel Category MC (%) Tensile Strength Tensile MOE Average (MPa) SD (MPa) CV 5.78 30.25 9.15 15° 5.77 21.81 30° 5.83 45° Average Density (g/cm ) Average (MPa) SD (MPa) CV 0.3 13438 3094 0.23 0.487 6.92 0.317 9562 2581 0.27 0.493 16.1 5 0.31 7710 1976 0.26 0.490 5.79 14.87 4.46 0.3 6556 1809 0.28 0.493 60° 5.88 10.59 2.98 0.281 4978 1285 0.26 0.492 75° 5.81 9.05 2.74 0.302 4716 1249 0.27 0.498 90° 5.81 8.01 2.28 0.285 4365 1261 0.29 0.504 5.9 15.57 4.68 0.3 7003.41 1846 0.26 0.556 5.84 16.43 4.89 0.297 7284.77 1819 0.25 0.603 Density=0.5 g/cm 'i 3 Angle 0° Density=0.55g/cm 3 Angle 45° Density=0.6 g/cm 3 Angle 45° 27 Table 3.2. Partially oriented flake assembly properties change trend (density = 0.5/cm ) 3 Decrease (%) Tensile M O E (MPa) Decrease (%) Angle Tensile Strength (MPa) 0° 30.25 15° 21.81 39 9562 40 30° 16.1 35 7710 24 45° 14.87 8 6556 17 60° 10.59 75° 9.05 90° 8.01 13438 40 17 4978 4716 13 31 5 8 4355 28 Figure 3.1 Relationship between tensile strength and orientation angle range 16000 Tensile MCIE (MPa) 14000 < • 12000 i 10000 8000 6000 I 4000 2000 0 0 i i i i i i i i i i 15 30 45 60 75 90 Partial Orientation Range (degree) Figure 3.2 Relationship between tensile M O E and orientation angle range 0 -0.5 H> l Figure 3.3 Relationship between n and tensile strength error Figure 3.4 Relationship between n and tensile M O E error ••— Experiments H Hankinson 0 0 20 40 60 80 100 Partial Orientation Range (degree) Figure 3.5 Comparison of Hankinson's equation and tensile strength experiments -•— Experiments ct ^ - Hankinson 15000 10000 o 5000 c H 0 20 40 60 80 100 Partial Orientation Range (degree) Figure 3.6 Comparison of Hankinson's equation and tensile M O E experiments 31 Figure 3.7 Relationship between density and tensile strength 11000 y= 11383x+ 1247.1 R = 0.5301 2 10000 ea w 9000 8000 o 7000 c H 6000 H 5000 A 4000 0.45 0.5 0.55 Density 0.6 0.65 (g/cm3) Figure 3.8 Relationship between density and tensile MOE 33 3.2 Mechanics model to predict the tensile stiffness and tensile strength of oriented strand board As mentioned in the Chapter 1, it is generally accepted with adequate evidence, that M O E increases with panel density when all other factors (flake size, orientation, species, adhesive content, pressing conditions, etc.) are constant. The surface density strongly affects bending M O E because bending stresses and strains are higher at surfaces. Consequently, bending M O E values are highly dependent upon V D D . Different M O E values can be obtained for equal average board densities but with different V D D (Kelly, 1977). However very limited research has been carried out on how V D D influences the tensile M O E . Figure 3.5 is a typical V D D plot reflecting the higher density face layers and the lower density core layer. One can assume that the V D D is symmetrical about the panel mid-depth. Based on available information on V D D , the ratio of average density of the face layer over that of the core layer can be established. Usually, this ratio ranges from 1.1 to 1.4 (Geimer, 1975; Kelly, 1977; Harless, 1987; Xu, 1994). In this model, a discrete threedensity zone, symmetrical about the specimen mid-depth, will be assumed along the vertical direction. 34 / \ d K si; M id-depth Figure 3.9 A typical VDD plot reflecting the higher density face layers, lower density core layer and the nearly symmetrical density profile about the panel middepth. In Section 3.1, experiments on the thin layer (4 mm) partially oriented assemblies were performed and a database on the tensile stiffness and strength was set up to provide input on the mechanical properties of individual layers for the current model. 35 3.2.1 Calculation of tensile strength and stiffness of layered flakeboard A schematic view of three layer symmetric orthotropic laminates (face and bottom layers are identical) is shown in Figure 3.6 A x2 Figure 3.10 Three-layer symmetric orthotropic laminate If a uniformly distributed tensile strain is applied to the laminate in the X i direction (Figure 3.7), the bonding adhesive constrains all laminae to deform the same amount. From compatibility consideration, one has ei=e, = f c e i (4) where: £i = strain of the laminate in X i direction £i = strain of the face layer in X i direction f £i° = strain of the core layer in X i direction 36 Figure 3.11 Three-layer symmetric orthotropic laminate under uniform axial stress in the x, direction The stresses in the face and core are given by Hooke's Law a, =E, ei f ai =E, f c ; (5) c e i where: a/ = stress in the face layer (MPa) o~i° =stress in the core layer (MPa) The cross-sectional area of the laminate A\ (area normal to the 1 direction) is the product of specimen width w (mm) and thickness t (mm): Ai= wt - (6) Similarly, the cross-sectional areas of the face laminae and core are given by A , = 2wt and A i = wt f c c (7) From equilibrium, the total force P on the laminate is defined by P= 2 P + P , f c (8) where: P = force applied on the face layer f P = force applied on the core layer c But p = A,=ai A! +ai A, f f c (9) c ai For the entire laminate one has the nominal stress as 0-1 = E,e, (10) substituting Eqs.(lO) and (5) into Eq. (9) gives EieiAi = E , e A f 1 f 1 + E, e A c 1 (11) c 1 Rearrangement leads to an equation for E i , the modulus of elasticity of the laminate in the X i direction: Ej= E / ( A / / A , ) + E ^ C A ^ / A O (12) Since the width of all laminae is the same, Eqs. (6) and (7) can be substituted into Eq. (12) to give a modified version for Ej: Ei= - ( 2 E ! t + E i t ) t f f c (13) c From Eq. (9), we can get Oi=Oi (A /Ai) + Oi (A f f c 1 c 1 /A ) (14) 1 From Eq. (5), we can get C^CEfVErW (15) a^CEfVEfW (16) substituting Eq.(16) into Eq. (14) gives a!= (E, / E, ) 0! (A, / AO + a, (A, / A,) f c c f c c (17) substituting Eq.(17) into Eq. (14) gives 38 CJi=o-i ( A / / A , ) + ( E , 7 E, ) a , ( A , / A , ) f f (18) 0 f So, the actual stress G i w i l l be the minimum of E q . (17) and E q . (18). Using the mean tensile M O E and tensile strength information of the thin partially oriented assembly, Eqs. (13), (17) and (18) are used to predict the mean the tensile M O E and tensile strength of three-layer partially oriented flakeboards. For the three-layer partially oriented flakeboards with different layer density and layer thickness, the average board density is: Paverage = Pcore ( x where: ~ ) + 2 pface ( ~ x ) e = density of the core layer (g/cm ) pface = density of the face layer (g/cm ) p cor paverage t (19) = average density of the board (g/cm ) 3 = thickness of the core layer (mm) core tface = thickness of the face layer (mm) t = thickness of the board (mm) A s mentioned above, usually the ratio of face layer density to core layer density ranges from 1.1 to 1.4. So, combining Eqs. (13), (17) and (18) with E q . (19) at a given ratio of _^£L t o r Issr*. t rlcore__ t j _ 2x-2^-), we can calculate the predicted tensile stiffness and t tensile strength of three-layer partial oriented flakeboards. For example, in this study, ^ ^ - = 0 . 3 and layer density). pf e= aC 1-3 p c o r e (1.3 was chosen as the ratio of face layer density to core From Eq.(19), we have Paverage = Pcore X 0.3 + 2.6 p c o r e X 0.35 (20) Finally we can get Eq. (21) and Table 3.3 Paverage 1-21 = pcore (21) Table 3.3 Layer density information of three-layer partially oriented flakeboards with 0.7: 0.3 as the ratio of face layer to core layer Average density (g/cm ) (p era e) av g 3 Density of the core layer (Pcore ) (g/cm ) Density of the face layer (pface) (g/cm ) 0.41 0.46 0.50 0.53 0.60 0.65 3 0.50 0.55 0.60 3 There is only tensile M O E and tensile strength data available for panel at 45° orientation angle range with different densities, and there is no tensile M O E and tensile strength data available for panels formed at other ranges of orientation with different densities except 0.5 g/cm . The linear regression information from Figures 3.7 and 3.8 3 was used to adjust tensile M O E and tensile strength of the higher density face layer and lower density core layer. Table 3.4. Regression results for tensile strength and tensile M O E of panel at 45° orientation range Average density (g/cm ) Tensile Strength (MPa) 0.65 0.60 0.55 0.50 0.45 19.71 18.00 16.28 14.57 12.86 3 Ratio 1.10 1.11 1.12 1.13 Tensile M O E (MPa) 8646 8076 7507 6938 6369 Ratio 1.07 1.07 1.08 1.09 40 As shown in the Table 3.4, an average ratio 1.11 is used for tensile strength and 1.08 is used for tensile M O E for other panels with different densities at orientation ranges. For example, the tensile strength of panel at 15° orientation range with 0.60 g/cm is 1.11 3 times the tensile strength of panel with 0.55 g/cm at 15° orientation range, which is 1.11 times the tensile strength of panel with 0.50 g/cm at 15° orientation range. 3 The calculated predicting tensile properties of three-layer partially oriented flakeboards are listed in Tables 3.5 to 3.7. Included in Table 3.8 are data showing the experimental results of three-layer partial oriented flakeboards with average density of 0.55 and 0.6 g/cm and in Table 3.9 are data 3 showing the comparison of the predicted and experimental tensile strength and tensile M O E . Despite the crude estimate of some panel's information such as the face layer density and core layer density, this mathematical model did reasonably well in predicting the tensile strength and M O E . Experimental tensile strength and M O E results are all lower than the corresponding predicted ones, ranging from 83 to 97 percent of the predicted values except for tensile M O E at density of 0.60 g/cm . 3 Part of the differences between the predicted and measured tensile strength can be explained by Weibull theory (Weibull, 1939). The essence of Weibull theory of strength is that the worst defect or weakest link controls the strength of a material. By applying this theory to different sizes of the same material, a size effect is recognized. Strength decreases as the specimen size increases, simply because more defects are expected in a larger volume. This has been observed for wood in tension parallel to grain (Madsen and 41 Buchanan, 1986), in bending (Bohannan, 1966), in shear (Foschi and Barrett, 1976), in tension perpendicular to grain (Barrett, 1974) and others (Lam et a l , 1990, Lam et al, 1994). The expression of this size effect according to Weibull's (1939) 2-P distribution is: a=I nW 1/k K (22) where, a and V are the strength and volume of material, k and m are the shape and scale parameters of 2-P Weibull distribution model which can be determined experimentally, e" dz was used by Weibull (1939) in the statistical manipulations. In this z 0 study, the volume of OSB panel is three times that of partially oriented assembly, this may partially explain why the predicted tensile strengths of three layer panel based on the partially oriented assembly are higher than the experimental results. Even though the difference between predicted and calculated values is not big, there is a still potential for the accuracy to be improved. a) From Figure 3.5, we can see that the V D D within each panel is not that of a truly discrete three-layer construction (i.e., a sandwich construction). It is more likely that an even higher precision could be reached if a multilayer analysis model is investigated, such as 5-layer or 7-layer. b) Each layer density was based on the assumption and the actual value is still unknown. c) The ratio of top and bottom layer to middle was based on the assumption and the actual value is still unknown. d) The strength properties of partially oriented assembly associated with higher density information were estimated from the regression equation. But scatter of experimental data about the regression indicates that there is a large range for the face layer's 42 maximum tensile strength and M O E . This lack of accuracy in prediction layer's strength directly affects the accuracy of the model. Interest was also the variability of the experimental and model results. The C V is given in Table 3.8 for each panel's type. While these values are based on only 8 replications, they do give an indication of specimen variability from which some general observation can be made. One such observation is that the three-layer experimental OSB panels were more consistent in tensile strength and M O E than the thin layer partially oriented assemblies. In other words, they are more consistent than the material from which they are made. For example, the C V range in Table 3.1 is from 28% to 50% while C V range in Table 3.8 is from 20% to 30%. There is big improvement in tensile MOE's consistency when panel's thickness was increased. This seems to be a consequence of he so-called "averaging effect" of composite materials. 43 Table 3.5. Prediction of tensile strength and stiffness (average density=0.5 g/cm ) 3 Panel structure: top and bottom layer 0°; middle layer 90° (random) Ratio of top and bottom to middle 0.3: 0.7 0.4: 0.6 0.5: 0.5 0.6: 0.4 0.7: 0.3 Tensile Strength (MPa) 13.01 14.67 16.33 18 19.67 Tensile Stiffness (MPa) 7087 7994 8901 9809 10716 Panel structure: top and bottom layer 15°; middle layer 90° (random) Ratio of top and bottom to middle 0.3: 0.7 0.4: 0.6 0.5: 0.5 0.6: 0.4 0.7: 0.3 Tensile Strength (MPa) 10.87 11.82 12.75 13.70 14.68 Tensile Stiffness (MPa) 5924 6444 6964 7483 8003 Panel structure: top and bottom layer 15°; middle layer 75° Ratio of top and bottom to middle 0.3: 0.7 0.4: 0.6 0.5: 0.5 0.6: 0.4 0.7: 0.3 Tensile Strength (MPa) 12.69 13.69 14.68 15.68 16.67 Tensile Stiffness (MPa) 6170 6654 7139 7624 8108 44 Table 3.6. Prediction of tensile strength and stiffness (average density =0.55 g/cm ) 3 Panel structure: top and bottom layer 0°; middle layer 90° (random) Ratio of top and bottom to middle 0.3: 0.7 0.4: 0.6 0.5: 0.5 0.6: 0.4 0.7: 0.3 Tensile Strength (MPa) 13.45 15.41 17.37 19.33 21.30 Tensile Stiffness (MPa) 7455 8629 9803 10978 12151 Panel structure: top and bottom layer 15°; middle layer 90° (random) Ratio of top and bottom to middle 0.3: 0.7 0.4: 0.6 0.5: 0.5 0.6: 0.4 0.7: 0.3 Tensile Strength (MPa) 11.74 13.12 14.51 15.92 17.30 Tensile Stiffness (MPa) 6098 6820 7542 8264 8987 Panel structure: top and bottom layer 15°; middle layer 75° Ratio of top and bottom to middle 0.3: 0.7 0.4: 0.6 0.5: 0.5 0.6: 0.4 0.7: 0.3 Tensile Strength (MPa) 12.54 13.76 14.97 16.19 17.41 Tensile Stiffness (MPa) 6646 7290 7934 8578 9222 45 Table 3.7. Prediction of tensile strength and stiffness (average densty=0.60 g/cm ) 3 Panel structure: top and bottom layer 0°; middle layer 90° (random) Ratio of top and bottom to middle 0.3: 0.7 0.4: 0.6 0.5: 0.5 0.6: 0.4 0.7: 0.3 Tensile Strength (MPa) 14.93 17.23 19.51 21.86 24.15 Tensile Stiffness (MPa) 8134 9390 10647 11904 13160 Panel structure: top and bottom layer 15°; middle layer 90° (random) Ratio of top and bottom to middle 0.3: 0.7 0.4: 0.6 0.5: 0.5 0.6: 0.4 0.7: 0.3 Tensile Strength (MPa) 12.23 13.64 15.06 16.46 17.87 Tensile Stiffness (MPa) 6669 7435 8205 8971 9741 Panel structure: top and bottom layer 15°; middle layer 75° Ratio of top and bottom to middle 0.3: 0.7 0.4: 0.6 0.5: 0.5 0.6: 0.4 0.7: 0.3 Tensile Strength (MPa) 13.26 14.65 16.05 17.46 18.86 Tensile Stiffness (MPa) 6921 7653 8385 9117 9850 46 Table 3.8. Tensile strength and stiffness of three-layer partial oriented flakeboards with average density of 0.55 g/cm and 0.60 g/cm (SD = standard deviation, C V = coefficient of efficiency) 3 Panel Category Density=0.55 g/cm Top and Bottom Layer: 15° Middle layer: 75° Density=0.55g/cm 3 3 MC Tensile Strength SD % MPa MPa 5.45 15.47 3.09 6.2 14.4 5.58 16.1 CV Tensile MOE SD CV MPa MPa 0.20 8901 2130 0.24 3.06 0.21 8719 1683 0.19 3.72 0.23 10214 2783 0.27 3 Top and Bottom Layer: 15° Middle layer: 90° Density=0.6 g/cm Top and Bottom Layer: 15° Middle layer: 75° 3 47 Table 3.9. The comparison of the predicted and experimental results Density=0.55 g /cm 3 Experiment Predicted Ratio of result properties Experiment to predicted (%) Panel structure: Top & Bottom layer: 15° Middle layer: 90° Tensile M O R (MPa) 14.40 17.30 83.2 Tensile M O E (MPa) 8719 8987 97.0 Panel structure: Top & Bottom layer: 15° Middle layer: 75° Tensile M O R (MPa) 15.47 17.41 88.8 Tensile M O E (MPa) 8901 9222 96.5 Panel structure: Top & Bottom layer: 15° Middle layer: 75° Tensile strength (MPa) 16.10 18.86 85.4 Tensile M O E (MPa) 10215 9850 103.7 Density=0.55 g /cm Density=0.60 g /cm 3 3 48 3.3 Finite element analysis of the tensile strength properties of the partially oriented flakeboard In the past several years, numerical modeling has been increasingly applied in the field of wood science, especially by means of finite element analysis (Triche and Hunt, 1993; Lam, 1994; Wang and Lam, 1995). The objective of this part of the study is to perform finite element analysis that predicts the probabilistic distribution of tensile strength in a three-layer partially oriented flakeboard. A database on the tensile M O E and M O R for single layer assemblies and three-layer partially oriented flakeboards was developed as presented in Chapters 2 and 3. In the finite element analysis, the structures of the flakeboard to be considered are as follows: Symmetrical three layer with partial orientation Face and bottom layer orientation: 15° Core layer orientation: 75° Proportion of face layer to core layer: 0.7:0.3 A commercial finite element analysis (FEA) program — A N S Y S , was used in this study. In A N S Y S , a three-dimensional element SOLID 45 was used to calculate the stresses of simulated members under various levels of axial loads. Comparing with the simulated strength of the members yielded the failure probability distribution. The element is defined by eight nodes having three degrees of freedom at each node: translations in the nodal x, y, and z directions. The mesh size for the analysis was 50 x 50 x 4 mm (length x width x thickness). 49 In the model, the finite element mesh of a parallel three-ply assembly with homogenous flake element is shown in Figure 3.12. Figure 3.12. Finite element mash of a three-layer flake assembly In wood, elastic properties vary with species, moisture content, orientation and many other factors. In this thesis, the ratio of elastic properties for aspen in the Longitudinal (L), Radial (R), and Tangential (T) directions is taken from Bodig and Jayne (1982) as: E L : E R : E T = 20:1.6:1; G „ : G p. „ =0.37; u. =0.50; \i tr RT :G L T =0.67; \i TR R T =10:9.4:1; E : G „ =14:1 =0.33; [i L RL =0.044; \i TL =0.027 (23) (24) where E represents modulus of elasticity, G represents modulus of rigidity and jx represents Poisson ratio. The experimental cumulative probability density distributions of tensile strength and tensile M O E for 15° and 75° orientation range assemblies and simulated ones are shown in Figures 3.13 and 3.14. M S Excel statistics functions (Normal distribution) are used to generate the simulated ones using experimental average and standard deviation values. During the simulation, negative values (2-3 percent of total data) are removed. The figures show that both tensile strength and M O E are very close to normal distribution. Tensile Strength (MPa) Figure 3.13. Cumulative probability of tensile strength of flake assemblies • 15 Degree Experiment CDF • 75 Degree Experiment CDF 15 Degree Simulation CDF 75 Degree Simulation CDF 20000 Tensile MOE (MPa) Figure 3.14. Cumulative probability of tensile M O E of flake assemblies 51 The average tensile strength and standard deviation of tensile strength values of single layer assembly with orientation 15° and 75°are 21.81 MPa and 6.92 MPa, and 9.05 M P a and 2.47 M P a respectively. In setting up the finite element model, the tensile M O E and tensile strength assigned to each element are required. The database on the tensile strength for single layer assembly was firstly obtained from the experiment as mentioned above. The relationship between experimental tensile M O E and tensile strength can be established and shown in Figure 3.15. So the tensile M O E value could be determined by the regression equation in Figure 3.15 for a given tensile strength value. After tensile M O E of each element was generated, the other elastic properties were obtained to construct the stress-strain relationship. The procedure ignored the randomness in tensile M O E for a given tensile M O R considering the relative high R value (0.88) for the relationship between tensile M O E and tensile 2 M O R . This procedure is adequate. For simplicity, linearity was only considered in the finite element analysis. So, for each simulation, the calculated stresses in all element for each layer at particular load level were compared with associated simulated strengths. In the case when the calculated stress exceed the simulated strength, the layer was considered as broken and the panel was considered as broken. The simulation sample size in the fitting procedure is 150. Based on the information of single layer flake assemblies, the failure probabilities of three-layer partially oriented flakeboards with single size of finite element meshes were simulated. Figures 3.16 and 3.17 show the experiment and simulation results for tensile strength and tensile M O E . The relationships between experimental tensile strength and tensile M O E , and between simulated tensile strength and tensile M O E of three layer panel are shown in Figures 3.18 and 3.19. Generally speaking, good agreement between the simulation and the testing data is obtained when the load is below the median level. When the load exceeds the median level, the difference between simulation and testing data is increased. Even for the experimental data, there is a big variation for the ultimate loads comparing with the tensile M O E . This trend is common for both solid wood and composites wood 20000 -| 0 40 1 , , 10 20 30 < 40 1 50 Tensile MOR (MPa) Figure 3.15. Relationship between experimental tensile M O E and tensile strength of thin assemblies. 53 materials since failure is associated with a localized weakness in the material and stiffness is measured over a much larger region, thus allowing for an averaging effect. Good agreement between the model prediction and the test data of three-layer partially oriented flakeboard is obtained. Although, simulation results match the test data generally, there are still potential for improving the simulation. a) As mention in 3.2, the V D D within each panel is not that of a truly discrete threelayer construction (i.e., a sandwich construction). It is more likely that an even higher precision could be reached if a multilayer analysis model is investigated on the condition that very thin assembly with thickness under 4 mm could be made. b) Size effect is not considered here. A l l the F E A element size is smaller than the experimental specimen size. According to the Weibull's theory mentioned in 3.2, the elements should have a slightly higher tensile strength than they were used in the analysis. c) Nonlinearity was not considered in the study. Whenever the calculated stresses exceed the simulated strength, the layer will be considered as broken. However, the member still has some reserve capacities as the load may be transferred from the broken layer to the unbroken layer. So a stresses updating procedure is used to account for the nonlinearity, it is expected to get higher precision. As mentioned in Chapter 2, manufacturing conditions were controlled in this study. Thus eliminating this variable from this model. To make this model more useful, a broader range of manufacturing conditions should be investigated. 54 •8 O 0.6 H 5 0.4 ^3 • 0.8 Experiment Simulation S 0.2 u 10 20 30 Tensile Strength (MPa) Figure 3.16. Cumulative probability distribution of simulation and experiment result for tensile strength of three-layer partially oriented flakeboards Figure 3.17. Cumulative probability distribution of simulation and experiment result for tensile M O E of three-layer partially oriented flakeboards 55 14000 y = 432.55x + 2207.2 ^ 12000 R = 0.9737 | 10000 ~ 8000 | 6000 •3 4000 I 2000 2 F- 1 0 5 10 15 20 25 30 Tensile Strength (MPa) Figure 3.18. Relationship between experimental tensile strength and tensile M O E of three-layer partially oriented flakeboards Figure 3.19. Relationship between simulation tensile strength and tensile M O E of three layer partially oriented flakeboards 56 3.4 CONCLUSIONS Within the parameters of the materials used and partially oriented wood assemblies and panels constructed in this study. It has been found that: 1. Partially oriented wood assemblies and panels provide us with information about how flake orientation affect panel's mechanical properties such as tensile strength and MOE. 2. The flake orientation change can dramatically affect panel properties. Flake alignment served to increase tensile strength and M O E in the direction of alignment to 3.78 and 3.08 times that of random configurations. 3. Panel's density is another important factor affecting panel's tensile strength and M O E . With the increasing of density, the tensile mechanical properties increased dramatically. 4. Three-layer partially oriented OSB panel's tensile strength and M O E can be predicted from its layer thickness' and stiffness' properties. Prediction precision varies with the board structures and densities. Prediction errors are within 20% for all boards tested. In order to increase the prediction precision, the multiple-layer of analysis could be used in the further study. 5. The mathematical model developed in this study could be another approach to predict the tensile strength and tensile M O E properties based on the each layer properties that were obtained by experiments using a robot system. The advantage for this approach is easy and simple. 6. A finite element analysis model was developed to model the probability of failure and probabilistic distribution of tensile strength of OSB. The probabilistic distributions of 57 tensile strength and the load capacity probabilistic distributions for three-layer partially oriented OSB were predicted successfully. Good agreement between predictions and experimental data was observed. CHAPTER 4. SUMMARY AND CONCLUSIONS The concept of H D D was first proposed by Suchsland in 1959 to analyze particleboard technology. A gravimetric method involving a drilling technique was shown to be capable of detecting the micro-density variation in particleboard, as well as distinguishing particleboards in this regard. It was found that this variation decreased as specimen size, used in determining this density nonuniformity, increased. If standard deviation was used as the index of nonuniformity, the expression S =a(l/A) was shown to be appropriate to b relate density variation (S) to specimen size (A). Parameter b related to correlations of density points, and a range between 0 and 0.5 was determined for this parameter. It was found that b decreased as particle size increased. Raw material characteristics influenced HDD. Generally, with larger specimen sizes, particleboards made with larger particles exhibited greater density variations. With smaller specimen sizes, particleboards made with smaller particles showed smaller density variations. The size and number of voids were identified as responsible for these results of mats provided under hand-forming operations. The recognition of these two aspects of voids suggests that forming method plays a significant role in determining the magnitude of H D D . Any method that reduces the size of voids in between particles would improve board uniformity. In the past years, considerable research has been carried out on how flake orientations influence OSB mechanical properties, but nearly all of them used the "pure oriented" method namely, all flakes in the panel were oriented in one specific direction or two. With the help of the robot system, partially oriented flake assemblies in this study can 59 very closely represent commercial OSB panels and provide us with more details about how flake orientation affects panel's tensile mechanical properties. The current development of available structural wood composites has largely resulted from costly trial-and-error based laboratory experiments. A n analytical model capable of predicting the structural properties of a wood composite material based on the properties of its ingredients would be an invaluable tool. The real value of modeling lies in its capacity to simplify description of properties and to identify the common characteristics of different composites. In this study, a layer concept was developed and used to evaluate the effects of layer's properties such as density, flake orientation on the tensile strength and M O E of partially oriented OSB. Thin layer flake assemblies (4 mm in thickness) with different orientations and densities were formed by a robot and tested. A mathematical analysis model was derived based on this concept to predict those properties. The verification process shows that all of the experimental results were lower than this model's predicted values. Furthermore, a finite element analysis 3D model was developed to model the probabilistic distribution of tensile strength of three-layer partially oriented flakeboard, with the help of a commercial finite element analysis (FEA) program — A N S Y S . The simulation results also show good agreement between the model prediction and the test data of three-layer partially oriented flakeboard. 60 4.1 Future Developments Scientists and engineers have been more successful in designing synthetic fiberreinforced composites than wood-based composites, mainly because of the complexity of the microstructures and the inherent variability of the latter. Particleboard could be viewed as a three dimensional nonuniform structure. With the establishment of the concept of horizontal density distribution, a concept of a three dimensional density distribution could be developed by combining the knowledge of vertical density profile. This concept could lead to the development of a general theory on short-fiber wood composites, similar to the laminate theory for continuous fiber composites. Particleboard is a highly complex material, as physical, structural and chemical mechanism all contribute to its ultimate properties. The particleboard structure is not uniform, nor are physical and chemical properties. The understanding of how structure interacts with other properties is not only important to fully comprehend the influence of H D D and V D D , but to the development of future particleboard design values for engineering calculations and reliability analysis. The determination of the effect of manufacturing parameters on strand properties is possible for further study. Manufacturing conditions were controlled in this study. Thus eliminating this variable from this model. A broader range of manufacturing conditions such as adding wax, different pressing time and temperature should be investigated. 61 The greatest potential for modeling and simulating the tensile strength and M O E may lie in its ability to be expanded to allow prediction of other properties such as compression and bending. Further advancement of the model to permit modeling of commercial products is also possible. The final contribution will be to improve and optimize short-fiber composite products, utilize wood resource more efficiently and make less experimental testing. 6 2 REFERENCE 1. American Society for Testing and Materials. 1994. Standard test methods for evaluating properties of wood-base fiber and particle panel materials. Philadelphia, PA. 2. Badejo, S.O.O. 1988. Effect of flake geometry on properties of cement-bonded particleboard from mixed tropical hardwoods. Wood Sci. Technol., 22(3):357-370. 3. Barrett, J.D. 1974. Effect of size on tension perpendicular to grain strength of Douglas-fir. Wood and Fiber. 6(2): 126-143. 4. Bodig, J. and B . A . Jayne. Mechanics of Wood and Wood Composites. Van Nostrand Reinhold Company Inc. 1982. p i 15-118. 5. Hohannan, B . 1966. Effect of size on bending strength of wood members. U S D A Fores Service, Research Paper R P L 56. 6. Bowyer, J. 1995. The raw material for 21 century. Forest Prod. J., 45(5): 12-18 7. Canadido, L.S., F. Saito and S. Suzuki. 1988. Effect of particle shape on the orthotropic properties of oriented stand board. Mokuzai Gakkaishi, 34(1): 21-27. 8. Canadido, L.S., F. Saito and S. Suzuki. 1988. Influence of strand thickness and board density on the orthotropic properties of oriented stand board. Mokuzai Gakkaishi, 36(8): 632-636. 9. Cha, J.K. and R.G. Pearson. 1994. Stress analysis and prediction in 3-layer laminated veneer lumber: Response to crack and grain angle. Wood Fiber Sci., 26(1): 67-72. 10. Dai, C and S. Chen. 1996. Simulation of mat formation for wood strand composites processing. Third Pacific Rim Bio-Based Composite Symposium. Japan, p i 17-124. 11. Dai, C and P.R. Steiner. 1993. Compression behavior of randomly formed wood flake mats. Wood Fiber Sci., 25(4): 234-240. 12. Dai, C and P.R. Steiner. 1993. Spatial structural of wood composites in relation to processing and performance characteristics: Part 1. Rationale for model development. Wood Sci. Technol., 28(1): 45-51. 13. Dai, C and P.R. Steiner. 1994. Spatial structural of wood composites in relation to processing and performance characteristics: Part 2. Modeling and simulation of a randomly-formed flake layer network. Wood Sci. Technol., 28(2): 135-146. 14. Dai, C and P.R. Steiner. 1994. Spatial structural of wood composites in relation to processing and performance characteristics: Part 3. Modeling the formation of multilayered flake mats. Wood Sci. Technol., 28(3): 229-239. 63 15. Dai, C and P.R. Steiner. 1995. Analysis and implication of structure in short fiber wood composites. Second Pacific Rim Bio-Based Composite Symposium, November 6-9, Vancouver, Canada, pi7-24. 16. Ellis, S. C. 1989. Some factors affecting the flow and penetration of powdered phenolic resin into wood. Ph.D. thesis. The University of British Columbia, Department of Wood Science. 17. Engels, K . 1978. Latest European development in blending. Proceedings, Thirteenth Washington State University International Symposium on Particleboard. T . M . Maloney, Ed., Washing State University, Pullman, W A , pl03-l 15. 18. Foschi, R. A . and J. D . Barrett. 1976. Longitudinal shear strength of Douglas-fir. Can. J. Civil Eng., 3(2): 198-208. 19. Geimer, R.L., H . M . Montrey and W.F. Lehmann. 1975. Effects of layer characteristics on the properties of three-layer particleboard. Forest Prod. J., 45(3): 19-29. 20. Geimer, R.L. 1975. Effects of layer characteristics on the properties of three layer particleboards. Forest Prod. J., 25(3): 19-29. 21. Geimer, R.L. 1978. Construction variables considered in the fabrication of a structural flakeboard. General technical report WO-5. U S D A Forest Service. Washington, D. C. p69-80. 22. Geimer, R.L. 1979. Data basic to the engineering design of reconstituted flakeboard. Proceedings, Thirteenth Washington State University International Symposium on Particleboard. T . M . Maloney, Ed., Washing State University, Pullman, W A , pp. 105125. 23. Geimer, R.L. 1981. Predicting flakeboard properties: improvement in vending properties by aligning a mixture of flakes. Proceedings, Fifteenth Washington State University International Symposium on Particleboard. T . M . Maloney, Ed., Washing State University, Pullman, W A , pi05-125. 24. Geimer, R.L. 1981. Predicting shear and internal bond properties of flakeboard. Holz als Roh-und Werkst., 39(10): 409-415. 25. Geimer, R.L. 1982. Dimensional stability of flakeboards as affected by board specific gravity and flake alignment. Forest Prod. J., 32(8): 44-52. 26. Hall, P. 1988. Introduction to the theory of coverage processes. New York: John Wiley & Sons. 27. Hankinson, R.L. 1921. Investigation of Crushing Strength of Apruce at Varying Angles of Grain. Air Service Information Circular No. 259, U.S. A i r Service. 64 28. Hansel A and P. Niemz. 1988. Development of a model for cross-sectional density profile in three layer particleboard. Holz als Roh und Werk., 46(4): 125-132. 29. Harless, T.E.G. 1987. A model to predict the density profile of particleboard. Wood Fiber Sci., 19(1): 43-51. 30. Heath, L. 1985. Fundamentals of Robotics-Theory and Applications. Reston Publishing Company Inc. Reston. 31. Heebink, B.G., W.F. Lehmann and F.V. Hefty. 1972. Reducing particleboard pressing time: Exploratory study. U S D A For. Serv. Res. Pap. F P L 180. For. Prod. Lab., Madison, Wis. 32. Hoover, W.L., M.O. Hunt, R.C. Lattanzi, J.H. Bateman and J.A. Youngquist. 1992. Modeling mechanical properties of single-layer, aligned, mixed-hardwood strand panels. Forest Prod. J., 42(5): 12-18. 33. Hunt, M.O. and S.K. Suddarth. 1974. Prediction of elastic constants of particleboard. Forest Prod. J., 24(5): 52-57. 34. Kamke, F.A. and M.P. Wolcott. 1991. Fundamentals of flakeboard manufacture: wood-moisture relationships. Wood Sci. Technol., 25(1): 57-71. 35. Kelly, M . W . 1977. Critical literature review of relationships between processing parameters and physical properties of particleboard. Gen. Tech. Rept. FPL-10. U S D A forest Serv., Forest Prod. Lab., Madison, Wis., p2-38. 36. Kollman, F.F.P. and W.A.Jr. Coe. 1968 Principles of Wood Science and Technology, I Solid Wood. New York, Springer-Verlag Inc., p252-261. 37. Lam, F. and E. Varoglu. 1990. Effect of length on the tensile strength of lumber. Forest Prod. J., 40(5): 37-42. 38. Lam, F., Y.T. Wang and J.D. Barrett. 1994. Simulation of correlated non-stationary lumber properties. J. Material Civil Eng., 6(1): 34-53. 39. Lang, E . M . , and M.P. Wolcott. 1996. A model for the viscoelastic consolidation in wood strand mats. Part I. Structural characterization of the mat via Monte Carlo simulation. Wood Fiber Sci., 28(1): 100-109. 40. Lau, P. W - C . 1982. Numerical approach to predict the Modulus of elasticity of oriented waferboard. Wood Sci., 14(2): 73-85. 41. Lu, C. 1995. Composite mat simulation program. Directed study report. Department of Wood Science, The University of British Columbia, April, plO-29. 42. Lu, C , P.R. Steiner, and F. Lam. 1998. Simulation study of wood-flake composite mat structure. Forest Prod. J., 48(5):89-93. 65 43. Madsen, B . and A. H . Buchanan. 1986. Size effect in timber by a modified weakest link theory. Can. J. Civil. Eng., 13: 218-232. 44. Maloney, T . M . 1970. Resin distribution in layered particleboard. Forest Prod. J., 20(1): 45-50. 45. McNatt, J. D . 1973. Basic engineering properties of particleboard. U S D A Forest Service, Forest Product laboratory, Research Paper FPL-206. 46. McNatt, J. D. 1994. Static bending properties of structural wood-based panels: large panels versus small specimen tests. Forest Prod. J., 34(4): 50-54. 47. McNatt, J. D., L. Bath and R. W. Wellwood. 1992. Contribution of flake alignment to performance of strandboard. Wood Sci. Technol., 26(2): 257-265. 48. Moslemi, A. A . 1974. Particleboard. Volume I: Materials. Southern Illinois University Press. Carbonale and Edwardsville, pl23-158. 49. Oudjehane, A., F. Lam and S. Avramidis. 1998. Modeling the influence of the formation process on engineering properties of flakeboards. Engineered System Using Structural Panels, Forest Products Society (Eds.), Madison, WI, pl-5. 50. Oudjehane, A. and F. Lam. 1998a. Modeling the profile of engineering properties within random wood based composite panels. ICCE/5, Las Vegas, p693-694. 51. Oudjehane, A. and F. Lam. 1998b. On the density profile within random and oriented wood-based composite panels: horizontal distribution. Composites, Part B , 29B, 687694. 52. Shaler, S.M and P.R. Blankenhorn. 1990. Composite model prediction of elastic moduli for flakeboard. Wood Fiber Sci., 22(3): 246-261. 53. Shen, K . and M . N . Calloll. 1969. A new method for evaluation of internal strength of particleboard. Forest Prod. J., 19(8): 17-21. 54. Smith, D . 1980. Consideration in press design for structural board. Proceedings, Fifteenth Washington State University International Symposium on Particleboard. T . M . Maloney, Ed., Washing State University, Pullman, W A . , pl45-159. 55. Strickler, M . D . 1959. Effect of press cycle and moisture content on properties of Douglas fir flakeboard. Forest Prod. J., 9(7): 203-205. 56. Steenberg, B . 1966. Formation and properties of fiber network. In Consolidation of the Paper Web. (Francis Bolam, Ed.). Transactions of the symposium held at Cambridge. Vol. 1:177-186. 66 57. Steiner, P.R and C. Dai. 1993. Spatial structure of wood composites in relation to processing and performance characteristics. Part 1. Rationale for model development. Wood Sci. Technol., 28(1): 45-51. 58. Steiner, P.R and W. Xu. 1995. Influence of flake characteristics on horizontal density distribution of flakeboard. Forest Prod. J., 45(4): 61-66. 59. Suchsland, O. 1959. A n analysis of the particleboard process. Q. Bull., Mich. Agr. Exp. Sta., Mich. State Univ. 42(2): 350-372. 60. Suchsland, O. 1962. The density distribution in strandboard. Michigan Quart. Bull., 45(1): 104-121. 61. Suchsland, O. 1967. Behavior of a particleboard mat during the pressing cycle. Forest Prod. J., 17(2): 51-57. 62. Suchsland, O. and H . Xu. 1989. A simulation of the horizontal density distribution on a flakeboard. Forest Prod. J., 39(5): 29-33. 63. Suchsland, O. and H . Xu. 1991. Model analysis of flakeboard variables. Forest Prod. J., (11/12): 55-60. 64. Suo, S.M. and J.L. Bowyer. 1994. Simulation modeling of particleboard density profile. Wood Fiber Sci., 26(3): 397-411. 65. Triche, M . H . and M.O. Hunt. 1993. Modeling of parallel-aligned wood strand composites. Forest Prod. J., 43(11/12): 33-44. 66. Vanmarcke, E. 1993: Random fields: Analysis and synthesis. Combridge: The MIT press. 67. Wang, Y.T. and F. Lam. 1996. Tensile strength modeling of parallel-aligned strand based wood composites. Proceedings of Third Pacific Rim Bio-Based Composite Symposium, Koyto, Japan, Dec. 1996. p234-239. 68. Wang, Y.T. and F. Lam. 1997. Computational modeling of material failure for parallel-aligned strand based wood composites. Comput. Material Sci. (11): 157-165. 69. Weibull, W. 1939. A statistical theory of the strength of materials. Proc, Royal Swedish Inst., No. 151, Stockholm, Sweden. 70. Winistorfer, P.M., T . M . Young and E. Walker. 1996. Modeling and comparing vertical density profiles. Forest Prod. J., 46(2): 35-40. 71. Wolcott, M P . 1990. Fundamentals of flakeboard manufacture: Viscoelastic behavior of wood composites. Wood Fiber Sci. 22(4): 345-361. 72. X u , W. 1994. Horizontal density distribution of particleboard: origin and implication Ph.D. thesis. Faculty of Forestry, The University of British Columbia, p23-56. 67 73. Xu, W . and P. R. Steiner. 1995. Rationalizing internal bond and thickness swell test specimen size. Wood Fiber Sci., 27(4): 389-394. 68
- Library Home /
- Search Collections /
- Open Collections /
- Browse Collections /
- UBC Theses and Dissertations /
- A model to predict tensile mechanical properties of...
Open Collections
UBC Theses and Dissertations
Featured Collection
UBC Theses and Dissertations
A model to predict tensile mechanical properties of robot formed wood flakeboard Chen, Guanqi 2002
pdf
Page Metadata
Item Metadata
Title | A model to predict tensile mechanical properties of robot formed wood flakeboard |
Creator |
Chen, Guanqi |
Date Issued | 2002 |
Description | Compared with solid wood, one of the advantages of wood-based composites is the great potential for the design of material properties through manipulation of manufacturing variables. Large strides are presently being made in the design of non-veneer structural panels such as oriented strand board (OSB) by using material science and engineering principles. Scientists and engineers have been more successful in designing synthetic fiber-reinforced composites than wood-based composites, mainly because of the complexity of the microstructures and the inherent variability of the wood composites. In this study, recent research in modeling and predicting the properties of flakeboards has been summarized. The relationships among the structure in terms of void volume, density distribution, and the properties of the panels are discussed. With the help of a robotic system, very thin partially oriented wood assemblies were made and tested. The relationship between flake orientation and tensile strength and tensile MOE were determined. The relationship between density and tensile strength and tensile MOE were also examined. Based on layer properties, a three-layer mathematical model was derived to predict the tensile strength and tensile MOE. Partially oriented three-layer OSB panels were made and tested to verify this mathematical model. Very good agreement was found between this model and experimental results. Furthermore, a 3D finite element model was developed to simulate the probability of failure and probabilistic distribution of tensile strength of OSB. The probabilistic distributions of tensile strength and the load capacity probabilistic distributions for three-layer partially oriented OSB were predicted successfully. Good agreement between predictions and experimental data was observed. |
Extent | 3247689 bytes |
Genre |
Thesis/Dissertation |
Type |
Text |
File Format | application/pdf |
Language | eng |
Date Available | 2009-08-12 |
Provider | Vancouver : University of British Columbia Library |
Rights | For non-commercial purposes only, such as research, private study and education. Additional conditions apply, see Terms of Use https://open.library.ubc.ca/terms_of_use. |
DOI | 10.14288/1.0090134 |
URI | http://hdl.handle.net/2429/12009 |
Degree |
Master of Applied Science - MASc |
Program |
Forestry |
Affiliation |
Forestry, Faculty of |
Degree Grantor | University of British Columbia |
Graduation Date | 2002-05 |
Campus |
UBCV |
Scholarly Level | Graduate |
Aggregated Source Repository | DSpace |
Download
- Media
- 831-ubc_2002-0039.pdf [ 3.1MB ]
- Metadata
- JSON: 831-1.0090134.json
- JSON-LD: 831-1.0090134-ld.json
- RDF/XML (Pretty): 831-1.0090134-rdf.xml
- RDF/JSON: 831-1.0090134-rdf.json
- Turtle: 831-1.0090134-turtle.txt
- N-Triples: 831-1.0090134-rdf-ntriples.txt
- Original Record: 831-1.0090134-source.json
- Full Text
- 831-1.0090134-fulltext.txt
- Citation
- 831-1.0090134.ris
Full Text
Cite
Citation Scheme:
Usage Statistics
Share
Embed
Customize your widget with the following options, then copy and paste the code below into the HTML
of your page to embed this item in your website.
<div id="ubcOpenCollectionsWidgetDisplay">
<script id="ubcOpenCollectionsWidget"
src="{[{embed.src}]}"
data-item="{[{embed.item}]}"
data-collection="{[{embed.collection}]}"
data-metadata="{[{embed.showMetadata}]}"
data-width="{[{embed.width}]}"
async >
</script>
</div>
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.831.1-0090134/manifest