UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Influence of panel structure on wood to flakeboard nail connection properties Wang, Sunguo 2001

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

Item Metadata

Download

Media
831-ubc_2001-0132.pdf [ 8.99MB ]
Metadata
JSON: 831-1.0075171.json
JSON-LD: 831-1.0075171-ld.json
RDF/XML (Pretty): 831-1.0075171-rdf.xml
RDF/JSON: 831-1.0075171-rdf.json
Turtle: 831-1.0075171-turtle.txt
N-Triples: 831-1.0075171-rdf-ntriples.txt
Original Record: 831-1.0075171-source.json
Full Text
831-1.0075171-fulltext.txt
Citation
831-1.0075171.ris

Full Text

INFLUENCE OF PANEL STRUCTURE ON WOOD TO FLAKEBOARD NAIL CONNECTION PROPERTIES By SUNGUO W A N G M . Eng., Northeastern Forestry University, China, 1987 B. Sc., Nanjing Forestry University, China, 1984 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in F A C U L T Y OF G R A D U A T E STUDIES Department of Wood Science FACULTY OF FORESTRY We accept this thesis as conforming lo the required standard The University of British Columbia October, 2000 © Sunguo Wang, 2000 I n p r e s e n t i n g t h i s t h e s i s i n p a r t i a l f u l f i l m e n t o f t h e r e q u i r e m e n t s f o r an advanced degree a t t h e U n i v e r s i t y of B r i t i s h C o l u m b i a , I a gree t h a t t h e L i b r a r y s h a l l make i t f r e e l y a v a i l a b l e f o r r e f e r e n c e and s t u d y . I f u r t h e r agree t h a t p e r m i s s i o n f o r e x t e n s i v e c o p y i n g of t h i s t h e s i s f o r s c h o l a r l y p u r p o s e s may be g r a n t e d by t h e head o f my department o r by h i s o r h e r r e p r e s e n t a t i v e s . I t i s u n d e r s t o o d t h a t c o p y i n g o r p u b l i c a t i o n of t h i s t h e s i s f o r f i n a n c i a l g a i n s h a l l not be a l l o w e d w i t h o u t my w r i t t e n p e r m i s s i o n . Department o f The U n i v e r s i t y of B r i t i s h C olumbia Vancouver, Canada ABSTRACT Oriented Strand Board (OSB) or flakeboard is widely used in the building industry as different components such as shear walls, floors, roofs and underlayments. The performance of OSB to lumber connections has been investigated by many researchers, but its relationship to OSB or flakeboard panel structure has never been systematically studied. The study presented in this thesis focuses on this research scope. The research project was divided into two parts. Preliminary tests (Phase I) on OSB-to-lumber nail connections were conducted using 11 mm commercial OSB panels as side members and Spruce-Pine-Fir (SPF) lumber as main members. Several combinations of OSB specimen sizes, nailing patterns and test set-ups were investigated. Tensile loads were applied statically along the longitudinal direction of the lumber member, but perpendicular to the nail shank for all specimens. Both the single nail and the two-nail combination patterns were examined in OSB specimen I (50x240x11 mm) and specimen II (240x240x11 mm). Loading directions relative to OSB face flake orientation were studied for specimen II. The results showed that the chosen test jigs were suitable for small sized OSB-to-lumber nailed connections. The new set-up with specimen II was more efficient for small scale nail connection testing since the specimen can be easily adjusted to study the influence of the loading directions, nailing patterns and multiple nailing; hence, more information could be obtained. Two main failure modes, pull-through and pull-out, were observed in the preliminary tests. The second part of the project (Phase II) included the main tests. Three principal processing parameters, flake orientation, flake thickness and board density, were considered in the experimental design of flakeboard structures. A Monte Carlo computer program WinMat® was used to simulate mat structure patterns and their corresponding horizontal density profiles. A robot-based formation system was applied to build flakeboard mats, which ensured exactly the same mat structures as defined in the computer program. Predefined and laboratory-manufactured oriented and random flakeboards were then conditioned and assembled with 38x89 mm SPF lumber into nail connections. Single nail lateral resistance tests were conducted to study the effects of n failure modes, panel types and loading directions on nail-connection properties. The results showed that: 1) most nail properties for the specimens that failed in the pull-out mode were significantly different from those in the pull-through mode; 2) the specimens that failed in the pull-out mode had higher initial stiffness and connection strength (maximum, yield and ultimate loads) than those in the pull-through mode; 3) compared to OSB panels, random panels had higher connection strength for the pull-through mode, larger maximum displacement for the pull-out mode, and higher maximum and ultimate strain energies, and larger ultimate displacement for both failure modes; 4) the 90° loading direction in OSB panels indicated significantly different nail properties for both pull-out and pull-through modes, compared with the 0° and 45° loading directions, but there were no significant differences in nail properties between 0° and 45° loading directions under the pull-through mode; 5) there was significant difference in connection strength between 0° and 45° loading directions under the pull-out mode; 6) from regression analyses, most of the OSB or random flakeboard to SPF lumber nail connection properties were affected by different combinations of panel local density (LD), board to flake thickness ratio (TR), and lumber specific gravity (G); 7) a parametric study was carried out to show a potential application of the information developed in this paper; generally, higher lumber specific gravity and panel local density mostly showed better initial stiffness and connection strength (loads) within the regression ranges and fixed lumber or flakeboard properties. However, the effect of panel to flake thickness ratio is comparatively complex. Different types of connection or loading conditions may produce opposite trends. Hankinson's equation predicts very close initial stiffness and maximum load to measured values at 45° loading angle based on nail properties along and across OSB face flake alignment, and may also have good predictions on the nail performance at any loading angle, which will be verified in the further study. in TABLE OF CONTENTS ABSTRACT ii TABLE OF CONTENTS iv LIST OF TABLES vi LIST OF FIGURES viii ACKNOWLEDGEMENTS x CHAPTER 1 LITERATURE OVERVIEW 1 1.1 INTRODUCTION 1 1.2 RESEARCH BACKGROUND 1 1.3 RESEARCH GOALS 9 1.4 REFERENCES 10 CHAPTER 2 PHASE I. PRELIMINARY TESTS ON COMMERCIAL OSB PANELS 14 2.1 INTRODUCTION 16 2.2 MATERIALS AND METHODS 18 2.2.1 Materials 18 2.2.2 Methods 18 2.2.2.1 Nailing Patterns and Test Set-ups 18 2.2.2.2 Load-Slip Curves and Definitions of Nail Connection Properties 25 2.2.2.3 Lumber Specific Gravity and OSB Local Density 28 2.3 RESULTS AND DISCUSSION 29 2.3.1 Results 29 2.3.2 Discussion 30 2.3.2.1 Test Set-ups 30 2.3.2.2 Failure Modes 36 2.3.2.3 Loading Directions 38 2.3.2.4 OSB Local Density, Lumber Specific Gravity and Moisture Content 41 2.4 CONCLUSIONS 43 2.5 REFERENCES 44 CHAPTER 3 MAT PATTERN SIMULATION AND PANEL MANUFACTURE 47 3.1 INTRODUCTION 48 3.2 INPUT D A T A FOR M A T STRUCTURES 49 3.3 M A T SIMULATION 51 3.4 PANEL M A N U F A C T U R E 58 3.4.1 Manufacturing Process 58 iv 3.4.2 Materials 58 3.4.3 Robot-Based Mat Formation 59 3.4.4 Panel Pressing 60 3.5 RESULTS AND DISCUSSION 61 3.6 CONCLUSIONS 61 3.7 REFERENCES 62 CHAPTER 4 PHASE II. NAIL CONNECTION TESTS ON ROBOT-FORMED PANELS 64 4.1 INTRODUCTION 66 4.2 MATERIALS AND METHODOLOGY 68 4.2.1 Materials 68 4.2.2 Methodology 70 4.2.2.1 Nailing Pattern and Experimental Procedures 70 4.2.2.2 Panel Local Density Used for Regression Analysis 73 4.2.2.3 Statistical Analysis of Testing Results 75 4.2.2.3.1 T-tests 75 4.2.2.3.2 Multivariate Regressions 76 4.3 RESULTS AND DISCUSSION 77 4.3.1 Results 77 4.3.2 Discussion 77 4.3.2.1 T-tests on Means of Nail-Connection Properties 77 4.3.2.1.1 Failure Modes 78 4.3.2.1.2 Panel Types 90 4.3.2.1.3 Loading Directions in OSB-Lumber Joints 91 4.3.2.2 Multivariate Regression Models 93 4.3.2.3 Parametric Study of Regression Models 104 4.3.2.3.1 Impact of Lumber Specific Gravity 104 4.3.2.3.2 Influence of Panel Local Density 106 4.3.2.3.3 Impact of Panel to Flake Thickness Ratio 107 4.3.2.4 Prediction of Nail Connection Properties Using Hankinson's Formula 109 4.4 CONCLUSIONS 112 4.5 REFERENCES 113 CHAPTER 5 CONCLUSIONS 116 v LIST OF TABLES Table 2.1 Number and type of tests conducted 24 Table 2.2 Two different nailing formats (Specimens I and II) 31 Table 2.3 Effects of loading directions on nail properties (Specimen II) 32 Table 3.1 Factors and three levels 50 Table 3.2 Taguchi's L 9 (34) array 50 Table 3.3 Comparisons of target and actual board density 60 Table 4.1 Simulation and experimental local density values for some mat patterns 74 Table 4.2 Effect of two main failure modes on nail connection properties 80 Table 4.3 Nail connection properties of oriented flakeboard (OSB)/lumber connection 81 Table 4.4 Nail connection properties of random panels/lumber connection 82 Table 4.5 Nail connection properties when loading along face flake orientation 83 Table 4.6 Nail connection properties when loading at 45 degree to face flake orientation Table 4.7 Nail connection properties when loading across face flake orientation Table 4.8 Percentages of failure modes for random and oriented panels and loading directions in OSB panels 86 Table 4.9 T-tests on two failure modes: pull-through and pull-out 87 Table 4.10 T-tests on performance means under pull-through failure mode 88 Table 4.11 T-tests on performance means under pull-out failure mode 89 Table 4.12 Regression models of oriented panel nail connection properties 94 Table 4.13 Regression models of random panel nail connection properties 95 Table 4.14 Regression models when loading along face flake alignment 96 Table 4.15 Regression models when loading at 45° to face flake alignment 96 Table 4.16 Regression models when loading across face flake alignment 97 84 85 v i Table 4.17 Effect of lumber specific gravity (G) on initial stiffness and maximum load when TR=11.9 and LD=0.630 g/cm 105 Table 4.18 Impact of panel local density (LD) on initial stiffness and maximum load when TR=11.9 and G=0.430 106 Table 4.19 Impact of panel to flake thickness ratio (TR) on initial stiffness and maximum load when G-0.430 and LD=0.630g/cm3 108 Table 4.20 Comparisons of measured and Hankinson's formula-predicted initial stiffness and maximum load 110 vn L I S T O F F I G U R E S Fig. 1.1 A new set-up for nail connection tests 5 Fig. 2.1 Nailing format for 240x50x11 mm OSB 20 Fig. 2.2 Nailing pattern for 240x240x11 mm panels 20 Fig. 2.3 Test set-up diagram for 240x50x 11 mm OSB 21 Fig. 2.4 Test set-up diagram for 240x240x 11 mm panels along two diagonals 22 Fig. 2.5 Test set-up photo for 240x240x11 mm panels 23 Fig. 2.6 Load-displacement curve for nail connection tests 25 Fig. 2.7 Typical load-displacement curve for pull-out failure 27 Fig. 2.8 Typical load-displacement curve for pull-through mode 28 Fig. 2.9 Relationship of test set-up to stiffness and maximum load per nail 30 Fig. 2.10 Impact of test set-up on three displacements 34 Fig. 2.11 Influence of test set-up on strain energy per nail 35 Fig. 2.12 Effect of test set-up on ductility factors 35 Fig. 2.13 Relationship of failure modes to stiffness and maximum load per nail 37 Fig. 2.14 Effect of loading direction on stiffness and maximum load 39 Fig. 2.15 Loading direction and displacements 39 Fig. 2.16 Influence of loading direction on strain energy 40 Fig. 2.17 Relationship of loading direction to ductility factors 40 Fig. 2.18 OSB local density and failure modes 42 Fig. 3.1 Flake location and alignment in a mat 51 Fig. 3.2 Two-dimension graphs of density distribution 55 Fig. 3.3 Three-dimension graphs of density distribution 56 Fig. 3.4 Simulation panels with three main flake alignments 57 Fig. 3.5 Robot-formed panels with three main flake alignments 57 Fig. 3.6 Robot forming system 59 Fig. 4.1 Nailing pattern for laboratory-based panels 70 Fig. 4.2 Simulated and measured panel local density comparisons 73 vm Fig. 4.3 Nail shapes in pull-through and pull-out failure modes 78 Fig. 4.4 Relationship of measured and predicted initial stiffness in OSB to lumber connection 99 Fig. 4.5 Relationship of measured and predicted maximum load in OSB to lumber connection 99 Fig. 4.6 Relationship of measured and predicted initial stiffness in random flakeboard to lumber connection 100 Fig. 4.7 Relationship of measured and predicted maximum load in random flakeboard to lumber connection 100 Fig. 4.8 Relationship of measured and predicted initial stiffness when loaded along face flake alignment in OSB to lumber connection 101 Fig. 4.9 Relationship of measured and predicted maximum load when loaded along face flake alignment in OSB to lumber connection 101 Fig. 4.10 Relationship of measured and predicted initial stiffness when loaded at 45° to face flake alignment in OSB to lumber connection 102 Fig. 4.11 Relationship of measured and predicted maximum load when loaded at 45° to face flake alignment in OSB to lumber connection 102 Fig. 4.12 Relationship of measured and predicted initial stiffness when loaded across face flake alignment in OSB to lumber connection 103 Fig. 4.13 Relationship of measured and predicted maximum load when loaded across face flake alignment in OSB to lumber connection 103 Fig. 4.14 Relationship of initial stiffness with loading angle in OSB to lumber connection 111 Fig. 4.15 Relationship of maximum load with loading angle in OSB to lumber connection 111 ix ACKNOWLEDGEMENTS I would like to thank my supervisors Dr. Frank Lam and Dr. Stavros Avramidis, Department of Wood Science, UBC, for their invaluable advice, tremendous suggestions and help throughout this research. Thanks also go to my supervisory committee members: Dr. David Barrett and Dr. Simon Ellis for their guidance. Gratitude is extended to Dr. Robert Kozak for his direction in Statistics. Acknowledgement is given to Bob Myronuk and Avtar Sidhu for their laboratory assistance. Other helps from Dr. Congjin Lu, Dr. Chunping Dai, Dr. Liping Cai, George Lee, Henry Ming He, Yintang Wang, Bingning Zhou, Pablo Garcia, Bingye Hao, Stephanne Fabris, Kaiyuan Wang, Jianhe Wang, Peggi Clouston, Wilson Lau, Julia Laber, Winnie Louie and Dayna Furst are deeply appreciated. I also wish to thank Ainsworth Lumber Co. for contributing testing materials, C A E Industries and UBC Carpenter Shop for providing stranding equipment. Financial assistance from the Natural Sciences and Engineering Research Council of Canada is gratefully acknowledged. Finally, my greatest gratitude goes to my wife - Jiwen and daughter - Ying for their love, patience, encouragement, understanding and support during my educational studies. x C H A P T E R 1. L I T E R A T U R E O V E R V I E W CHAPTER 1. LITERATURE OVERVIEW 1.1 INTRODUCTION Oriented Strand Board (OSB) is a mat-formed panel made of strands, which are sliced along the longitudinal direction from small-diameter, fast-growing round wood logs, and then bonded with an exterior-type resin under heat and pressure. Due to its specific structure and manufacturing process, OSB has found uses as an engineered wood product that possesses unique mechanical properties. As a matter of fact, OSB is widely used in the construction industry as load-carrying parts like wall sheathing, roof sheathing, subflooring, underlayment, single-layer flooring, structural insulated panel components, and engineered wood I-joists (Smulski, 1997; SBA, 1998; Website information, 1998). In these applications, OSB shear and nail connection characteristics are as important as some other physical and mechanical properties like modulus of rupture (MOR), modulus of elasticity (MOE), internal bond (IB) and thickness swell (TS). It is necessary to investigate OSB to lumber nail connection properties systematically and find out the effect of OSB structural patterns on the nail connection performance. 1.2 RESEARCH BACKGROUND Nails and other type of fasteners are used in wood structures for three main purposes. Firstly they are applied to hold two or more pieces of material together. Secondly, they 1 C H A P T E R 1. L I T E R A T U R E O V E R V I E W are used to transfer the load from one member to another. Finally, they are used to join two or more members together in such a way that they act as a unit. To understand the ability for fasteners to withstand applied loads, usually two kinds of tests are carried out. The first is called withdrawal test with the load acting along the fastener axis, and the second, lateral loading test (or lateral resistance test) with the load acting perpendicular to the fastener axis. This study focused on the lateral resistance properties of OSB to lumber connections. There are a number of publications about wood joints with common nails as fasteners (Kuenzi, 1955; Mack, 1963 and 1966; Leach, 1964; Golebrowski and Mielezarek, 1966; Albert and Johnson, 1967; Noren, 1968; Morris, 1970; Wilkinson, 1971; Patterson, 1973; Debonis and Bodig, 1975; McClain, 1975; Antonides et al, 1979; Jenkins et al, 1979; Stone et al, 1980; Liu and Soltis, 1984; Lau and George, 1987; Blass, 1994; Mohammad and Smith, 1994, 1996; He, 1997; Sieber et al; 1997; Rogerson, 1998). In summary, the following factors were found to impact the properties of a wood joint, either a wood/wood connection or a wood/composite panel connection: • Wood: species, specific gravity, moisture content, grain alignment and straightness, resistance to splitting, thickness of member • Properties of Side Members: e.g. wood, plywood, hardboard, flakeboard, OSB • Nails: size, shape, mechanical performance of nail material, surface condition • Spacing of Nails • Type of Connection: single, double, or multiple shear, pressure between contact surfaces 2 C H A P T E R 1. L I T E R A T U R E O V E R V I E W • Direction of Nailing: radial, tangential or end grain, at 90 degrees or inclined to the surface • Direction of Loading: rate, time between joint assembly and testing, type of loading, e.g. static or dynamic loads • Environmental Factors: e.g. temperature and humidity • Creep in Nailed Joints Many studies have been recently conducted on the lateral resistance properties of commercial OSB (side member) to lumber connections with different types of fasteners (Blass, 1994; Mohammad and Smith, 1994 and 1996; He, 1997; Sieber et al; 1997; Rogerson, 1998). Mohammad and Smith (1996) investigated the effect of lumber moisture content and subsequent fluctuating moisture content on single shear lumber to OSB panel nail connections. They found significant decreases in connection stiffness, especially for kiln-dried lumber, when the connections were exposed to changing moisture content. Sieber et al., (1997) used 50 mm common nails and spiral nails to conduct monotonic and cyclic lateral resistance tests in lumber to OSB panel connections. They found that: 1) the conditioning of the specimen after assembling was important, especially for common nails; 2) the load carrying capacity and the ability to store energy of connections with both nail types were observed to be almost the same, but common nails achieved significantly higher ductility values due to their higher initial stiffness, and 3) the angle of force to lumber grain orientation showed no significant effect on the above results. Their experiments showed that connections with spiral nails subjected both to monotonic and cyclic loads failed in a mixed failure mode, but common nails tended to fail with a slightly higher probability in OSB pull-through than 3 C H A P T E R 1. L I T E R A T U R E O V E R V I E W withdrawal (pull-out). Finally Sieber et al., (1997) discovered that a change of nail spacing from 5 mm to 150 mm had no obvious influence on the basis of nail connection performance. He (1997) studied the failure modes in nail connections and concluded that the nail connections failed in a mixed failure mode, either by nail withdrawal (pull-out) from the lumber or by nail pull-through from OSB panels, and that the wood density and local mechanical properties of panel affected the failure modes. Specimens with lower wood density were more likely to fail in withdrawal, but as the wood density increased, the occurrence of pull-through failure went up. This tendency was observed to be stronger for common nails. Rogerson (1998) carried out a series of lateral nail resistance tests on OSB to lumber connections using 9.5 mm and 12.7 mm OSB panels, 38x89 mm western hemlock lumber and 63.5 mm hand driven common nails. A new test set-up approach was tried (Fig. 1.1). The OSB samples were first cut into 150 mm squares. Along either an edge or a diagonal of the OSB specimen, nails were applied to fasten the OSB panel and lumber together by using four different levels of nailing distances from 9.5 mm to 38.1 mm. This simple set-up made both diagonal and edge nailing tests possible in the same specimen since the square specimen could be turned easily around the central hole. A nailing distance of 6.35 mm was also tried, but significant breakage of the OSB corners happened during nailing, especially with the 9.5 mm thick OSB. The failure modes were divided into the following categories: • Lumber splitting • Nail head pull-through 4 C H A P T E R 1. L I T E R A T U R E O V E R V I E W • Na i l bending • Na i l pull-out • Na i l break • Layer shear of O S B • Break-out of the O S B (to the extent of lateral destruction through to the nearest edge) Load head connection A Rotate to test new positions Lumber/bed connection Fig. 1.1 A new set-up for nail connection tests 5 C H A P T E R 1. L I T E R A T U R E O V E R V I E W Also, the relative frequency of these failure modes was totaled for each set of replications and reported as a percentage of nails causing such failures. The nail connection properties including maximum load, maximum deflection and energy to break were compared according to loading directions. The results showed that when the nailing distance from edge or corner is more than 25.4 mm, high maximum load and energy to break values can be achieved, and that the 12.7 mm OSB had higher maximum load and energy to break values compared to 9.5 mm thick OSB. Wood composites offer the potential to increase both the utilization and the value of low to moderate quality wood resources. To fully understand wood composite structure and its effect on properties is of much importance in making efficient use of wood materials and decreasing the production costs. Several researchers (Harless et al., 1987; Humphreys and Bolton, 1989; Suchsland and Xu, 1989; Kamke and Wolcott, 1991) have used mathematical models to describe wood composite systems. Efforts have also been focused on modeling the strength and elastic properties of short fibre wood composites (Harless et ah, 1987). Although many approaches have been applied, most of the theoretical models are based on assumptions of a perfect alignment of wood elements, optimum bonding between elements, continuous gluelines, and uniform structure. Since these assumptions are not valid for products such as commercial flakeboard, discrepancies are often found between predicted and experimental results (Harless et al., 1987). The influence of random flakeboard and OSB structures on some physical and mechanical properties has also been investigated by several researchers like Dai and Steiner (1993, 1994a, b, c), Wang and Lam (1997, 1999), and Lu (1999). Dai and Steiner 6 C H A P T E R 1. L I T E R A T U R E O V E R V I E W (1993, 1994a, b, c) showed that randomly formed flake mats with random flake positions and orientations can be fully described on a probabilistic basis. Mat structural parameters such as flake centroids, flake coverage and between-flake void sizes are Poisson distributed random variables. Non-uniform flake coverage distribution is an inherent feature of a randomly-formed mat. That is why a horizontal density variation always exists in a random flakeboard. Because of point-to-point spatial correlation of local flake coverage, the variation of flakeboard density averages in finite sampling zones depends on the zone size, flake size, flakeboard thickness and compaction ratio (Dai, 1993). Lu (1999) investigated the structural characteristics of partially oriented flakeboard mats based on a simulation computer program, such as 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 degrees of orientation of flakes in both simulated and experimental mats. The relationships between thickness swelling and mat structures in robot-formed flakeboard mats were also examined. Wang and Lam (1997, 1999) studied the influence of the spatial organization of wood elements inside a three-layer oriented flake mat on the OSB manufacturing process, horizontal density distribution (HDD), and OSB properties, considering three important processing parameters: flake slenderness ratio, flake ratio and board density. Response Surface Method (RSM) was used to build up the relationship between physical and strength properties of robot formed panels and the predefined OSB structures due to different combinations of the three processing parameters. Regression models were built 7 C H A P T E R 1. L I T E R A T U R E O V E R V I E W up relating panel performance including MOE, MOR, IB and TS with flake aspect ratio, flake orientation and board density. Even though there are many studies about OSB to lumber nail connection properties, OSB structure characteristics, and the impact of OSB structure patterns on some OSB physical and mechanical properties, the effect of flakeboard structural features on flakeboard (including OSB) to lumber nail connection characteristics has not been systematically investigated up to date. In this study, preliminary research on OSB to lumber connections (small size sheathing) was conducted first. Based on these results and findings on OSB structural behavior and OSB nail connection properties (large size sheathing) by previous researchers, three principle processing variables, flake thickness, surface flake orientation and board density, were selected in the main experimental design to obtain various OSB structures by different combinations of these three parameters. Considering the important application of OSB as shear walls, OSB nail connection performance must be affected by three parts: a frame made of lumber, sheathing materials (OSB), and fasteners (nails, staples or screws). Therefore, in the main research part of this thesis, besides OSB and/or random flakeboard structural characteristics such as local density values around nailing positions and board to flake thickness ratio, lumber specific gravity, loading directions and failure modes are all discussed in order to set up the relationship of nail connection properties to OSB and/or random flakeboard structures. 8 C H A P T E R 1. L I T E R A T U R E O V E R V I E W 1.3 RESEARCH GOALS The objectives of this study include: • To determine the lateral resistance of nail connection properties of different structure panels. • To find the relationships of 50 mm common nail connection characteristics with OSB local density values, board to flake thickness ratios, lumber specific gravity values, loading directions, and failure modes. Two different nailing sample test set-ups, and the nail connection properties of commercial OSB panels with single nail and double nail connections are firstly presented in Chapter 2. In Chapter 3, an experimental design is introduced on laboratory-based panel structures used for flakeboard to lumber connections in main tests. Mat simulations and laboratory-based panel manufacture processes are also discussed. Chapter 4 focuses on the flakeboard to lumber connection assembly, on tests of various nail connections comprised of SPF lumber and robot-formed panels with different structures, and on the data analysis of nail testing results including t-tests and regression analyses. Chapter 5 draws conclusions from the overall research scopes and also proposes some future potential research work. 9 1 . 4 R E F E R E N C E S Albert, T.J. and J.W. Johnson. 1967. Lateral Holding Capacity of Power-Driven Fasteners. Forest Prod. J., 17(9): 59-67 Antonides, C.E. , M.D. Vanderbilt, and J.R. Goodman. 1979. Interlayer gap effects on nail slip modulus. Struct. Res. Rept. No. 22, Civil Eng. Dept., Colorado State Univ., Ft. Collins, Colo., USA Blass, H.J. 1994. Variation of load-slip behavior in nailed connections: variation parallel to the grain. Forest Prod. J. 44(1): 15-20 Blass, H.J. 1994. Variation of load-slip behavior in nailed connections: variation perpendicular to the grain. Forest Prod. J. 44(2): 30-34 Dai, C. 1993. Modeling structure and processing characteristics of a randomly formed wood-flake composite mat. Ph.D. thesis. Department of Wood Science, UBC, Vancouver, BC, Canada Dai, C. and P.R. Steiner. 1994a. Spatial structure of wood composites in relation to simulation of a randomly formed flake layer network. Part 2. Modeling and simulation of a randomly formed flake layer network. Wood Sci. and Technol., 28 (2): 135-146 Dai, C. and P.R. Steiner. 1994b. Spatial structure of wood composites in relation to processing and performance characteristics. Part 3. Modeling the formation of multi-layered random flake mats. Wood Sci. and Technol., 28 (3): 229-239 Dai, C. and P.R. Steiner. 1994c. Analysis and implication of structure in short fiber wood composites. Second Pacific Rim Bio-Based Composites Symposium, Nov. 6-9, Vancouver, BC, Canada, pi7-24 10 C H A P T E R 1. L I T E R A T U R E O V E R V I E W Debonis, A . L . and J. Bodig. 1975. Nailed wood connections under combined loading. Wood Sci. and Technol. 9(2): 129-144 Golebrowski, Z . and Z . Mielezarek. 1966. Stiffness of Nailed Joints. Forest Prod. J. 16(3): 64 Harless, E . G . and F . G . Waguer. 1987. A model to predict the density of particleboard. Wood Fiber Sci. , 19(1): 81-92 He, M . H . 1997. A Study of Wood Based Shear Walls Sheathed with Oversize Oriented Strand Board Panels. M.Sc . thesis. Department of Wood Science, U B C , Vancouver, B C , Canada Humphrey, P .E. and 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 Jenkins, J .L. , A . Polensek, and K . M . Bastendorff. 1979. Stiffness of nailed wall connections under short and long term loads. Wood Sci. 11(3): 145-154 Kamke, F . A . and M . P . Wolcott. 1991. Fundamentals of flakeboard manufacture: wood-moisture relationships. Wood Sci. and Technol., 25: 57-71 Kuenzi , E . W . 1955. Theoretical Design of a Nailed or Bolted Joint under Lateral Loads. U S Forest Products Lab, Report No. 1951. Madison, Wisconsin, U S A . 31pp. Lau, P . W . C . and P. George. 1987. Development of a load-slip test apparatus for nailed connections. Forest Prod. J. 37(11/12): 39-44 Leach, K . E . 1964. A Survey of Literature on the Lateral Resistance of Nails . Department of Forestry Publication No. 1085. Ottawa, Canada. 11pp. 11 C H A P T E R 1. L I T E R A T U R E O V E R V I E W L i u , J . Y . and L . A . Sol t i s . 1984. Latera l resistance o f na i led connections — a test method. Forest P r o d . J . 34(1): 55-60 L u , C . 1999. Organ iza t ion o f W o o d Elements i n Par t ia l ly Oriented F lakeboard M a t s . P h . D . thesis. Department o f W o o d Science, U B C , Vancouve r , B C , Canada M a c k , J .J . 1963. Study o f Creep i n N a i l e d Joints. C S I R O , Aus t . D i v . For . P rod . T e c h n o l . Paper N o . 27. Aus t r a l i a M a c k , J .J . 1966. The Strength and Stiffness o f N a i l e d Joints under Shor t -Dura t ion L o a d i n g . C S I R O , Aus t . D i v . For . P rod . Techno l . Paper N o . 40. Aus t r a l i a M c L a i n , T . E . 1975. Curv i l i nea r load-s l ip relations i n laterally loaded na i led connections. P h . D . dissertation. Co lo rado State U n i v . , Ft. C o l l i n s , C o l o . , U S A M o h a m m a d , M . A . H . and I. Smi th . 1994. Stiffness o f N a i l e d O S B - t o - L u m b e r Connect ions . Forest P rod . J . 44(11/12): 37-44 M o h a m m a d , M . A . H . and I. Smi th . 1996. Effects o f multi-phase moisture cond i t ion ing on stiffness o f na i led O S B - t o - l u m b e r connections. Forest Prod . J . 46(4): 76-83 M o r r i s , E . N . 1970. A n A n a l y s i s o f the L o a d - S l i p C u r v e for a N a i l e d Joint and the Effects o f Mo i s tu r e Content. J . Inst. W o o d S c i . 5(1): 3-9 N o r e n , B . 1968. N a i l e d Joints — The i r Strength and R i g i d i t y under Shor t -Term and L o n g - T e r m L o a d i n g . M e d d . Svenska Traforskn Inst. (Tralenik) N o . 158B. 80pp. Patterson, D . W . 1973. N a i l e d W o o d Joints under Lateral Loads . M . s c . thesis. Co lo rado State Un ive r s i ty , For t C o l l i n s , Co lo rado , U S A Rogerson , D . E . 1998. Corner and edge na i l ing test results on O S B and hemlock lumber w i t h 2 .5" c o m m o n nails . M B Research M e m o . M a c M i l l a n B l o e d e l L t d . , Vancouve r , B C , Canada 12 C H A P T E R 1. L I T E R A T U R E O V E R V I E W Sieber, D., Lam, F. and H. Prion. 1997. The Behaviour of Nailed Sheathing-to-Frame Connections under Static and Cycc Load. Research Report. Department of Wood Science, UBC, Vancouver, BC, Canada Smulski, S. 1997. Engineered Wood Products: A Guide for Specifiers, Designers and Users. PFS Research Foundation. Madison, USA Stone, J.L., M.D. Vanderbilt, M E . Criswell and J. Bodig. 1980. Generalized load-slip curve for nailed connections. Struct. Res. Rept. No. 30. Civil Eng. dept., Colorado State Univ., Ft. Collins, Colo., USA Structural Board Association (SBA). 1998. OSB Performance by Design™, Printed in Canada Suchsland, O. and H. Xu. 1989. A simulation of the horizontal density distribution in a flakeboard. Forest Prod. J., 39(5): 29-33 Wang, K. 1997. Robot-based research on three-layer oriented falkeboards. M.Sc. thesis. Department of Wood Science, UBC, Vancouver, BC, Canada Wang, K. and F. Lam. 1999. Quadratic RSM Models of Processing Parameters for Three-Layer Oriented Flakeboards. Wood Fiber Sci., 31(2): 173-186 Website information. 1998. Wood-Based Panel Products Technologies IV. Oriented Strand Board (OSB). Address: http://strateKis.ic.gc.ca/SSG/fbol 135e.html Wilkinson, T.L. 1971. Theoretical Lateral Resistance of Nailed Joints. Journal of the Structural Division, ASCE: 1381-1397 13 C H A P T E R 2. P H A S E I. P R E L I M I N A R Y T E S T S O N C O M M E R C I A L O S B P A N E L S C H A P T E R 2. P H A S E I . P R E L I M I N A R Y T E S T S O N C O M M E R C I A L O S B P A N E L S ABSTRACT Preliminary tests on nailed OSB-to-lumber connections were conducted using 11 mm thick commercial OSB panels as side members and SPF lumber as main members. Several combinations of OSB specimen sizes, nailing patterns and test set-ups were investigated. Tensile loads were applied statically along the longitudinal direction of the lumber member, but perpendicular to the nail shank for all specimens. Both the single nail and the two-nail combination patterns were examined in OSB specimen I (50x240x11 mm) and specimen II (240x240x11 mm). Loading directions relative to OSB face flake orientation were studied for specimen II. The results showed that the chosen test jigs were suitable for small sized OSB-to-lumber nailed connections, and the test set-ups with specimen II were more efficient for small scale nail connection testing since they were more flexible to adjust to the loading directions, nailing patterns and multiple nailing; hence, more information could be obtained. Two main failure modes, pull-through and pull-out, were observed in the preliminary tests. The dominating failure mode in an OSB-to-lumber connection depended on the side member (OSB) dimensions, nailing patterns, OSB density variation (local density) and lumber specific gravity. Loading directions had different impacts on various nailing characteristics. OSB local 14 C H A P T E R 2. P H A S E I. P R E L I M I N A R Y T E S T S O N C O M M E R C I A L OSB P A N E L S density and lumber specific gravity obviously affected nail connection properties, which needed further study in the main tests. 15 C H A P T E R 2. P H A S E I. P R E L I M I N A R Y T E S T S O N C O M M E R C I A L O S B P A N E L S 2.1 INTRODUCTION The use of wood and wood-based materials in many structural and other applications often involves mechanical fasteners, which can be divided into two main types: dowel and bearing. Dowel type connections, such as nails, screws, and bolts, transmit either lateral or withdrawal loads. Lateral loads are transmitted by bearing stresses developed between the fastener and the members of the connection, whereas withdrawal loads are axial loads parallel to the fastener axis transferred through friction or bearing to the connected materials. Bearing-type fasteners, like shear plates and split ring connectors, transmit shear forces through bearing on the connected materials (FPL, 1999). Nails are one of the principle fasteners widely used in light-frame building components, such as in OSB-to-lumber connections. They are defined as straight, slender, usually pointed and headed fasteners, designed to be driven and to hold two or more pieces together or to act as support (ASTM F547, 1990). According to A S T M , over one hundred types of nails are classified in light of their forms, materials and applications. Like other dowel connectors, nails can be used to resist withdrawal loads and/or lateral loads. Both withdrawal and lateral resistance are affected by the properties of wood or wood-based products, the nail, and the conditions of use (FPL, 1999). The strength and stiffness of light-frame building structures largely depend on the lateral resistance of the connections between the connected members. Besides the A S T M method (ASTM D 1761, 1995), there are several other test arrangements which can be applied to determine the lateral 16 C H A P T E R 2. P H A S E I. P R E L I M I N A R Y T E S T S O N C O M M E R C I A L O S B P A N E L S resistance of nailed connections consisting of wood framing and various panel sheathing materials under different types of loading conditions (Patterson, 1973; Debonis and Bodig, 1975; McClain, 1975; Antonides et al, 1979; Jenkins et al, 1979; Stone et al, 1980; Liu and Soltis, 1984; Lau and George, 1987; Blass, 1994; Mohammad and Smith, 1994 and 1996; He, 1997; Sieber et al, 1997; Rogerson, 1998). To sum up, nail connection properties vary with different types of wood (species, specific gravity, moisture content, and resistance to splitting etc.), nails (size, shape, surface condition, and mechanical properties), connections (single shear, double shear, or multiple shear etc.), nailing patterns (spacing, edge and side distance, number of nails per row, and number of rows), nailing status (penetration, manually or mechanically, pre-drilling, clinching, gaps between members, and direction of nailing, etc.), direction of loading (relative to lumber grains or composite orientation), duration of loading (loading rate, time between nail-driving and testing, type of loading), test set-up methods, side member material (sheathing like plywood, hardboard, OSB) properties, and creep in nailed connections. The objectives of the preliminary tests are: • To verify the feasibility of a modified jig in a small sized OSB-to-lumber connection testing; • To choose a side member (OSB specimen) dimension and nailing patterns which can make full efficient use of laboratory-based small sized panels (240x240 mm); • To briefly discuss the impacts of failure modes, OSB density variation (local density) and lumber specific gravity on nail connection properties. 17 C H A P T E R 2. P H A S E I. P R E L I M I N A R Y T E S T S O N C O M M E R C I A L O S B P A N E L S Commercial OSB panels were cut into different sizes (max size: 240x240 mm, equal to laboratory-based panel dimensions in the main tests), and used as side members of the OSB-to-lumber connections, and SPF lumber was chosen as main members of the connections. Common nails fastened OSB and lumber together according to various experimental requirements. The connections were tested in a universal material testing machine (MTS) to determine the connection performance. 2.2 MATERIALS AND METHODS 2.2.1 Materials OSB: 11 mm thick OSB panels, nominal board density 0.650 g/cm , and measured average density 0.643 g/cm Lumber: SPF, 38x89x420 mm (cut along the longitudinal direction 2x4 lumber), uniform, no defects and knots around nailing locations, moisture content (LMC) of 11-12%, specific gravity (G, oven-dried) of 0.40-0.50 Nail: 6d common nail, TREE ISLAND magnetic pack nails (ICBO no. 4266), length = 50 mm, diameter = 2.87 mm, and head diameter = 5.74 mm 2.2.2 Methods 2.2.2.1 Nailing Patterns and Test Set-ups A 1.22 m x 2.44 m OSB panel was cut into half along the central line of width. One half was used for the determination of nailed connection performance, and the other for the measurement of shear modulus (extra tests, not included in this thesis). 18 C H A P T E R 2. P H A S E I. P R E L I M I N A R Y T E S T S O N C O M M E R C I A L OSB P A N E L S Specimens (specimen I: 50x240x11 mm and specimen II: 240x240x11 mm) were randomly cut around the neighbor positions from the first half panel for connection property tests. In addition, three 75x75 mm specimens were sawn from the leftover material to measure OSB average moisture and density (CSA 0 4 3 7 , 1993). Typically the design of wood structures requires the provision of sufficient end distance, edge distance, and spacing in order to avoid unusual splitting (AITC, 1966; NFPA, 1971; FPL, 1999). Koch (1972) provided guidelines of a minimum end distance for the loaded end as 13-20 nail diameters, and a minimum nail spacing of 12-15 nail diameters in the loading direction. Ramos (1960) also provided information supporting the use of 10 nail diameters as the minimum edge distance. Except the edge distance (25 mm < 10x2.87=28.7 mm), all other connection spacing parameters in specimen I and specimen II met the above requirements in the preliminary and main tests. Two specimen dimensions and nailing patterns are shown in Figs. 2.1 and 2.2. After conditioning for more than two weeks at 20°C and 50% relative humidity (RH), the OSB and SPF lumber specimens were ready to be assembled into different types of connections. The 50 mm 6d nails were driven into OSB and lumber manually by hammer, keeping the nail heads flushed with the OSB surface to prevent OSB damages before loading. Normally, the connections were tested within 30 minutes after assembly. Two nail test set-ups, shown in Figs. 2.3-2.5, were considered for two types of OSB specimens, respectively. 19 C H A P T E R 2. P H A S E I. P R E L I M I N A R Y T E S T S O N C O M M E R C I A L O S B P A N E L S <516 50 240 50 50 <D16 50 240 50 A l l dimensions in mm 50 50 (A) Two-nail combination (B) Single nail Fig. 2.1 Nailing format for 240x50x11 mm OSB -016. X 50 50 45 240 (A) Two-nail combination along diagonals 50 45 / ,J 240 (B) Single nail along diagonals A l l dimensions in mm (A) and (B) Single nail for two Sides Fig. 2.2 Nailing pattern for 240x240x11 mm panels 20 C H A P T E R 2. P H A S E 1. P R E L I M I N A R Y T E S T S O N C O M M E R C I A L O S B P A N E L S P/2 A A p/2 P/2 A Top bolt 016 mm 50 mm common nail i-f SPF lumber 38x89x420 mm Bottom clamp system OSB L V D T (A) Two-nail combination per row (B) Single nail per row Fig. 2.3 Test set-up diagram for 240x50x11 mm OSB 21 C H A P T E R 2. P H A S E I. P R E L I M I N A R Y T E S T S O N C O M M E R C I A L O S B P A N E L S C H A P T E R 2. P H A S E 1. P R E L I M I N A R Y T E S T S O N C O M M E R C I A L O S B P A N E L S The SPF lumber located at the end of the connection was fixed to the MTS 810 universal testing machine table, while OSB panels situated at the upper end of the connection were connected to the MTS loading unit. A unidirectional tensile load was laterally applied to the specimen to cause a shear deformation. The loading rate was 2.54 mm/min for all tests. The displacements between OSB panel and lumber were measured by a Linear Voltage Displacement Transducer (LVDT) mounted on the connection. For specimen I, a tensile load was applied along the SPF grain direction and OSB orientation either in the single nail pattern or the two-nail combination format. For specimen II, a specific nailing pattern was prepared to gain information on nail connection properties along two diagonals (45° and 135° relative to OSB orientation) and two sides (parallel and perpendicular to OSB face flake alignment respectively), using the single nail and the two-nail combination patterns for two diagonals, and only the single nail format for two sides, as shown in Figs. 2.4 and 2.5. 23 C H A P T E R 2. P H A S E I. P R E L I M I N A R Y T E S T S O N C O M M E R C I A L O S B P A N E L S To ensure the feasibility of the new jig in small OSB sample testing and data reliability, eccentricity and gap between main member (lumber) and side member (OSB) were examined and recorded qualitatively for each run. Other conditions were: L V D T range: ±25.4 mm; Machine displacement range: ± 37.2 mm; Max load: ± 25000 kNf A list of variables in the preliminary tests is shown in Table 2.1. Table 2.1 Number and type of tests conducted OSB specimen Specimen I: 50x240xllmm Specimen II: 240x240x11mm Nailing pattern See Fig. 2.1 See Fig. 2.2 Number of nails per measurement (1) The single nail (2) The two-nail combination (1) The single nail (2) The two-nail combination Loading direction relative to OSB orientation Along flake orientation 0°(side), 45°(diagonal), 90°(side), 135°(diagonal) relative to OSB alignment No. of specimens 24 (12 for the single nail, 12 for the two-nail combination) 12 (6 for the two-nail combination along two diagonals plus the single nail along two sides, 6 for the single nail along two sides and two diagonals) No. of tests 36 (24 for the single nail, 12 for the two-nail combination) 100 (24 for the two-nail combination, 76 for the single nail) No. of nails 48 (24 for the single nail, 24 for the two-nail combination) 124 (76 for the single nail, 48 for the two-nail combination) 24 C H A P T E R 2. P H A S E I. P R E L I M I N A R Y T E S T S O N C O M M E R C I A L O S B P A N E L S 2.2.2.2 Load-Slip Curves and Definitions of Nail Connection Properties Under static loads, OSB to lumber connections deformed due to OSB and lumber bearing damage and nail bending. The changes of loads and displacements with time were recorded automatically using a PC based data acquisition system. Typical load-slip curves are shown in Figs. 2.6-2.8. Original load-slip curves were analyzed using a computer program to extract the key nail connection properties automatically. A Binomial smooth procedure was also applied to the original data to remove some noises by filtering. Fig. 2.6 Load-displacement curve for nail connection tests 25 C H A P T E R 2. P H A S E I. P R E L I M I N A R Y T E S T S O N C O M M E R C I A L O S B P A N E L S Definitions of nail connection properties are as follows (Fig. 2.6): K: initial slope or stiffness of load-displacement curve, N/mm Pmax' maximum load, N displacement corresponding to maximum load, mm Umax- strain energy at maximum displacement, the integration of the load-displacement curve from zero displacement to Amax, N.mm Pyid-' yield load, corresponding to 50% maximum load, N Ay/a: displacement corresponding to yield load, mm Uyu: strain energy at yield displacement, the integration of the load-displacement curve from zero displacement to Ayu, N.mm Pui,: ultimate load, corresponding to the post peak load at 80% of maximum load, N Auit: displacement corresponding to ultimate load, mm U„it: strain energy at ultimate displacement, the integration of the load-displacement curve from zero displacement to A„ih N.mm Z>/ : ductility factor one = Aui/Amax D2: ductility factor two = Amax/Ayid Two principal failure modes, pull-through and pull-out, were observed. They were defined from characteristics of the specimen damage (especially OSB part). Pull-out failure means that the connection failed because of significant bending of the nail with the nail eventually pulling out of the lumber. Typically the nail head remained on or 26 C H A P T E R 2. P H A S E I. P R E L I M I N A R Y T E S T S O N C O M M E R C I A L O S B P A N E L S above the top surface of the OSB member and the OSB nailing area had neither damage nor other distinctive changes. Pull-through failure means the connection failed because of the nail head intruded significantly into the OSB panel resulting in the loss of load-carrying capacity in the connection. Usually, a certain degree of the nail withdrawal or pull-out accompanied the ultimate pull-through failure mode, but for simplicity these contributions were generally ignored if they were not obvious enough to be differentiated from two main failure modes. Typically, the load-slip curve of a pull-out mode was relatively smooth before maximum (peak) load and showed fluctuating post peak responses (see Fig. 2.7). The load-slip curve of a pull-through mode, however, had an abrupt, sometimes irregular change in loads after peak load was reached, as displayed in Fig. 2.8. 1400 1200 1000 _ 800 •D ra ° 600 400 200 0.8 Max: 987 19.076 20388.8 31 0.5 Max: 616 .432 174.4 1.0 Max: 1233 9.779 10465.5 j . i i . J i_ 10 15 Displacement (mm) 20 Fig. 2.7 Typical load-displacement curve for pull-out failure 25 27 C H A P T E R 2. P H A S E I. P R E L I M I N A R Y T E S T S O N C O M M E R C I A L O S B P A N E L S Displacement (mm) Fig. 2.8 Typical load-displacement curve for pull-through mode 2.2.2.3 Lumber Specific Gravity and OSB Local Density After the connection tests, two 38x89x20 mm blocks were cut from every tested main member (lumber) to measure its moisture content and specific gravity (ASTM D 2395, 1993). Method B - volume by water immersion was applied. The blocks were weighed, and then put into a 103±2 °C oven for drying until a stable weight reached. The dried blocks were immersed into water (method B-II) to determine their volumes by measuring the volume of water displaced or by determining the weight of water displaced. The weight in grams is numerically equal to the volume in cubic centimeters. Finally lumber 28 C H A P T E R 2. P H A S E I. P R E L I M I N A R Y T E S T S O N C O M M E R C I A L O S B P A N E L S moisture content and specific gravity were calculated respectively according to weights and volume measured. In order to explore the impact of OSB local density on nail performance, 15x15 mm OSB specimens were cut by means of a small narrow band saw. Before sawing, exact square boundaries were drawn around the nailing position so as to locate representative nailing areas. Only pull-out OSB specimens and pull-through OSB specimens without loss of materials during nail property testing were considered for local density data analysis. All the local density values are based on air-dried weight of tested OSB specimens. 2.3 R E S U L T S A N D DISCUSSION 2.3.1 Results The preliminary tests consisted of two parts. One was 240x50x11 mm small rectangular OSB specimen (specimen I) testing, and the other 240x240x11 mm square OSB specimen (specimen II) testing. The single nail and the two-nail combination patterns were set up for both cases. The nail properties of the connections with different OSB specimen sizes and nailing patterns are summarized in Table 2.2. Loading directions relative to OSB face orientation are compared for specimen II, as shown in Table 2.3. Due to the small sampling size, only basic statistics were obtained here in preliminary tests. Various tendencies may be easily observed from histograms, and other graphs (Figs. 2.9-2.18). 29 C H A P T E R 2. P H A S E I. P R E L I M I N A R Y T E S T S O N C O M M E R C I A L O S B P A N E L S From Tables 2.2 and 2.3, it can be seen that different nail properties have varied coefficient of variation (COV) . Initial stiffness had the largest C O V . Loads that include maximum load, yield load and ultimate load had the smallest C O V . Coefficients of variation are dependent on many factors such as material variances, testing errors and other different experimental conditions. Here in this chapter, average nail connection properties w i l l be chiefly discussed. 2.3.2 Discussion 2.3.2.1 Test Set-ups Four different set-ups with varied specimen sizes and number of nails per row led to diverse nail connection properties (see Table 2.2 and Figs. 2.9-2.12). • Initial stiffness • Max load 2500 2000 S 1 5 0 0 u M 3 IOOO •si I 500 - 1600 - 1400 - 1200 1000 800 load, N 600 Max -- 400 1- 200 1 o Specimen I: Specimen II: Specimen I: Specimen II: single single combination combination Set-up Fig. 2.9 Relationship of test set-up to stiffness and maximum load per nail 30 Vi - J w Z < PH CC Vi O < U fi* UJ O U z o t/5 H Vi w H >H 05 < Z | w Pi CM W tn a PM r i Pi W H CH < 33 -a fl CS I/! fl OJ u o a w 03 a u a bxi fl C fl o s-.4> T3 O H r-j a 3 H o QH 60 a c to .1 o <u O . \Vi S? 5? LMC, i oo o ov o in -3-o o CN u S _i TI-CS cK m o HI o o o J s 3 m o oS m o O T1-o o J S _ J Tt CA Tt o Tt o o Tt rs CN es o o o (N o I T ) <n 00 m o rs CA O Tt VJD (N O cn vO o O O o o CN oo o o TT o Tf Tt o o o o o o O o O o rs "E u "Si a* . J o VO Ov O O cn cn o OO E o a' Co m O N O Tt o E u " M a' o cn rs OO CA E a o m m CA O vo Tf Tt 3 o O o 3 o o O O O o rs 3 <f E < I a r~ in ov cn CN cn VO r-o o CN 1 < I a m oo o o ^ ' E < 1^  a o CO O ^ ' E < I Q OO oo \6 OO Tt OO rs o m o Tt rs i E i < I a in vo cn o vo O CN O CN < I a m vn O rs o m <d> o I a Tt Tt o r-(N O < I fi ^t- CO cn o ON rs o Tt rs E E z 3 00 Ov r~~ •sr cn vo 00 cn vO CN O CN E £ Z 3 o\ CO o o m rn r-00 in rs d o E E z 3= TJ-od o (N in o m o E E Z CA O CA O cn CN cn OO rs CO CA m cn o Tt rs c <u c o u E E < CN cn oo CN CN O CN o u. l_ E E < oo r--o o o o J _ 1— <D D-E E < O r~i cn (N m o r» o 1— Li 1> Q-t-n 'c5 o H CD E E < CA O oo rs o vO rs o Tt rs <u L. 3 "S E Z CH= vn o O0 cn «n o CN <U Q. CO ' 3 Z CH (N o o c Z CH= r-r-o cn o (N oo CO o r-Z Cu m cn CA <n CA rs cn «n o Tt rs Xi E _3 T E E Z 3" -a-cn cn CN •n cs r~ o CN c o H e a> E E Z 3 r-O <N Q\ cn cn O o _c <u •S E E Z 3' rn (N cn CO ON o cn >n o r-E E Z S rt CA r-cn r-m r-m vo Tt o Tt rs i u 3 IT XI E B c E o d> O H Vi E — E E <" VO o o o VO VO O CN 'o ID O . E E < oo o (N o' o o cn O o E E <" r-o O CA tN OO o \o r-' o (D CL, C/) E E <" Tt o rs o vO CO rn o Tt rs i . c/) o Z a." oo CN CN cn in o CN z CH" oo m O o Z CH" m MD (N oo oo o \o Z CH" oo o CN Tt OO cn in O Tt rs > n u oc u a 1 o « « E E Z i 3 cn •* o o •^ r vb in cn cn cn cn o •a-CN E E z 1 3 oo in rs (N oo (N o Tt o o E E Z i 3 O o' MD CO CA m Tt CA Tt cn o TJ-O E E Z E 3 cn o ro CO CO o *o 00 r-Tt O Tt rs DD E E E E < cn oo CN Ov M-CN O CN E E 1 < cK O o o o o E E E < 00 rs (N o 00 (N O r-E E E < VO rn rs Tt m cn o Tt rs "« c •a c 3 O z E OH VO in cn CN CN cn in o CN Z E CH o r-r-(N CT\ rN m o o z i CH •o Tt m ro oo oo o \o r-z E CH r-Tt <N CA vO CI m m o Tt rs ea & c C I •a K, N/mm vo in ov ov »n Ov 00 in o CN E E Z v-> o\ CO m Tt Tt LT) o *n cn o o E E Z *^ o o o o in CO m cn r-o E E Z m r-^  m (N cn 00 Tt r-cn rs rs Tt O Tt rs CM u o a O a 3 Statistics O > < > a H > P O Z Vi _o 73 (>5 o > < > H > O z y ca 55 O > Q H > O u « 55 o \> \< > U J Q H > O Z V o Z fi E u a w SB U <U a o u a '3 fi fi o fi O a> -3 CD fi -3 « o « -2 '3 | E o -*-> j 5 15 I OH c 'I 1,3 LMC, % n Ov 30 [N t d o d 5? i 30 X> vTl •N o J S ON' d O d i - i ON m -o d ON -o O d N U Ov 5 -—i O d Ov Z> d LD O ON tN O O ~-o d d rN o d o d rN vO d so ro O d ON O d TT •N "E "Si ef - i o r-VO O •3-(N O in 00 d E u "£b ef J 0 0 O m o o o CN d s d _j ro d vo d ON OO d CN E DX ef in ON m d oo d ON ON d CN <r i Q r~ CN O VO Ov •*r r^ d ^ ' < II Q r- in ON d ^ ' i e> OO ro vo d O 0 ro VD d CN ^" f a ro O 0 od o »n ON ON d fN E <] < II Q oo >r> •<t o 00 CN O •* E < < II a r- m >n o CN O ro d E < II D C C r- CO d ON d "3-CN E < II a o r-. r-d vo CN d CN E E Z 3 Os CO VO OO o vd r^ o CN CN d o O ON c o *-> o <u u E E z 3 in in cn (N r-CN d E E Z 3 ON CN o vo d ro oo t^-CN ro d CN £ E z 3 = oo in oo C C '3-oo o m ON ro ro d CN e E 3 < co co o CN co >n o 5 a o B e E E < ro d a o "cd o E E < CN CN wo VD d CN c o E E < ON CN CN CN CN d CN z 6? O co o O vo VO O .S *c O ^ z O o (N o OO d Z a.= ON O 0 O VO ON o OO d t^" CN Z C= o o ro ro CN CN d t^ CN E E Z 3 oq o o CO co od co Ov d r^ o 3 o E E Z 3 rO CN (N (N CN (N d o 1 0 CO E E Z 3 " CN ro ro CN C O vo d CN o 1 o 1 o in E E Z 3 ' r— CN ro ro CN ON ro d Tt CN E E < r~-o d vo *t vo O T3 C CD O . <u OH 0 0 . 3 -a cu o E E <" r-o ro O CN d E E <' d in d CN VO d CN E E <' d OO d ro O N O Tt CN Z BM" VD vo o vo VO O Z e." m fN o oo d Z Cu"' 0 0 \o CN CN o 00 d 'Tf CN Z sT ON V D V D in r-CN d Tt CN E E Z E 3 VO CN 00 CN co co o o >—) E E Z i 3 O O ON rs UO •^ t ro ro ro ON uo ro d E E Z E 3 CN VD ro C O O r-^  oo ro ro m O d CN E E Z i 3 in d ro ON C O in od CN ON ro o d Tt CN E E 1 < CN Ov r-CN d r^ E E i < CO O CN ro CN d E S i < ON CN VO CN d CN E E E < O od vq CN ON ro d Tt CN Z 1 a. 00 oo CN co CN VO VO d Z i P. ON m >/-) (N o 0 0 d Z E £. CN V D ro m CN o oo d Tt CN z E 0. oo ro ro ON CN CN d Tt CN K, N/mm 00 tN VO CO vo vq d t E £ Z ON CN od ON ON o C O d E E Z -<f vo in CN CN in in 0 0 ON vq d CN E E Z vq CN o m C O ro C O o C O d Tt CN Statistics O > < > hi a r-> O o Z « 5 5 > > Q H > O U z. o « 5 5 u > < > UJ Q H > o u z 5 5 O > < > o H in > O Z CN CO u. a. o C H A P T E R 2. P H A S E I. P R E L I M I N A R Y T E S T S O N C O M M E R C I A L O S B P A N E L S For the connection of the single nail per row, specimen I has slightly higher maximum load, but much bigger initial stiffness than specimen II. For the connection of the two-nail combination per row, specimen I has higher initial stiffness and maximum load than specimen II (Fig. 2.9). Due to their proportional relationships to maximum load, yield and ultimate loads have the same tendencies as maximum load. On the whole, OSB specimen dimensions have obvious impacts on three kinds of loads and initial stiffness. The small OSB specimen (50x240x11mm) may have tighter contact with the piece of lumber (main member), which produces higher friction between the two members, leading to higher loads and initial stiffness. Two different nailing patterns, the single nail and the two-nail combination, have little change in maximum load per nail, but large variation in initial stiffness per nail. For specimen I, the single nail pattern has very close loads per nail to that of the two-nail combination but lower initial stiffness than the latter. For specimen II, loads per nail of the single nail pattern are larger than those of the two-nail combination, but initial stiffness is much smaller than that of the latter instead, which seems conflicted with the results from some researchers (Ramos, 1960; Leach, 1964), but consistent with others (Sieber etal., 1997). As shown in Fig. 2.10, for the single nail pattern there is not much difference in the three types of displacements between the two specimen sizes. However, for the two-nail combinations significant differences exist in these slips between specimen I and specimen II. For specimen I, the two-nail combination pattern has higher maximum, yield and ultimate displacements than the single nail pattern. In contrast, for specimen II, the two-nail combination pattern has smaller maximum, yield and ultimate displacements 33 C H A P T E R 2. P H A S E I. P R E L I M I N A R Y T E S T S O N C O M M E R C I A L O S B P A N E L S than the single nail pattern. Here, the displacements for the two-nail combination are the total displacements. 16 14 12 S 1 0 I * . K 6 4 2 0 U Max • Yield • Ultimate Specimen I: Specimen II: Specimen I: Specimen II: single single combination combination Set-up Fig. 2.10 Impact of test set-up on three displacements For the single nail pattern, yield and ultimate strain energies of specimen I are similar to those of specimen II. Higher max strain energy is found for specimen I compared to specimen II. For the two-nail combination, the three strain energies of specimen I are all higher than those of specimen II. In specimen I, the strain energies of the single nail pattern are all lower than those of the two-nail combination pattern. However, specimen II has higher energies for the single nail pattern compared to those for the two-nail combination (Fig. 2.11). 34 C H A P T E R 2. P H A S E I. P R E L I M I N A R Y T E S T S O N C O M M E R C I A L O S B P A N E L S • Max • Yield • Ultimate 18000 16000 1 14000 £ 12000 go 10000 o s tu s 2 1/1 8000 6000 4000 2000 0 Specimen single Specimen II: single Specimen I: combination Specimen II: combination Set-up Fig. 2.11 Influence of test set-up on strain energy per nail • Ultimate/max • Max/yield 1 1 1 Specimen I: Specimen II: Specimen I: Specimen II: single single combination combination Set-up Fig. 2.12 Effect of test set-up on ductility factors 35 C H A P T E R 2. P H A S E I. P R E L I M I N A R Y T E S T S O N C O M M E R C I A L O S B P A N E L S From Fig. 2.12, it is clear that ductility factor one (ultimate displacement/maximum displacement) is very close for all test set-ups, but ductility factor two (maximum displacement/yield displacement) has distinct differences between the two test set-ups, which demonstrates that at the beginning of tests there were more differences in slip changes before maximum load appeared; afterwards the load-slip curves tended to be flat (Figs. 2.6-2.8). For the single nail pattern, specimen I has larger ductility factor two than specimen II, but for the two-nail pattern, specimen I received lower value. For specimen I, the single nail pattern obtained higher ductility factor two dhan the two-nail combination; however-, for specimen II the ductility factors of the two patterns were close. 2.3.2.2 Failure Modes As discussed above, two main failure modes, pull-through and pull-out, were defined according to OSB surface damage and nail position along OSB thickness direction after connection testing. For the two failure modes, different specimen sizes and nailing patterns exhibit varying nail properties, as shown in Fig. 2.13. In general, pull-through and pull-out modes occurred in every type of connection with different specimen sizes and nailing patterns. The single nail pattern of specimen I has 46% pull-through mode and 54% pull-out failure mode, and the two-nail combination pattern of specimen I has 70% pull-through and 30% pull-out. In specimen II, however, both the single nail and the two-nail combination patterns have about 43% failures from pull-through and 57% from pull-out. 36 C H A P T E R 2. P H A S E I. P R E L I M I N A R Y TESTS O N C O M M E R C I A L OSB P A N E L S Fig. 2.13 Relationship of failure modes to stiffness and maximum load per nail P U L L - T H R O U G H A: Specimen I (the single nail); B: Specimen I (combination); C: Specimen II (the single nail); D: Specimen II (combination) P U L L - O U T E: Specimen I (the single nail); F: Specimen I (combination); G: Specimen II (the single nail); H: Specimen II (combination) For comparison purpose, initial stiffness and maximum loads for the two-nail combination are expressed as per nail basis. For the pull-through mode in specimen I (A and B), the two-nail combination pattern has larger initial stiffness, but lower maximum load than the single nail pattern. For specimen II (C and D), both the single nail and the two-nail combination exhibited similar maximum loads; however, the two-nail combination has much higher initial stiffness than the single nail pattern. For the same single nail pattern (A and C), specimen I has higher initial stiffness and maximum load 37 C H A P T E R 2. P H A S E I. P R E L I M I N A R Y T E S T S O N C O M M E R C I A L O S B P A N E L S than specimen II; nevertheless, for the two-nail combination (B and D), similar tendency between two specimen sizes could be observed. For the pull-out mode in specimen I (E and F), the single nail and two-nail patterns have almost same maximum loads and initial stiffness. For specimen II (G and H), the two-nail combination has lower maximum load but higher initial stiffness than the single nail pattern. For the single nail pattern (E and G), specimen I has similar maximum load but much larger initial stiffness compared to specimen II. For the two-nail combination (F and H), both initial stiffness and maximum load of specimen I are higher than those of specimen II. For the same test set-ups (A and E, B and F, C and G, E and H), there are obvious differences between the two main failure modes. Except C and G (specimen II with the single nail pattern), other set-ups all have higher initial stiffness and maximum load for the pull-through mode than those for the pull-out mode. Other nail connection performance such as different displacements, strain energies and ductility factors also changes with failure modes for the different specimens and nailing patterns. This information, however, is not discussed in detail in this section. 2.3.2.3 Loading Directions For specimen II, different loading directions were applied. In Figs. 2.14-2.17 the mean nail properties are plotted as a function of loading directions (degrees to face flake alignment). All specimens were assembled using one-nail pattern along two diagonals and two sides. 38 C H A P T E R 2. P H A S E I. P R E L I M I N A R Y T E S T S O N C O M M E R C I A L O S B P A N E L S Fig. 2.14 Effect of loading direction on stiffness and maximum load 16 14 12 | 10 « 6 4 2 0 • Max • Yield —A— Ultimate 45 90 L o a d i n g d i rec t ion to or ien ta t ion , degrees Fig. 2.15 Loading direction and displacements 39 C H A P T E R 2. P H A S E I. P R E L I M I N A R Y T E S T S O N C O M M E R C I A L O S B P A N E L S •Max 9 Yield —A— Ultimate 18000 16000 14000 12000 cu a CD s S £jo 10000 8000 • 6000 4000 2000 0 45 90 L o a d i n g d i rec t ion to or ientat ion, degrees Fig. 2.16 Influence of loading direction on strain energy • Ultimate/max • Max/yield e 20 18 16 14 + 12 10 2 * 0 45 Load ing direct ion to orientation, degrees 90 Fig. 2.17 Relationship of loading direction to ductility factors 40 C H A P T E R 2. P H A S E I. P R E L I M I N A R Y T E S T S O N C O M M E R C I A L O S B P A N E L S The angles of 0° and 90° stand for loading parallel and perpendicular to two sides of specimen II, and that of 45° (or 135°) for loading along two diagonals at such an angle relative to the OSB face flake alignment. Maximum load increases with the increased angle and at 90°, reaches the peak value. Maximum initial stiffness is reached at 90° and minimum stiffness is found at 45° (Fig. 2.14). No distinct differences can be observed in various displacements between different loading directions (Fig. 2.15). The strain energies increase with the increase of angles but have different changing rates. Maximum and ultimate strain energies exhibit more obvious changes with loading directions (Fig. 2.16). Ductility factor one changes with loading directions, and its maximum value occurs at 45°. However, ductility factor two seems not affected by loading directions (Fig. 2.17). 2.3.2.4 OSB Local Density, Lumber Specific Gravity and Moisture Content As shown in Fig. 2.18, the pull-through failure mode has lower OSB local density (LD) values than the pull-out failure mode regardless of specimen size and nailing patterns. Also, lumber specific gravity (G) varied amongst the different specimen testing set-ups, even though the lumber was randomly chosen and cut, and then applied in the connection assembly. Different nail properties are caused by many variables in the OSB-Lumber nailed connection system. OSB local density and lumber specific gravity are two important factors worth further studying in the main experiments. Due to the limited number of specimens and nails, it is hard to find an exact relationship of OSB local density and lumber density to nail properties from preliminary experiments even though the relation 41 C H A P T E R 2. P H A S E 1. P R E L I M I N A R Y T E S T S O N C O M M E R C I A L O S B P A N E L S may exist theoretically and objectively. In the main tests, these two factors w i l l be considered and discussed in detail. 1 2 3 4 5 6 7 8 Failure modes and nailing patterns Fig. 2.18 OSB local density and failure modes PULL-THROUGH: 1—specimen I, the single nail; 2—specimen II, the single nail; 3—specimen I, the two-nail combination; 4—specimen II, the two-nail combination P U L L - OUT: 5—specimen I, the single nail; 6—specimen II, the single nail; 7—specimen I, the two-nail combination; 8—specimen II, the two-nail combination Individual O S B specimen and lumber moisture contents were also measured. The average O S B moisture content was around 6.5% and the lumber moisture content was about 9.0%). The coefficients of variation of O S B and lumber moisture content were 4.2% and 4.5%, respectively. Moisture conditions of lumber and O S B members were very stable for all connections. This is important because higher and inconsistent moisture content 42 C H A P T E R 2. P H A S E I. P R E L I M I N A R Y T E S T S O N C O M M E R C I A L O S B P A N E L S may affect nail performance. In fact, Mohammed (1994) found OSB jointed to dried lumber could be up to three times stiffer than that jointed to green lumber. Therefore, conditioning OSB and lumber for a certain time to arrive into an equilibrium status is critical before assembling the OSB-to-lumber connection. In the main tests, OSB and lumber moisture contents were not taken as variables. 2.4 CONCLUSIONS From the preliminary tests, two different specimen sizes (240x50x11 mm rectangular and 240x240x11 mm square) and two nailing patterns (the single nail and the two-nail combination) were studied using self-designed jigs in the MTS testing machine. The following conclusions could be made: • The jig is suitable for small OSB-lumber nailed connection tests without much eccentricity. • Specimen I had more contact with lumber leading to more friction, which caused higher initial stiffness and loads. • Specimen II was more flexible for different nailing patterns, and more nails can be nailed in a certain area compared to specimen I, therefore, more information on nailing characteristics can be obtained. • Nail connection properties were influenced by failure modes to a certain degree, and loading directions had different impacts on varied nail properties. 43 C H A P T E R 2. P H A S E I. P R E L I M I N A R Y T E S T S O N C O M M E R C I A L O S B P A N E L S • OSB local density and lumber specific gravity may impact nail properties, which needs to be further studied in the main experiments. 2.5 REFERENCES American Institute of Timber Construction (AITC). 1966. Timber Construction Manual. John Wiley and Sons, Inc., New York. American Society for Testing and Materials. 1990. Standard Terminology of Nails for Use with Wood and Wood-Base Materials. A S T M F 547-77 (Reapproved 1990), A S T M , West Conshohocken, Pa., USA American Society for Testing and Materials. 1993. Standard Test Methods for Specific Gravity of Wood and Wood-Base Materials. A S T M D 2395-93, A S T M , West Conshohocken, Pa., USA American Society for Testing and Materials. 1995. Standard Test Methods for Mechanical Fasteners in Wood. A S T M D 1761-88 (Reapproved 1995), A S T M , West Conshohocken, Pa. Antonides, C.E., M.D. Vanderbilt, and J.R. Goodman. 1979. Interlayer gap effects on nail slip modulus. Struct. Res. Rept. No. 22, Civil Eng. Dept., Colorado State Univ., Ft. Collins, Colo., USA Blass, H.J. 1994. Variation of load-slip behavior in nailed connections: variation parallel to the grain. Forest Prod. J. 44(1): 15-20 Blass, H.J. 1994. Variation of load-slip behavior in nailed connections: variation perpendicular to the grain. Forest Prod. J. 44(2): 30-34 44 C H A P T E R 2. P H A S E I. P R E L I M I N A R Y T E S T S O N C O M M E R C I A L O S B P A N E L S Canadian Standards Association. 1993. Standards on OSB and Waferboard: Forest products. 0437 Series-93. Rexdale (Toronto), Ontario, Canada. Debonis, A .L . and J. Bodig. 1975. Nailed wood connections under combined loading. Wood Sci. and Technol. 9(2): 129-144 Forest Products Laboratory (FPL), Forest Service. 1999. Wood Handbook. USDA Forest Serv. Madison, Wis., USA He, M.H. 1997. A Study of Wood Based Shear Walls Sheathed with Oversize Oriented Strand Board Panels. M.Sc. thesis. Department of Wood Science, UBC. Vancouver, BC, Canada Jenkins, J.L., A.Polensek, and K . M . Bastendorff. 1979. Stiffness, of nailed wall connections under short and long term loads. Wood Sci. 11(3): 145-154 Koch, P. 1972. Utilization of the Southern Pines. Agriculture Handbook No. 420. USDA-Forest Services. U.S. Government Printing Office. Lau, P.W.C. and P. George. 1987. Development of a load-slip test apparatus for nailed connections. Forest Prod. J. 37(11/12): 39-44 Leach, K.E. 1964. A survey of literature on the lateral resistance of nails. Dept. of Forestry Publication No. 1085. Ottawa, Ont., Canada Liu, J.Y. and L.A. Soltis. 1984. Lateral resistance of nailed connections — a test method. Forest Prod. J. 34(1): 55-60 McLain, T.E. 1975. Curvilinear load-slip relations in laterally loaded nailed connections. Ph.D. dissertation. Colorado State Univ., Ft. Collins, Colo., USA Mohammad, M.A.H. and I. Smith. 1994. Stiffness of nailed OSB-to-lumber connections. Forest Prod. J. 44(11/12): 37-44 45 C H A P T E R 2. P H A S E I. P R E L I M I N A R Y T E S T S O N C O M M E R C I A L O S B P A N E L S Mohammad, M.A.H. and I. Smith. 1996. Effects of multi-phase moisture conditioning on stiffness of nailed OSB-to-lumber connections. Forest Prod. J. 46(4): 76-83 National Forest Products Association (NFPA). 1971. National Design Specification for Stress-Grade Lumber and Its Fastenings. Washington, D.C., USA Patterson, D.W. 1973. Nailed wood connections under lateral loads. M.Sc. thesis. Colorado State Univ., Ft. Collins, Colo., USA Ramos, A.N. Jr. 1960. Spacing of Sixpenny and Eightpenny Wire Nails in Douglas-Fir Multi-Nail Connections. U.S. Forest Products Laboratory. Report No. 2155. Madison, Wis., USA Rogerson, D.E. 1998. Corner and edge nailing test results on OSB and hemlock lumber with 2.5" common nails. MB Research Memo. MacMillan Bloedel Ltd. Vancouver, BC, Canada Sieber, D., Lam, F. and H. Prion. 1997. The Behaviour of Nailed Sheafhing-to-Frame Connections under Static and Cycle Load. Research Report. Department of Wood Science, UBC. Vancouver, BC, Canada Stone, J.L., M.D. Vanderbilt, M.E. Criswell and J. Bodig. 1980. Generalized load-slip curve for nailed connections. Struct. Res. Rept. No. 30. Civil Eng. dept., Colorado State Univ., Ft. Collins, Colo., USA 46 C H A P T E R 3. M A T P A T T E R N S I M U A L T I O N A N D P A N E L M A N U F A C T U R E C H A P T E R 3. M A T P A T T E R N S I M U L A T I O N A N D P A N E L M A N U F A C T U R E ABSTRACT Three principal processing parameters, flake orientation, flake thickness and board density, were considered in the experimental design of flakeboard structures. A Monte Carlo simulation computer program WinMat® written by Lu (1999) was used to simulate mat structure patterns and their corresponding horizontal density profiles. The results have shown that simulated flakeboard structures can be easily illustrated in 2-D horizontal density distribution (HDD) graphs, and that flakeboard structure changes due to varied processing conditions can be reflected in the graphs. The robot-based formation system was used to build flakeboard mats, which ensured exactly the same mat structures as defined in the computer program. 47 C H A P T E R 3. M A T P A T T E R N S I M U A L T I O N A N D P A N E L M A N U F A C T U R E 3.1 INTRODUCTION The physical and mechanical properties of OSB depend on the characteristics of its constituents and the manufacturing process. Special processing procedures such as flake alignment allow OSB to achieve some of its unique features. An important issue to OSB production is the control of variability of physical and mechanical properties within and between panels as this impacts both quality and production costs. Some of the causes of the within panel variability can be attributed to the distribution of voids and element contact area which are related to the OSB structure. Therefore, a better understanding of OSB structure and the variables that influence performance such as OSB to lumber nail connection properties will benefit its production and marketing in the future by choosing proper panel structural patterns to decrease variability. A two-dimensional mathematical model was developed to describe flakeboard structure features using uniform flake size and random process (Dai, 1993; Dai and Steiner, 1993, 1994a and 1994b). The non-uniformity of wood elements coverage was thought as the most considerable structural characteristic of a random mat. Different number of flakes were randomly overlapped at varied locations within a mat, which affect the final horizontal panel density distribution (HDD). Steiner and Dai (1993) summarized the spatial structure features of wood-based composites as follows: • The flake geometry has a strong impact on the relative volume in a mat. • The variation in horizontal density distribution (HDD) can be reduced by improving flake packing behavior and element alignment. 48 C H A P T E R 3. M A T P A T T E R N S I M U A L T I O N A N D P A N E L M A N U F A C T U R E • Element orientation plays an important role in optimizing directional strength properties. Dai (1993) also proposed a 2-D mathematical model based on the theory of random fibrous network in paper (Dodson, 1971) that represented the flake deposition and the area coverage with Poisson process. However, he did not discuss the real mat formation process such as partial orientation of flakes and multi-layer OSB structure. Lu et al, (1998) developed a Monte Carlo simulation program WinMat® and a mathematical model to characterize structures of wood flake mats. A robot control system was also built up, by which flake placements can be exactly controlled for experimental mats to match the simulated ones from the computer program. The simulation program linked with the robot allows verification of the theory and random generation of new database of mats with defined input processing parameters and mat structures. The program has been operational for a few of years and proved to be efficient (Wang, 1998; Lu and Lam 1999). The current research uses the computer program (WinMat®) in simulation of defined flakeboard mat structures and verifies the exactness of local density by comparing the values from simulated data with actual measurements. 3.2 INPUT DATA FOR MAT STRUCTURES Genichi Taguchi's approach (Hicks, 1993) was applied in the experimental design of robot-formed mat patterns. This experimental design examines at most four factors. Three key variables: flake thickness, flake orientation and board density, were chosen as 49 C H A P T E R 3. M A T P A T T E R N S I M U A L T I O N A N D P A N E L M A N U F A C T U R E main factors influencing flakeboard structures since they strongly impact OSB packing arrangements, and the manufacture of robot-based oriented flakeboards (Wang and Lam, 1998). The main series contained 9.5 mm thick panels. Three additional test cells were considered with panels of 11 mm thickness. Tables 3.1 and 3.2 include factors and the experiment layout, respectively. Table 3.1 Three factors and three levels Factors /levels 1 2 3 Orientation (face/core/face)* Random 0°/Random/0° 457Random/45° Board density, g/cm3 0.50 0.65 0.80 Flake thickness, mm 0.70 0.80 1.0 *Face/core/face: 25/50/25 Table 3.2 Taguchi's L 9 (34) array Mat A B C D patterns (Density, g/cm3) (Orientation, degree) (Flake thickness, mm) 1 1 (0.50) 1(R) 1 (0.70) 1 2 1 (0.50) 2 (0/R/0) 2 (0.80) 2 3 1 (0.50) 3 (45/R/45) 3 (1.00) 3 4* 2 (0.65) 1(R) 2 (0.80) 3 5* 2 (0.65) 2 (0/R/0) 3 (1.00) 1 6* 2 (0.65) 3 (45/R/45) 1 (0.70) 2 7 3 (0.80) 1(R) 3 (1.00) 2 8 3 (0.80) 2 (0/R/O) 1 (0.70) 3 9 3 (0.80) 3 (45/R/45) 2 (0.80) 1 Three other types of panels, 4a, 5a, and 6a, were made following panel 4, 5, and 6's conditions respectively except with board thickness of 11mm; Replications: 3 boards per combination, except that 4a, 5a and 6a have 2 replications 50 C H A P T E R 3. M A T P A T T E R N S I M U A L T I O N A N D P A N E L M A N U F A C T U R E 3.3 M A T S I M U L A T I O N A two-dimensional model of a wood flake mat can be considered as multi-layers of flakes in a defined area with each flake having a certain position in the horizontal plane of the mat. The flake centroid location can be defined by its x and y coordinates and flake orientation by an angle 9 (-90°< 9 < 90°), which is named as the angle between x axis and the longitudinal axis of the flake (Fig. 3.1) (Lu, 1999). Fig. 3.1 Flake location and alignment in a mat The mat simulation program WinMat® can randomly generate a database on the centroidal coordinates and flake orientation 9 of each flake in a mat to reflect the 51 C H A P T E R 3. M A T P A T T E R N S I M U A L T I O N A N D P A N E L M A N U F A C T U R E structural characteristics of flakeboards. This information can be analyzed by WinMat® to study flake coverage, free flake length, void size, and the horizontal density distribution. The UBC robot control system links the computer simulation process together with a robot to form mats with defined mat patterns. Experimental tests can be carried out to verify the simulated mat structures, such as measurements of local density values. The simulation program needs the following inputs: mat size (length, width and thickness, mm), flake dimension (length, width and thickness, mm), flake density (g/cm3), flake alignment (-90°< 0 < 90°), flake centroid locations, and final panel density (g/cm3). The calculations are carried out with the formuli as follows: Ka, =pbxLxWxT (3.1) where: Wma{. weight of a formed mat; L: mat length, equal to 240 mm; W: mat width, equal to 240 mm; T: target board thickness, mm; and pt: target board density, g/cm Wf = pf xlxwxtx Nf (3.2) where: 52 C H A P T E R 3. M A T P A T T E R N S I M U A L T I O N A N D P A N E L M A N U F A C T U R E Wf. Weight of flakes in the mat, g; pf. flake density, 0.380g/cm3; /: flake length, equal to 100 mm; w. flake width, equal to 20 mm; t: flake thickness, mm; and N/. number of flakes in the mat If we assume that flake dimension is uniform, and mat weight does not change during manufacture, then Eqs. 3.1 and 3.2 yield: N P±±LxWxT ( 3 3 ) pf X / XW X t LxW n,=- (3.4) IX w 1 , = ^ (3.5) Pf xt where: rif. number of flakes per layer, and Lm: total layers in the mat The flake overlap (O) is referred to as the number of flake layers at any local area, and it is related to the local density by the equation as follows (Lu and Lam, 1999): 0 = l-P- = RiRd (3.6) t pf 53 C H A P T E R 3. M A T P A T T E R N S I M U A L T I O N A N D P A N E L M A N U F A C T U R E where: O: flake overlaps at a certain area; p: local density at the area; Rf. thickness ratio of panel to flake; and Rd. density ratio of local panel to flake Much information can be obtained from the simulation program, such as horizontal density (overlaps) distribution, local density or overlap values, void size and distribution. Fig. 3.2 is a graph for three principal flake alignment cases. As a matter of fact, the density contour maps for each of the twelve mat patterns are different due to different combinations of target board density, target board thickness, flake thickness and flake orientation angle (Fig. 3.3). Three flake alignments are shown in Figs. 3.4 and 3.5. Random panels (1, 4 and 7) show the same 2-D graphs, but different 3-D graphs since they have varied nominal density (0.50, 0.65 and 0.80 g/cm3 respectively). The oriented panels including 0°/random/0° (2, 5 and 8) and 457random/45° (3, 6, and 9) also have the same tendencies in horizontal density profiles as random panels do. The 0°/random/0° and 45°/random/45° panels should have identical panel structures if the first ones were turned clockwise at 45°. But it is worthwhile to notice that for the two kinds of panels, local density values corresponding to the same x, y coordinates would be different since they both follow the same nailing patterns as mentioned in Chapter 2. Panels 4, 5 and 6 (9.5mm) and 4a, 5a and 6a (11mm) have correspondingly the same 2-D and 3-D graphs (0.65 g/cm ), but flake overlaps are obviously different due to their varied thickness, which should be considered in the analysis of flakeboard nailing performance. 54 C H A P T E R 3. M A T P A T T E R N S I M U A L T I O N A N D P A N E L M A N U F A C T U R E Random Panel Density Distribution 220- Wb 180 0 20 40 60 80 100 120 140 160 180 200 220 X Direction (mm) 0.90 0.80 0 70 - 0.60 • 0.50 • 0.40 • 0.30 - 0.20 0.10 Fig. 3.2 Two-dimension graphs of density distribution 55 C H A P T E R 3. M A T P A T T E R N S I M U A L T I O N A N D P A N E L M A N U F A C T U R E Random Panel Density Distribution Fig. 3.3 Three-dimension graphs of density distribution C H A P T E R 3. M A T P A T T E R N S I M U A L T I O N A N D P A N E L M A N U F A C T U R E Fig. 3.5 Robot-formed panels with three main flake alignments 57 C H A P T E R 3. M A T P A T T E R N S I M U A L T I O N A N D P A N E L M A N U F A C T U R E 3.4 PANEL MANUFACTURE 3.4.1 Manufacturing Process Aspen Logs Debarking Stranding Drying w w w ....... .p. Manually Blending (PPF) —• Robot-Based Forming —• Pressing Trimmed to 240x240 mm 3.4.2 Materials Aspen is a mostly used species in the OSB manufacture. In this research, aspen was also selected. Its average density was 0.380 g/cm based on the oven-dry wood weight. Aspen logs with diameters of 120-150 mm were cut into about 100 mm long sections along the longitudinal direction and sent to C A E for flaking. The chosen logs were flaked into rectangular flakes of 100x20 mm with three different classes of nominal thicknesses: 0.70, 0.80 and 1.00 mm. After a couple of weeks' air-drying, flake moisture content was lowered to 8-9%. The measured average thicknesses (sampling size: 200) of three types of flakes are: 0.714, 0.794 and 0.966 mm, respectively. The flakes with uniform color and thickness were selected as experimental materials. Powder phenol formaldehyde (PF), CASCOPHEN® W735A, was chosen as glue in the experiments. The resin content was 5%, based on oven-dry wood weight. Air dried flakes and powder PF resin were manually blended in a sealed plastic bag. To ensure a uniform 58 C H A P T E R 3. M A T P A T T E R N S I M U A L T I O N A N D P A N E L M A N U F A C T U R E resin distribution and reduced flake damage, shaking force and frequency were carefully controlled for every plastic bag. 3.4.3 Robot-Based Mat Formation The robot-based mat formation system consists of three main subsystems, PC computer, robot controller, and robot mat former (Fig. 3.6), which links the computer simulation and experimental mats together. The centroid location and orientation of each flake can be automatically procured by simulation and recorded in a file. The robot-based system gains access to the file and the information can be transferred to the robot controller in the form of robot commands. One by one, flakes can be put in a defined position with a certain orientation using the robot arm in order to create a mat with a predefined structure. The whole system can substantiate the simulation procedures and build up a test database for every mat with high repeatability (Wang and Lam, 1998; L u , 1999). Aspen flakes with same dimensions and qualities were used for both face and core layers, and the ratio of face to core was 50:50. Computer Robot controller Robot mat former Fig. 3.6 Robot forming system 59 C H A P T E R 3. M A T P A T T E R N S I M U A L T I O N A N D P A N E L M A N U F A C T U R E 3.4.4 Panel Pressing 3 3 3 Target board density: 0.50 g/cm , 0.65 g/cm and 0.85 g/cm , on the basis of the calculation for simulation program WinMat with mat dimension 240x240x9.5 (or 11) mm, actual board thickness and actual board density (based on sample weight and volume at about 6% m.c.) values are listed in Table 3.3. Table 3.3 Comparisons of target and actual board density Mat patterns Target board density (g/cm ) Average board density (g/cm3) 1 0.500 0.522 2 0.500 0.538 3 0.500 0.529 4 0.650 0.642 5 0.650 . 0.648 6 0.650 0.655 7 0.800 0.799 8 0.800 0.832 9 0.800 0.826 4a 0.650 0.654 5a 0.650 0.658 6a 0.650 0.661 Pressing conditions were as follows: • Press machine: computer-controlled 300 x 300 mm hot press • Press pressure: 2.60 MPa • Pressure ramp rate: 4.14 MPa/min. 60 C H A P T E R 3. M A T P A T T E R N S I M U A L T I O N A N D P A N E L M A N U F A C T U R E • Press temperature: 185 °C • Press time: 5 min for 0.50 g/cm boards, 5.5 min for 0.65 g/cm boards, 6.5 min for 0.80 g/cm boards 3.5 RESULTS AND DISCUSSION Using the simulation program WinMat®, a database can be automatically built up for every defined mat structure. The robot-formed system makes it possible to transfer predefined and simulated image mats into real mat structures. Very good match can be found between simulation (target) and actual robot-formed panels. In general, if inputs for the computer simulation, such as flake density, flake dimension, are representative of the flakes used in the mat, the panels with repeatable structures can be built up as defined in the simulation, which has been verified by the results as shown in Table 3.3, and also by others (Wang and Lam, 1998; Lu and Lam, 1999). 3.6 CONCLUSIONS The defined flakeboard structure patterns can be simulated accurately using the computer simulation program WinMat®. The flakes can be deposited into a mat by converting the simulation data file into robot commands, and then activating the robot arm to position flakes one by one in a predefined location and orientation. The average board density results confirm that the simulation program and the robot-based forming system are both working well to obtain the expected board structures. 61 C H A P T E R 3. M A T P A T T E R N S I M U A L T I O N A N D P A N E L M A N U F A C T U R E 3.7 REFERENCES Dai, C. 1993. Modeling structure and processing characteristics of a randomly formed wood flake composite mat. Ph.D. Thesis. Department of Wood Science, UBC, Vancouver, BC, Canada Dai, C. and P.R. Steiner. 1993. Compression behavior of randomly formed wood flake mats. Wood Fiber Sci., 25(4): 349-358 Dai, C. and P.R. Steiner. 1994a. Spatial structure of wood composites in relation to processing and performance characteristics. Part 2. Modeling and simulation of a randomly formed flake layer network. Wood Sci. and Technol., 28 (2): 135-146 Dai, C. and P.R. Steiner. 1994b. Spatial structure of wood composites in relation to processing and performance characteristics. Part 3. Modeling the formation of multi-layered random flake mats. Wood Sci. and Technol, 28 (3): 229-239 Dodson, C.T.J. 1971. Spatial variability and the theory of sampling in random fibrous networks. J. Roy. Statist. Soc. B. 33(1): 88-94 Hicks, C R . 1993. Fundamental concepts in the design of experiments. Saunders College Publishing, New York, USA Lu, C ; P.R. Steiner; and F. Lam. 1998. Simulation study of wood-flake composite mat structures. Forest Prod. J., 48(5): 89-93 Lu, C. 1999. Organization of wood elements in partially oriented flakeboard mats. Ph.D. Thesis. Department of Wood Science, UBC, Vancouver, BC, Canada Lu, C. and F. Lam. 1999. Study on the x-ray calibration and overlap measurements in robot formed flakeboard mats. Wood Sci. and Technol., 33(2): 85-95 62 C H A P T E R 3. M A T P A T T E R N S I M U A L T I O N A N D P A N E L M A N U F A C T U R E Steiner, P.R. and C. Dai. 1993. Spatial structure of wood composites in relation to processing and performance characteristics. Part 1. Rational for model development. Wood Sci. and Technol., 28 (1): 45-51 Wang, K. and F. Lam. 1998. Robot-based research on three-layer oriented flakeboard. Wood Fiber Sci., 30(4): 339-347 63 C H A P T E R 4. P H A S E II. N A I L C O N N E C T I O N T E S T S O N R O B O T - B A S E D F O R M E D P A N E L S C H A P T E R 4. P H A S E I L N A I L C O N N E C T I O N T E S T S O N R O B O T - F O R M E D P A N E L S ABSTRACT Predefined and laboratory-manufactured oriented and random flakeboards were assembled with 38x89 mm SPF lumber into nail connections. Single nail lateral resistance tests were conducted to study the effects of failure modes, panel types and loading directions on nail-connection properties. The results showed that: 1) most nail properties for the specimens that failed in the pull-out mode were significantly different from those in the pull-through mode; 2) the specimens that failed in the pull-out mode had higher initial stiffness and connection strength (maximum, yield and ultimate loads) than those in the pull-through mode; 3) compared to OSB panels, random panels had higher connection strength for the pull-through mode, larger maximum displacement for the pull-out mode, and higher maximum strain energy and ultimate strain energy, and larger ultimate displacement for both failure modes; 4) the 90° loading direction in OSB panels indicated significantly different nail properties for both pull-out and pull-through modes compared with the 0° and 45° loading directions, but there were no significant differences in nail properties between 0° and 45° loading directions under the pull-through mode; 5) there was significant differences in connection strength between 0° and 45° loading directions under the 64 C H A P T E R 4. P H A S E II. N A I L C O N N E C T I O N T E S T S O N R O B O T - B A S E D F O R M E D P A N E L S pull-out mode; 6) from the regression analysis, most of the OSB or random flakeboard to SPF lumber nail connection properties were impacted by different combinations of panel local density (LD), board to flake thickness ratio (77?), and lumber specific gravity (G); 7) a parametric study was carried out to show the potential application of the information developed in this paper; generally, higher lumber specific gravity and panel local density mostly showed better initial stiffness and connection strength (loads) within the regression ranges and fixed lumber and panel properties. The effect of panel to flake thickness ratio is comparatively complex. Different types of connection or loading conditions may produce opposite changing tendencies. Hankinson's equation pretty well predicts initial stiffness and maximum load to measured values at 45° loading angle on the basis of nail properties along and across OSB face flake alignment, and may have good predictions on nail properties at any loading direction, which should be verified in the further study. 65 C H A P T E R 4. P H A S E II. N A I L C O N N E C T I O N T E S T S O N R O B O T - B A S E D F O R M E D P A N E L S 4.1 INTRODUCTION Despite increased sophistication in construction techniques and greater need for engineered wood assemblies, common nailed joints still remain of paramount importance in wood structural applications. Unlike other fastening methods, a nailed joint relies on a unique interaction between the fastener and the jointed material(s) for its strength and stiffness (McLain, 1975). The nailed joint has always been, and most likely will continue to be one of the significant factors for safe and economical wood construction. In addition, the development of new wood composites and their new applications in wood structures motivates research on and design of nailed joints for these products to be further refined for higher efficiency. To meet this goal, a closer look is needed at the complex interaction between the fasteners and the properties of the wood-based members. Even though there are many studies on the performance of flakeboard nail connections (Mohammad and Smith, 1994, 1996; Sieber et al, 1997; He, 1997; Rogerson, 1998) and about the impact of Oriented Strand Board (OSB) structures on some of the panel's physical and mechanical properties (Suchsland and Xu, 1989; Kamke and Wolcott, 1991; Dai and Steiner, 1994a, b, c; Lu and Lam, 1998 and 1999; Wang and Lam, 1998, 1999), the effect of flakeboard structures on the lateral resistance of OSB or random flakeboard to lumber nail connection has not been considered. Now that the structure of OSB was verified to influence density distribution and some physical and mechanical properties of the panels (Lu and Lam, 1998, 1999; Wang and 66 C H A P T E R 4. P H A S E II. N A I L C O N N E C T I O N T E S T S O N R O B O T - B A S E D F O R M E D P A N E L S Lam, 1998, 1999), it is expected that the structure will also affect nail connection performance. The lateral resistance of OSB or random flakeboard to lumber nail connections is controlled by the properties of the framing lumber (main member), OSB or random flakeboard sheathing material (side member), and fasteners (common nails). In this study, it is assumed that the nails used in the tests have uniform shapes and mechanical properties. Therefore, only panel types (random or oriented flakeboards), panel local density values (LD) around nailing positions, board to flake thickness ratio (TR), loading directions in OSB panels, lumber specific gravity (G), and failure modes (FM) were examined. Based on the test results, relationships of nail properties to panel local density (LD), panel to flake thickness ratio (77?) and lumber specific gravity (G) have been evaluated under different conditions. The objectives of the study are: • To determine the lateral resistance of nail connections with different structural panels using a new test jig to apply lateral forces along the direction parallel to the longitudinal axis of the framing member (parallel to grain direction of the lumber). • To compare nail connection properties between different panel types (random and oriented flakebaords), loading directions (0°, 45° and 90° angles relative to the face flake alignment) in OSB panels, and failure modes (pull-through and pull-out) using T-tests. • To find the relationships of nail connection properties with OSB local density (LD), board to flake thickness ratio (TR), and lumber specific gravity (G) in terms of panel 67 C H A P T E R 4. P H A S E II. N A I L C O N N E C T I O N T E S T S O N R O B O T - B A S E D F O R M E D P A N E L S types: robot-formated small sized random or oriented flakeboard, and under different loading directions in OSB panels by regression analysis. 4.2 M A T E R I A L S A N D M E T H O D O L O G Y 4.2.1 Materials Flakeboard: Laboratory-based random flakeboards and oriented flakeboards (OSB), 240x240x9.5 mm and 240x240x11.0 mm, nominal board densities: 0.50, 0.65 and 0.85 g/cm , respectively. Twelve different mat structures of flakeboards with different combinations of flake thickness, panel density, flake orientation and panel thickness were simulated using WinMat® program. The centroid location and orientation of each flake can be automatically procured by simulation and recorded in a file. The robot-based system then gains access to the file and transfer the information to the robot controller in the form of robot commands. One by one, flakes can be put in a defined position with a certain orientation using the robot arm to create a mat with a predefined structure. Aspen flakes with uniform dimensions and qualities were used for both face and core layers of the mat. The ratio of face to core was 50:50. The flakes, stranded from aspen logs with diameters of 12-15 cm, had 100x20 mm sections, and three different nominal thicknesses: 0.70, 0.80 and 1.00 mm, correspondingly. The measured average thicknesses of these three types of flakes were: 0.714, 0.794 and 0.966 mm, respectively. 68 C H A P T E R 4. P H A S E II. N A I L C O N N E C T I O N T E S T S O N R O B O T - B A S E D F O R M E D P A N E L S Powder phenol formaldehyde (PPF), CASCOPHEN® W735A, was chosen as glue in the main experiments. The resin content was around 5%, based on oven-dry wood weight. Air dried flakes and power PF resin were manually blended in a sealed plastic bag prior to robot forming. Robot-formed mats with predefined structures were finally hot-pressed into panels with required densities. Pressing conditions were as follows: • Press machine: computer-controlled 300 x 300 mm hot press • Press pressure: 2.60 MPa • Pressure ramp rate: 4.14 MPa/min. • Press temperature: 185 °C • Press time: 5 min for 0.50 g/cm boards, 5.5 min for 0.65 g/cm boards, 6.5 min for 0.80 g/cm3 boards Lumber: SPF, kiln-dried lumber, 38x89x550 mm, uniform, no defects and knots around nailing areas, specific gravity ranges: 0.35-0.50, average moisture content: 11-12%. Nails: Common nails, 6d TREE ISLAND magnetic pack nails (ICBO no. 4266), nail length = 50 mm, diameter = 2.87 mm, and head diameter = 5.74 mm. Lumber and robot-formed panels were conditioned at 50% R.H. and 20°C for more than two weeks before the lumber/flakeboard joint assembly and nail property testing. 69 C H A P T E R 4. P H A S E II. N A I L C O N N E C T I O N T E S T S O N R O B O T - B A S E D F O R M E D P A N E L S 4.2.2 Methodology 4.2.2.1 Nailing Pattern and Experimental Procedures The adopted nailing pattern is shown in Fig. 4.1. The lumber located at the end of the connection was fixed to the M T S 810 universal testing machine table, while O S B panels situated at the upper end of the connection were connected to the M T S loading unit. A unidirectional tensile load was applied to the specimen to cause a shear deformation. The loading rate was 2.54 mm/min for all tests. The displacements between panel and lumber were measured by a Linear Voltage Displacement Transducer ( L V D T ) mounted on the connection. Fig. 4.1 Nailing pattern for laboratory-based panels 70 C H A P T E R 4. P H A S E II. N A I L C O N N E C T I O N T E S T S O N R O B O T - B A S E D F O R M E D P A N E L S The nail properties consist of all characteristics as defined below: K: initial slope or stiffness of load-displacement curve, N/mm Pmax' maximum load, N AmaX' displacement corresponding to maximum load, mm Umax- strain energy at maximum displacement, the integration of the load-displacement curve from zero displacement to Amax, N.mm Pyia-: yield load, corresponding to 50% maximum load, N Ayui: displacement corresponding to yield load, mm Uyui: strain energy at yield displacement, the integration of the load-displacement curve from zero displacement to Ayid, N.mm P„i,: ultimate load, corresponding to the post peak load at 80% of maximum load, N Auit: displacement corresponding to ultimate load, mm Uui,: strain energy at ultimate displacement, the integration of the load-displacement curve from zero displacement to Auu, N.mm Dp ductility factor 1 = Aut,/Amax D2: ductility factor 2 = Amax/Ayid 71 C H A P T E R 4. P H A S E II. N A I L C O N N E C T I O N T E S T S O N R O B O T - B A S E D F O R M E D P A N E L S Some additional definitions in this section include: Loading directions (relative to the face flake orientation): R: loading relative to random flakeboards 0°: loading parallel to the face flake orientation in OSB panels 45°: loading at 45 or 135 angle relative to the face flake alignment in OSB panels 90 : loading perpendicular to the face flake alignment in OSB panels LD: panel local density around a nailing location (15 mm x 15 mm area), from simulation data using the W i n M a t program TR: board to flake thickness ratio G: lumber specific gravity FM: failure modes of a panel (random flakeboard or OSB)-lumber joint, either pull-through or pull-out. Pull-out failure means that the connection failed because of significant bending of the nail with the nail eventually pulling out of the lumber. Typically, the nail head remained on or above the top surface of the panel member and the panel nailing area had neither damage nor other distinctive changes. Pull-through failure means the connection failed because of the nail head intruded significantly into the panel resulting in the loss of load-carrying capacity in the connection. Usually a certain degree of the nail withdrawal or pull-out accompanied the ultimate pull-through failure mode, but for simplicity these contributions were generally ignored if they were not obvious enough to be differentiated from two main failure modes. 72 C H A P T E R 4. P H A S E II. N A I L C O N N E C T I O N T E S T S O N R O B O T - B A S E D F O R M E D P A N E L S 4.2.2.2 Panel Local Density Used for Regression Analysis Shown in Table 4.1 are panel local density information measured around various nailing positions (15x15 mm) for three types of robot-formed mats, and that simulated using the WinMat® program around the same nailing locations (15x15 mm) for the three mats. This process is possible because the mats structures were known a priori through the robot formation system. Comparisons of the simulated and measured local density values indicate a reasonably good agreement; therefore, simulated local density values will be used in the subsequent analyses (Table 4.1 and Fig. 4.2). 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 Simulated local density, g/cm3 Fig. 4.2 Simulated and measured panel local density comparisons These simulated local density values were applied in the regression analysis as one of the three selected factors. By doing this, the real density measurements around nailing positions were not needed, as these measurements would be especially difficult to obtain accurately for the pull-through failure areas due to panel damage and loss of material. 73 s-o ix IN w os vq ro ro' Ov vo vo o ro CN VO CN ho 6 1«g J W Z < a W S DS o o W < pa i H o aa O Ci Z o H (/> fcd H Z o f-u w z z o u z W < X OS w H < X a I N <U + N - * N CCS a •** a Vi C U SB a cu T3 cu o C cu s a * cu T3 S ecs Ci o ' + N -2 £ t/5 i—i Tf CU CCS H cu e CCS cu cu fi o •*•> CCS fl o • P N .2 "s E 153 c« fl cu -a O s_ o s-i. IW fl o • P N -pJ -2 "3 £ I K fl cu O > <_ u o u s~ -fi o • P N "3 s a cu "5* fl cu T3 a cu "Sb fl cu -a a > < cu "5b Vi fl cu -a cu ~6JD £ cu "5b C/3 fl CU o CN CN O O cu "5b ox S fl o CCS Vi Z 2 CN O ov Ov o in CN Si Ov oo vo o o o VO o C--CN r--CN CN o CN r--oo in ro VO IT) VO C M o o m ON VO o VO m Ov IVO OV Ov VO VO < lis vo 00 00 m 00 < vo in in I oo vo ro oo OV O oo m o Ov vb VO vo o l ro 00 VO © 00 vo Si Ov 00 CO | oo © in CN o CN oo Ov vo o vb ro o vo Ov m t--Ov Ov O O I M vo © vo vo © cu o I > o c I-a < ,Z Ov in < IT) B o o oo o XI VO 3 E S £ D 3 — ,C3 "3 S S -2 BX ° "3 Z a. C H A P T E R 4. P H A S E II. N A I L C O N N E C T I O N T E S T S O N R O B O T - B A S E D F O R M E D P A N E L S 4.2.2.3 Statistical Analysis of Testing Results By Tauguchi's approach to experimental design, we can reduce the number of experiments and gain maximum information. Twelve different mat patterns were obtained from twelve combinations of three main processing factors: flake thickness, board density, and face flake orientation. Using the special nailing formats (Fig. 4.1) and the test set-up discussed in Chapter 2, tensile loads were applied along panel sides and diagonals, thus the influence of loading directions relative to the flake alignment (OSB) on the nail-connection performance could be studied. Each panel was assembled and tested with sixteen individual nail tests. Each time one nail was fastened and then tested, and a load-slip curve (a data set) was recorded automatically by a data acquisition program. The original load-slip data were smoothed using Binomial Filter Process (Lu, 1999) to remove some noise and acquire a relatively smooth curve. From the smoothed curve, all nail-connection properties including initial stiffness, loads (connection strength), displacements, strain energies and ductility factors were obtained. The entire data set was then sorted by different factors, and analyzed using different approaches to evaluate property differences under varied conditions, and finally to build up the relationships of nail-connection properties to the main member (SPF lumber) and the side member (flakeboards) characteristics. 4.2.2.3.1 T-tests T-tests are often applied when two or more varied experimental conditions need to be compared. If T-tests indicated that the properties under different treatments were significantly different, it means that the conditions would distinctly affect the test results. 75 C H A P T E R 4. P H A S E II. N A I L C O N N E C T I O N T E S T S O N R O B O T - B A S E D F O R M E D P A N E L S T-tests were conducted on the mean values of the various nail-connection properties under varied testing conditions. 4.2.2.3.2 Multivariate Regressions T-tests can only identify a property difference between two or more varied conditions, but cannot define the reasons of the difference. A multivariate regression approach was therefore applied to achieve this goal. After trying a few different models, including linear, exponential, power and log-linear models by plotting scatter-grams, a quadratic regression model was judged to best represent the data and was selected. Considering all the main factors influencing the flakeboard to lumber connection properties, five variables were chosen in the regression analysis which include lumber specific gravity (G), panel local density (LD), board to flake thickness ratio (flake overlaps, 77?), flakeboard types (random or oriented), and loading directions in OSB panels. To further simplify the regression analyses, the data were sorted into different groups according to flakeboard types and loading directions. The quadratic model for this part of the study takes the following form (Rawlings, 1988; Hicks, 1993): Y = p0 + p,xx + p2x2 + p,x3 + pX + p5x22 + fit*! + frXJi + fi,X,X3 + &x2x3 (4-1) where: Y: connection property Xf. panel local density (LD) X2. board to flake thickness ratio (77?) Xf. lumber specific gravity (G) Pf. regression coefficients (i = 0, 1,2, . . . , 9) 76 C H A P T E R 4. P H A S E 11. N A I L C O N N E C T I O N T E S T S O N R O B O T - B A S E D F O R M E D P A N E L S The SAS® program ( B A C K W A R D method) was used to analyze the data to obtain all the regression models (SAS, 1996). 4.3 RESULTS AND DISCUSSION 4.3.1 Results A total of 528 nail connections were tested with oriented or random flakeboards. The results are summarized in Tables 4.2-4.7, in which some statistics, such as average, standard deviation, coefficient of variation, maximum value, minimum value, and number of tests, are summarized according to the categories of failure modes, panel types (random and oriented flakeboards), and loading directions relative to the face flake alignment in OSB panels. Based on the original data and T-tests, each nail property under different conditions was compared. By means of regression analysis on the original data, the effects of lumber property (specific gravity) and sheathing material property (random and oriented flakeboards: local density and thickness ratio) were investigated. 4.3.2 Discussion 4.3.2.1 T-tests on Means of Nail-Connection Properties T-tests were carried out to examine the differences in nail-connection properties between the two main failure modes (pull-through and pull-out), between the random and oriented flakeboards (OSB), and between the loading directions in OSB panels. The results are shown in Tables 4.8-4.11. 77 C H A P T E R 4. P H A S E II. N A I L C O N N E C T I O N T E S T S O N R O B O T - B A S E D F O R M E D P A N E L S 4.3.2.1.1 Failure Modes Two main failure modes were observed in the nailed connections. The relative portion of the two failure modes varied with board types and with loading directions in OSB panels (Table 4.8). In general, the percentage of pull-through mode ranges from 23% to 31% of the total failures. Random panels had slightly more pull-through failures than OSB panels. Amongst the three loading directions in OSB panels, the 45° loading direction had the least percentage of pull-through failures and correspondingly the highest percentage of pull-out failures. The shapes of nails after failure were clearly different between the two failure modes: "S" or "Z," shape for the pull-out mode and relative straight shapes for the pull-through mode, as shown in Fig. 4.3. Fig. 4.3 Nail shapes in pull-through and pull-out failure modes For the whole data set including all test specimens, all nailing performance except maximum strain energy and ultimate strain energy was significantly different for the two 78 C H A P T E R 4. P H A S E II. N A I L C O N N E C T I O N T E S T S O N R O B O T - B A S E D F O R M E D P A N E L S failure modes. Among them, initial stiffness, connection strengths (loads) and ductility factor two (D2) under the pull-out failure mode were obviously higher than those under the pull-through mode. On the contrary, displacements, yield strain energy and ductility factor one (Df) under the pull-through mode were significantly higher than those under the pull-out mode (Tables 4.2 and 4.9). For oriented panels, all properties except maximum strain energy (Umax) and ultimate strain energy (Uuil) were significantly different between the two failure modes. Initial stiffness, connection strength and D2 under the pull-out mode were obviously higher than those under the pull-through mode. Other properties under the pull-out mode were obviously smaller than those under the pull-through mode (Tables 4.3 and 4.9). For random panels, most of the connection properties were statistically different between the two failure modes. Initial stiffness and strength under the pull-out mode were distinctly higher than those under the pull-through mode. However, displacements and yield strain energy under the pull-through mode were obviously higher than those under the pull-out mode, as shown in Tables 4.4 and 4.9. Loading directions in oriented panels also caused significantly different nail connection properties between the two failure modes. When a joint was loaded along the OSB face flake orientation (0° loading angle), initial stiffness, strength, yield and ultimate displacements, and D2 were significantly different under both failure modes. Among them, initial stiffness, strength, and D2 under the pull-out mode were higher than those under the pull-through mode. Other properties were comparatively smaller under the pull-out mode (Tables 4.5 and 4.9). 79 ii ii a, o u a c o u <u C c o u fi fi o (*> <u T3 O E a is s 03 s O 5 .a cn oo VD OV Ov cn CN VO OS in Tt o O Tt cn O d d o o VO so VO r-; in vo o 00 cn VO OS as vo CO —• oo vo CN —- CN d o O O o CN o Tt in Tt o Tt so Tt in vo o vo cn o CN 00 CN O Tt 00 in m in o — o o VO Tl H W 1/3 < < O UJ -J O X l<c s 5 IS! ON VO CN VO Ov O Tt 'A, _ t r~- . in _ o — ov „ Tt o > H 00 CN OS r— oo CN o in in CN — CN —[ in o OS *° VO o O oo cn — O OS — VO OS m Tt o VO CN O - J N •* "I n — so o cn X oo CN z tt CN cn OS OS cn cn OS in Tt o o Tt cn o o d o o — O — o cn cn cn Tt so <n 00 in — CN p CN Tt d o d — d oo Tf in CN vo VO r- o cn -~ Tt 00 in Tt vd O Tt cn 00 Tt cn CN o in OS 00 Tt in O Tt o d vb o vo Tt in scs Tt r- in o m in CN Tt 00 ON Tfr — r— d cn vo 00 Tt o — 0O — Tt O Tt — CN — VO O OS _ OS S CN — — CN — in d O £ <N O — — Tt °i oo cn Tt — O 00 — — VO Tt cn cn r- — Tt o Tt r- — — CN Tt Tt Tt m S cn N - & vo O -O CO . t VO Tt — CN O — CN — > W G H t/3 VO d > O U SO CN X < S z tt cn r- cn OS Tt cn CN vo Os t--Tt O o Tt cn 00 d d d d d cn VO r- vo f-; _J os in SO 00 d 00 cn o OS CN OS CN in CN r- '—' Tt t— r- —- '— Tt 00 d d d —' d cn CN cn o CN Tt Tt Tt CN in r- Tt O 00 —' vd d SO Tt cn vo 00 in wn Tf 1 — Tt CN Os O 00 —' d d cn — cn vo Tt OS _- O — cn S ^ t— ^ m — o — Tt d Tt OS so cn 00 Tt o cn ~ cn d m vo —! oo O OS „ oo 2 nv P Os O Tt d CN — CN SO OS ^- — o — SO vo oo X < r-oo cn O cn — z fi o u cu fi fi o u S-cu -O ffl g u « o .o cu « 5= X I cu •pH fi cu • PP p. o C H O CU a o u a & o cu fi fi o cu '« z — a H < o vO 60 5 £ "N 5 s s IS -N s cn cn Tt ON <N O NO NO o ON ON Tt ON cn CN in o O o o o cn CN oo 00 r-; in NO CN in — CN o — 00 c  o r-00 NO o Tt CN ON 00 CN CN in o o o — O c4.57 in ON t— t— Tt 1.40 o in CN in — NO o NO cTt 00 O NO NO CN cn o in Tt o CN in o © NO cTt ON — in oo Tt - § 2 ON CN ON „ NO o NO o Tt Tt Tt CN CN 00 r- oo in o o — CN vo — ^ s Tt 00 CN 2 Tt cn Tt © ' > fad Q H NO . CN Tt . in CN — cn 2 P © . ON NO O — CN FS 3 £ 8 s s v5 o CN Tt O N O N cn cn r- O N in Tt o o Tt cn o © o o © O N N O ON O N 1-; N O © —' in — CN o 00 O N CN cn ci N O o •n 00 in »—' cn © CN o o o © ' O N Tt «n in r- O N r- o CN cn Tt oo in — N O o cn O N N O r- Tt o in O N 00 Tt "n © © o N O O N o — r-oo Tt o N O Tt ON CN NO Tt o cn oo ON — cn l ON NO _ cn ON oo _ Tt o o ON 00 ON CN NO o Tt © 00 ON CN Tt © Tt — ON r- Tt > ° Q > > H O < cfl U 00 o CN o in CN CN —; „ in o ON o cn — — ON NO O r- NO — — CN ON NO CN Tt — NO ON X Z < K S S z in r-- cn ON ON r^  CN NO ON r-Tt O o Tt cn o o o O o o o oo CN CN in NO CN o — 00 ON ON ON ON NO in CN NO ON r- Tt o O o O . in o CN vo Tt CN in o Tt O o NO Tt o NO ^« Tt 00 Tt CN ON o o O cn ND CN N O CN cn Tt NO J_ cn CN — Tt Tt ~ — r-~ NO —« in o Tt in CN Tt — Tt ON <r, Jj! CN s ; cn oo 1 . m — o — Tt o N O o ON CN NO 00 NO ON NO NO CN NO ON — N O Jj _- — o — ON NO > faJ Q H Tt o NO CN NO CN NO CN CN — ON NO r- CN in — — NO — CN > 2 z 8 § s u l fi o '& u a fi fi o w u a XI S a ii fi cs a S © fi es I . © VI ii © U a-fi _ © <u fl fl o a CS H tt 1 -1 o s IS 8 5 s s IS xj § 1*7 s i i CO in Tt Tt co CN vo oo e'-Tt o o Tt en d d d d d cn o CN — CN O N — — d — 00 r-VO in in CN CN CN cn o o Tt NO r-d d d — d 14.96 5.98 0.399 41.55 4.66 VO ON in in i— o cn o Tt NO O vo I— d d O N O N cn cn r-O N vo r- t^ 00 "~™ vO o cn CN cn d t -vo . cn _ o — Tt d Tt CO Tt vo vo cn Tt oo g — vo r- ^  C N r-— o — cn — oo C N <N £ -O . NO — O oo 00 Tt t-~ C N Tt Tt VO Tt — vo cn O CN — — VO Tt O O C N VO > W Q H i CN vn N O . r- Tt t-O — >n — O N r- vo — O N C -cn O N — > o u r c tt C r « -| O c i. c cV : =5 -5 s £ ; S? ! mm e X? X ft o <c p o s BS I TH -1 sT CM s s T ? < ? , N.mi g s t> S | 5 5 Tt Tt cn 00 o o Tt cn d d d d cn r- CN O N o CN O N d d 00 r-co N O O N o o — vo Tt d d d d in o o r-N O oo Tt d N O in d cn Tt in t-~ o o O Tt Tt CN d d vd o r— N O d O Tt O N Tt oo o ov in Tt N O O N cn „ Tt t 1 O N . C N — o ~ r— C N C N Tt — Tt Tt CN ° i OV — o t--r-. — ^ d d CN UJ O H t/3 VO cn Tt r— Tt Tt Tt Tt Tt CN vo O N Ov > 2 z S S S — — o — 00 in CN CN cn CN r- cn Tt — — o Tt d d d — d Tt r- in CN Tt Ov in NO in o cn Tt vd d Tt cn Tt CN Tt CN VO Tt CN O ^ d d cn in O N in VO r- ON CN ON ON NO in Tt OO CN vo o cn VO r» ON cn o Tt o cn 00 Tt in d CN NO CN vo cn 00 r-in ON cn ' I cn ON av o in CN O d Tt d O N OO Tt — in C N — vo (N Tt r-Tt r-O N Tt Tt 00 d « cn ^ ~~. N O o oo cn C N Tt in oo cn O N — in > u a oo Tt vq cn — ("vi o — vo C N vo CN ° cn ^ - cn CN O — ON — o — VO CN N O O vo O N CN cn — «=> s Tt — m N O Tt CN r- — — o CN O cn — — 3 Z o M u S S z CN OO fl © fl CU • PN u o cu « C cu cu c« OX) fl o "« w> .fl •3 _o fl cu -e 5= <z> cu • PN -PN CU a o PH a fl _o v> cu cu fl fl o cu z IT; T t 2 c. H -5 CN ON 00 ON ON CN NO ON in Tt O o Tt ro 00 O o o o o 00 ON 00 CN in vo 00 CN o oo oo ON o ON Tt NO oo NO CN 00 NO CN O CN oo © o o © ' 00 ON NO o in in OO Tt CN Tt' in in O oo 00 o NO Tt oo o r- NO o o oo NO ro in —* oo —• o o vb — 00 5 5 K 5 s IS - N § 5 I o o Tt — ro in © o r-o Tt O CN —< CN NO — CN O 00 — in 00 ON ON — O — ON 00 Tt r-Tt in © — — oo 00 Tt Tt o — <N. vo ro oo — o r- — oo Tt o Tt © C N © ro © — r-CN CN CN ro oo CN CN CN O in NO — CN ro ^ > UJ Q > o u ON — 0O 00 c i vo oo < 5 NO vo ro ON ON ro ro 00 ON in Tt © o Tt ro Tt © o o o © CN NO Tt r-; in CN in NO Tt — CN O —• od CN ro CN o r-r— Tt o oo 00 Tt — ro vo CN Tt © O © ' © o CN ro NO Tt in r- ON t~- 00 in ro ro Tt Tt — Tt © CN Tt CN CN © NO ro O o o O in in Tt CN o vb CN 00 ON OO NO <N O Tt © ro 8 <*! — O 00 o P 2 ^ C N © C N o o > hi Q H ON „ o o 2, NO oo Tt m . © o CN O — CN °v Tt m o — — Tt NO —• O ON — —< ~ —! © i i n © ro O NO ro vo — NO o — X Tt CN ro CN . CN — m ^ ° ro NO O — i CN 00 NO o < « u S S z ro CN in ON ro CN VO ON r-Tt o o Tt ro O © o O O ON 00 r- r--in NO CN o 00 C N — r~ Tt ro r-~ r-~ — — Tt CN o Tt r- o o ON Tt CN in Tt in o Tt — ON O NO Tt CN o ro NO in ON NO 00 Tt CN NO Tt —' o o ro CN vo Tt vo oo ON ON 00 r— o o o — o — NO o r-in o > hi Q H c / i ON o ON Tt CN Tt C N NO ON NO r-. —. — O ro O m ON NO vo o n o< o — o in CN Tt CN Tt CN Tt CN Tt CN Tt £ 00 NO m " CN — CN . m _ _ o — o ro ^ S ON m ^ — — Tt O ro — CN > % z o a u S S z c o c © CU u CU cu s-W) cu •O <r, + H « ex -d « o fl a* -C si s-cu a o s-a & %-« cu cu fl fl o eu Z VO CU « 5 < o 5 S 5 "N in ro Tt 00 tN o in N O O 00 r-Tt Tt r-ro N O r-d d d d d 11.7 tN tN 0.188 15.7 00 V O r-tN tN ro t--ON NO ro in oo NO CN p tN d d d — d o oo in ON ro in r- o Tt oo Tt 00 in NO d Tt ro o ro ON tN >n in 00 ro NO p d d •n N O r-N O r-in o ro o S r- ro O N r- is ro Tt Tt O N ~ — 00 — in o Tt — N O ~ Tt S 00 S r-oo °! m _ o — r- N O oo tN — O V O °: in CN r-O CN CM — N O d o tN 00 d o — tN r- °. N O o N O — ri oo oo r-— O O N — — ro ^z, Tt ro CN O N ro N O CN' £ CN — — S r- O N N O . O N in r~~ O — ro Tt N O > ° O > > H O < v, u S S z o IN Tt in Tt ro ro r- r-~ Tt O o Tt ro o d d d d d Tt N O N O 00 d o in N O o CN d od Tt CN 00 ro r-CM N O tN in oo in ro o tN d d d d O N o in oo O tN o tN o N O od in O 00 d Tt ro Tt ro tN ON Tt in m O 00 Tt NO p tN d d in tN 00 — NO O 00 Tt - s -Tt m . tN CN O — CM CM ON O CN CM O N O O N O N O Tt Tt O Tt N O O N O Tt > a H 00 Tt r-o Tt Tt 00 o — Tt oo CM N O in tN IN CM CM d in o O N 00 r~ — o Tt o Tt ^ Od ^ o — — o Tt O N O N " m in o O — ro Tt > % z o < S u S S z N O ro Tt CM o tN N O o oo r— Tt oo ro N O ro d d d d d 12.1 CM CM 0.179 15.7 N O 00 N O ro O N O N 00 ro N O Tt in in O N — o Tt d d d d O N o in Tt in — oo ro N O ro in ro Tt O N ro N O d ro Tt ' 1 ro 00 in m O N oo Tt CN o p — d d C O , ro CM r- Tt I ~ ro CN m m . — t-— in o Tt in O N N O CM '11 r- Tt N O ro . in ro ro O N — O — N O — N O Tt N O Tt N O d d in N O 00 ro N O — N O N O N O CM ' ! O O O N ro oo o O N ro — °o O N N O — ro ro O O N CM ro —i Zl I~- tN N O O N O N ro o — r- — CM 8 5 MD O <n r-ON m m — u I ^  3 z > C o 2 a < So u S £ z a « s *c o 03 55 <u u Vi Vi O S-CJ 03 b£ fl -3 03 o fl <u J3 « s-o a o u a B u fl fl o U 03 z 03 H tt 1^ o XI 2: s 5 -a s £ C s 5 IS xi § c 0," S I CN ON 00 ON ON ro CN NO ON in Tt O o Tt co OO d d d d d OO ON r - 00 ^J, CN in NO 00 CN d 00 00 CN co o ON CO Tt NO CN •—i ON r - CN —H co oo d d d ' d 00 ON ON in oo NO r~ NO 00 CO CN ro ON oo — in d CN in oo CO CN i n o r - Tt 00 in CO ON O oo —~ d d CO ^ CO NO NO oo Tt in ro oo O 00 Tt o Tt 00 ro oo NO o CO NO 00 in Tt Tt ro VO CO mi NO CO Tt d CN NO oo CO 00 o Tt 00 r- Tt Tt ON 00 ON o NO 00 Tt — o O NO — NO 2 <-•) . 00 O 00 m CN NO d o Tt NO Tt t-00 CO o O N Tt Tt 00 CO d CN CO CO CO CN in N O CN r - r ~ d ON o CO CN r — o Tt 00 CO — N O o ro > VG TDE AO [AX < Vi U NO CO Tt NO NO 00 00 00 00 tt 00 o O N OO O N CN CO N O N O in Tt o o Tt CO r -d d d d d CN O N — 00 r - -N O in r--CN d 00 CN N O O 00 r - CO • Tt CN •— O N N O —' CN o CO r -d d d — d CN o o CO N O CN Tt CN CN CO N O ro OO in r-Tt d CN (-^  CN CN in CO r - O oo N O CO ON "~" d d CO CN Tt N O OO N O Tt o t-- O r -O N 00 Tt r -m o CO O N CN in CN o Tt 00 in CN in r -CO d CN ' CN CO O N _ oc CN m ' I Tt O N r -O N o N O CN — o 00 o r - Tt ON — Tt d o o . — o ON Tt CN O NO — Tt _; Tt d CN O N L. N O ON . r -— o — O ! NO O > Q ^ H O vi U NO CO Tt Tt O CN Tt ON — I--m CO NO X < NO Tt CN 00 00 o tt i~J Tt ON r- ON ON co CN NO ON r-Tt o o Tt CO ^_ d d d d d NO in o 00 CN CN iri NO CN d OO NO 00 NO ON r— ON Tt t-- o r- — — in d d d —> d NO , , NO in o in ON NO OO Tt in CO ON ,—. in d CN in NO ^ « CO ON ^— Tt 00 in CN ON O d d CO NO Tt 00 Tt Tt in NO > a o Tt NO Tt d — CO Tt d Tt 0 Tt £j CO 00 CN . c o i n — o — r- NO Tt d NO — NO NO CN __ —: d NO Tt ^ - ^  c o r - - — O O0 Tt NO o Tt NO Tt Tt d Tt CN NO 00 i n " . NO Tt — — o — ON NO o o -. CO CN — O CO Tt NO > o vi U X < 33 U C H A P T E R 4. P H A S E II. N A I L C O N N E C T I O N T E S T S O N R O B O T - B A S E D F O R M E D P A N E L S Table 4.8 Percentages of failure modes for random and oriented panels and loading directions in OSB panels Panel types/loading directions Percentage of pull-through mode, % Percentage of pull-out mode, % Whole data 26.7 73.3 (oriented + random panels) Random panels 28.4 71.6 Oriented panels (OSB) 25.8 74.2 0° loading in OSB 27.3 72.7 45° loading in OSB 22.7 77.3 90° loading in OSB 30.7 69.3 86 Vi -1 td z «I CM Q td ai o u. Q td r/5 < CO I H o o Pi z o Vi H r/5 td L-Z o H U td Z Z o u < z td ac pi td H CM < 3 O i 3 a -a c JS W) s o IN s a a> T3 o & s -o a © Vi ii I H o\ rr « -3.12 * * * -1.01 00 Z -3.10 * * * -2.80 * * * -1.89 * -0.36 OO Z o CN CN * * ro T t 00 * o CN oo Z CN o oo z S S t l MO OO © oo z t--T t OO z ro O i oo Z 'NO ro © i z T t 00 1 * T t ON CN * * * £ "S ON in * tt * o T t * * * O T t * * * CN f-; * ON ON O oo z in CN * * * § s oo CN * * * o (N * * ON © CN * CN CN OO z CN O oo z in T t CN * * s s m ro T t * * * Tt O ro * * * in ro ro * * r-oo * OO 00 * ro ro * * * s ON © OO Z ro o 00 CN ©' 00 z T t © © 1 cz> z ON O CN i * * ON oo * s 5 § 5 O ON ro * * * o CN * * oo oo CN * * * ON WO oo z. 00 ro o oo z m vq ro * * * § © ON t> i * * * ro o T t 1 * * * ro ON • * tt * in T t i tt * * oo ro in i * * * r-ro CN • 1 * * * K, N/mm < O i—i o t * * * *o in in • * * tt in 00 i * * * i * * * ON in • * * * ON 00 ro • * * tt Properties 03 « T3 CJ "© _fl Vi £ o -a OJ a a •r c O a fl T3 a _ © 0 o 6X) _fl -3 es 0 fl -a 03 _© 0 1 ^ a. o oo Vi Z H A H C a) JS it c 60 -a a a Xi a J: a o "3 0J} o 6 -a c o o H A H c •S 6 I I C cd o "3 ^ .SP c in cd C oj} ffl § Vi J W Z <! EH O td OS O u. Q W < 03 I H O pa O os z o Vi H </> td H Z o H U td z z O U J N H z td C/3 < X OS td CH < X u cu T5 o E cu s-3 60 3 © 3 a cu -a s s cn fl « cu a cu u fl ITS a c2 s-cu a A © m CU •<-> H o ^ H CU « O N co GO N O o GO rZ 5; © i GO o N O 1 GO V O co © • GO oo o © i GO oo o GO C O N O © GO s 5 •»•* in CN * * m CN © GO o CN i O N r~; * S s <N in CN * * in O N O GO in • GO C O CN o 1 g s s co © GO in o © GO ro CN O GO N O ro © GO s s O N co © i GO o N O © GO o N O © GO m CN GO s o o CN * * o N O o GO CN CN CN > * ro T—« 1 GO 5 § O N cn GO o GO N O T-T 1 GO © GO § O CN * * m O N o 1 GO CN O C O 1 * * * CN O N C O * * •se-K, N/mm < O 00 © GO N O in i GO O CN CN 1 * * CN N O C O 1 tt * * Properties Random to oriented panels 0° loading to 45° loading 45° loading to 90° loading 0° loading to 90° loading ue CO Z > ean o H £ A H cu c c CL> ca J 3 Q. T3 eu * . 2 c ' C o ' S o - o fi o ca o H T3 S cu ca E o 3 H C c -^^  ^ i ^ <u <U B _> cu nega CO . 2 nega H C c ca cu J 3 u Q . ^ ^ O D, O "ca s 3 " S - a '> tyi • 3 -t-> c c/l O c o on on u <U <u i H in o UIO o UIO H A H ived hen <D O — alue: * cn <u alue: * * 3 ?• val ean ;ant H E i n 00 oo co ca § I l -a 1 ca m ra 3 00 CQ C/> w Z <C 0-Q W O u-Q UJ C/5 <! 03 I H O 03 o oi z o </> H t/3 UJ H Z o H u UJ z z o u J z UJ oo < UJ H CM 33 -a o £ a s . S fl O i a u ii ~a a fl Vi C & o u fl 03 E u a ii o< © Vi ii re ii CS H 0.48 NS 0.22 NS 1.57 NS 1.24 NS -1.14 NS 0.68 NS -0.42 NS 0.18 NS s s J 2.58 •se-tt tt -1.40 NS 2.64 * •se-tt 1.24 NS S "3 2.20 •se-tt -0.25 NS 3.95 * •se-tt 3.35 * •se-tt s 5 0.77 NS , -0.27 NS 1.38 NS 1.11 NS S 5 -0.26 NS 0.42 NS 1.65 •se- 1.71 •ss-S S § 2.52 * * -1.49 NS 1.85 tt 0.53 NS s § 5 2.40 * * -0.67 NS 3.07 •se-tt * 2.13 * •se-i 1.43 NS -2.46 # * -1.34 NS -3.59 tt * •se-K, N/mm < o o • a OO -1.50 NS -2.87 •se-tt * -3.75 tt •se-tt Properties Random to oriented panels 0° loading to 45° loading 45° loading to 90° loading 0° loading to 90° loading c O . T3 CD ON OO •n O o c Cd 00 z H A H c JS c oO -a c 03 C CD CD JS it a cd o '3 00 o 3 T3 e o o CD H A H e J3 C C3 O tC c« Cd •i-» O o s oo C H A P T E R 4. P H A S E II. N A I L C O N N E C T I O N T E S T S O N R O B O T - B A S E D F O R M E D P A N E L S Loaded at the 45° direction relative to the face flake alignment, the connection had significant differences in many nail properties between the two failure modes. Initial stiffness, connection strength, maximum strain energy, ultimate strain energy, and D2 under the pull-out mode were all higher than those under the pull-through mode. Only yield displacement under the pull-out mode was smaller than that under the pull-through mode (Tables 4.6 and 4.9). Only two characteristics, Z)/ and D2, were not significantly different between the two failure modes when the OSB to lumber joint was loaded perpendicular to the OSB face flake orientation (90°). Initial stiffness and connection strength were both higher under the pull-out mode than those under the pull-through mode. All the other properties had comparatively smaller means under the pull-out mode (Tables 4.7 and 4.9). In summary, most nail-connection properties were significantly different between the two main failure modes, either considering the whole data set or individual panel types or loading directions in OSB panels. Initial stiffness and strength under the pull-out mode were always higher than those under the pull-through mode for all different conditions. 4.3.2.1.2 Panel Types Two panel types, random flakeboards and oriented flakeboards, were manufactured in the laboratory and then assembled into the panel-lumber joints after a few weeks of conditioning. Under the same nail-property testing conditions, the two panel types were found from T-test analysis to have significant differences in some properties for the two failure modes (Tables 4.10 and 4.11). 90 C H A P T E R 4. P H A S E II. N A I L C O N N E C T I O N T E S T S O N R O B O T - B A S E D F O R M E D P A N E L S For the pull-through mode, significant differences existed in connection strength, ultimate displacement, maximum strain energy and ultimate strain energy. OSB panels had smaller ultimate displacement, lower strength, and reduced maximum and ultimate energies compared to random flakeboards (Tables 4.3, 4.4 and 4.10). Under the pull-out mode, maximum and ultimate displacements, and maximum and ultimate strain energies were significantly different between random and oriented panels. All of the above properties were higher for random panels compared to OSB panels (Tables 4.3, 4.4 and 4.11). In general, OSB panels had smaller displacements, lower strengths and reduced strain energies compared to random panels either in the pull-through mode or in the pull-out mode. 4.3.2.1.3 Loading Directions in OSB-Lumber Joints Three loading directions in OSB panels were studied. T-test results are summarized in Tables 4.10 and 4.11 respectively. There was no significant difference in nail-connection properties between loading at 0° and 45° loading angles when nails failed in the pull-through mode (Table 4.10). Only connection strength was significantly different between the two loading directions in the pull-out failure mode. Connection strength, under the pull-out failure mode, at the 45° loading angle was higher than that at the 0° loading angle, e.g. maximum loads of 1221N and 1164N, respectively (Tables 4.5, 4.6, and 4.11). For specimens that failed in the pull-through mode, initial stiffness, connection strength, maximum strain energy and ultimate strain energy were significantly different between 91 C H A P T E R 4. P H A S E II. N A I L C O N N E C T I O N T E S T S O N R O B O T - B A S E D F O R M E D P A N E L S 45° and 90° loading angles. And all of the properties above had higher means at the 90° loading direction (Tables 4.6, 4.7 and 4.10). In comparison, specimens that failed in the pull-out mode had more significantly different nail-connection properties between the two loading directions. Among these characteristics, only initial stiffness at the 90° loading angle was higher, and other properties including maximum displacement, yield displacement, ultimate displacement, maximum strain energy and ultimate strain energy at the 90° loading angle was all lower than those at the 45° loading angle (Tables 4.6, 4.7, and 4.11). The specimens with the pull-through failure mode had three properties significantly different between 0° and 90° loading angles: initial stiffness, connection strength and ultimate strain energy (Table 4.10). All of the significant properties at the 90° loading direction were higher. However, the specimens that failed in the pull-out mode had more properties significantly different between the two loading directions: initial stiffness, connection strength, and (maximum, yield and ultimate) displacements. And initial stiffness and connection strength at the 90° loading angle were obviously higher. Other properties, i.e. all three displacements, at the 90° loading direction were smaller comparatively (Tables 4.5, 4.7, and 4.11). In conclusion, loading angles affected the nail-connection performance under the two failure modes. Usually, specimens with the pull-out mode had more significantly impacted properties. There was not significant difference in nail properties between 0° and 45° loading directions for the pull-through mode. But there did exist differences in some nail properties between 45° and 90° loading angles, and between 0° and 90° 92 C H A P T E R 4. P H A S E II. N A I L C O N N E C T I O N T E S T S O N R O B O T - B A S E D F O R M E D P A N E L S loading angles. Initial stiffness and/or strengths at the 90° loading direction were always higher than those at 0° and 45° loading directions for both failure modes. Connection strength at the 45° loading direction was higher than that at the 0° loading angle but close to that at the 90° angle (Tables 4.5-4.7). Tables 4.5-4.7 also summarize panel local density, thickness ratio and lumber specific gravity of the specimens under varied conditions. There was much difference in average local density between the specimens failing in the pull-through and pull-out modes for the whole data set, the random panels, the oriented panels (OSB), and the loading directions in OSB. In comparison, the differences in thickness ratio and lumber specific gravity were much less obvious. Under the same failure mode, there were some differences between random and oriented panels, and between different loading directions in local density, thickness ratio and lumber specific gravity. It was these differences under varied conditions that led to different nail connection properties. By regression analysis, these influencing factors from flakeboards (side member) and lumber (main member) and their relationships to nail-connection properties will be further discussed. 4.3.2.2 Multivariate Regression Models Three main factors, panel local density (LD), lumber specific gravity (G), and board to flake thickness ratio (TR), were chosen to set up multivariate regression models. Tables 4.12-4.16 show regression models under different conditions. Only significant regression models are shown below. 93 C H A P T E R 4. P H A S E II. N A I L C O N N E C T I O N T E S T S O N R O B O T - B A S E D F O R M E D P A N E L S Table 4.12 Regression models of oriented panel nail connection properties Regression models R2 F TV 55^ K(N1mm) = 4345.170-1935.736L/J 2 -13016.662G 2 - 833.14577? + 4 2 2 . 9 3 3 Z £ > x 77? +1295.79577? x G 0.3497 37.21 351 97212368.36 Pmm(N) = 281.953 -2069 .347LD 2 -2.900TR 2 + 2442.200ZZ) - 2449.822G + 2803.667LD x G + 168.75977? x G 0.6111 89.32 347 6763806.93 Umm (N.mm) = 59174.085 - 22966.574ZD 2 + 323269.461G2 + 62962.700ZD - 330224.013G - 2596.336ID x TR + 3968.73 \TRxG 0.1490 9.75 340 3764379003.42 Ayld (mm) = -1.199 + 0.42077? - 0.30\LD x TR + 6.372LDxG-QA9\TRxG 0.0713 6.38 336 74.49 Uyld(N.mm) = 3986.121-574.749ZD 2 +26011.246G 2 + 1903.721ZD + 65.03477? - 22222.356G - 98.308ZZ) x TR 0.1069 6.48 331 5492737.62 Aull (mm) = 11.280 - 23.425ID 2 + 61.5\1LD - 55.980G - 2.812LD x TR + 3.81877? x G 0.1119 8.12 327 5795.64 Uull(N.mm) = -8153.788 -42710.144ZD 2 + 95707.833ZD - 43686.878G - 2791.980ZD x TR + 3803.849Ti? x G 0.1837 14.40 325 7265910173.99 Note: All equations are significant at the 0.001 level, and all variables in the models are significant at the 0.10 level. All regression models were built up using the SAS Program. i?-square, F-value, valid data pairs (AO and sum of squares of error (SSE) were also automatically recorded 94 C H A P T E R 4. P H A S E II. N A I L C O N N E C T I O N T E S T S O N R O B O T - B A S E D F O R M E D P A N E L S through the regression process. It is necessary to note that the total number of valid data pairs (AO refers to the data points, which really contribute to the regression models. This is the outcome of reweighing the original data set such that some data were automatically removed during the regression procedure due to their studentized residuals either > 3, or <-3 (SAS, 1996). Table 4.13 Regression models of random panel nail connection properties Regression models R2 FN SSE K(N7mm) = -27257'Al'4 + 156.3 1 577?2 + 120735.710G + 905.668ZD x TR -19373.687ZD x G - 9212.97977? x G 0.1950 8.24 175 54102689.15 p m a x (AA) = -565.256 - 2043.727ZD 2 + 2268.020G 2 + 3427.35 ILD 0.3515 30.71 173 4568667.98 A m a x (mm) = -20.150 - 65.324ID 2 + 200.463ID - 198.912G - 10.442LD x 77? + 18.42977? x G 0.1170 4.43 172 2365.40 <Vmax (N.mm) = -73993.332 - 69648.585LD 2 + 332.6227T?2 + 49850.555G 2 +199519.833ZZ) - 9 3 3 9 . 1 4 2 Z £ > x 7 7 ? 0.1341 5.08 169 2939278952.65 A„„ (mm) = -31.466 - 87.345ID 2 + 246.783ZD - 214.127G - 12.425ZZ) x TR + 22.16177? x G 0.2640 11.26 162 2974.09 Uuh (N.mm) = -115775.877 -126189.556ZD 2 + 476.262Ti?2 + 90071.832G 2 +316517.106ID -13228.06517) xTR 0.2616 11.12 162 4706534826.81 Note: All equations are significant at the 0.001 level, and all variables in the models are significant at the 0.10 level. 95 C H A P T E R 4. P H A S E II. N A I L C O N N E C T I O N T E S T S O N R O B O T - B A S E D F O R M E D P A N E L S Table 4.14 Regression models when loading along face flake alignment Regression Models R2 F N SSE K(N 1 mm) = -822.168 -1242 .150LD 2 + 8422.01 177>x G 0.4874 40.41 87 15104178.05 p m a x (N) = -495 .129 - 2507.129Z73 2 + 4127.415773 - 45.95477? +110.64377? x G 0.7301 55.46 86 1215025.97 UyM(N.mm) = 7494.729 -1516 .959Z73 2 - 8.40877?2 + 4 0 0 1 4 . 3 3 2 G 2 + 208.21077? - 38312.558G + 4479 .87677>xG 0.2900 5.17 82 1641102.28 Uuh(N.mm) -= -14053.023 - 56929.839Z73 2 + 82557.990773 0.2733 15.04 82 1835506142.38 73, = 9.257 + 4.238Z73 2 - 2 5 . 9 0 5 G 2 -21 .309773 + 33.684773 x G 0.2044 4.95 81 29.70 Note: All equations are significant at the 0.001 level, and all variables in the models are significant at the 0.10 level. Table 4.15 Regression models when loading at 45° to face flake alignment Regression models R2 F N SSE K(Nlmm) = 6261.080 - 607.43877? - 1 2 3 7 3 . 1 2 1 G +193.31677) x 77?+ 1143.92177? x G 0.3073 18.97 175 49442306.08 Pmax(/V) -= 2 7 1 . 9 2 2 - 1 8 9 8 . 2 5 4 Z Z ) 2 -7 .90477? 2 +196.10677? - 3961.543G + 7916 .9847D x G 0.6418 59.83 172 3359896.27 A m a x (mm) = -21.443 + 43.161773 + 2.91777? - 4 . 1 6 3 7 7 » x 77? 0.1247 7.88 169 1558.21 Umax(N.mm) = -30841.474 + 54888.645773 + 3366.87677? -4700.114773x77? 0.2124 14.74 167 1717141108.91 UuU (N.mm) = -28633.966 + 63422.223773 + 3311.82577? -4855.727773x77? 0.2032 13.35 160 3684125088.25 73, = -2 .499 + 0.02177?2 +18 .117SG-1.28877? x G 0.1006 5.78 158 49.84 Note: All equations are significant at the 0.001 level, and all variables in the models are significant at the 0.10 level. 96 C H A P T E R 4. P H A S E II. N A I L C O N N E C T I O N T E S T S O N R O B O T - B A S E D F O R M E D P A N E L S Table 4.16 Regression models when loading across face flake alignment Regression models R2 F N SSE K(Nlmm) = -15149.499-3316.045LD 2 -141071.308G 2 -1347.60577? +102801.875G + 596.71877) x 77? + 2193.24677? x G 0.3853 8.46 87 25216927.35 P m a x ( T V ) = 56.438 -1404.61077) 2 + 2584.786ZZ) - 76.72977? +189.12177? x G 0.4014 13.75 86 1600779.66 A m a x (mm) = 68.150 - 8.027LD - 4.719177? -118.883G +10.30677? x G 0.2969 8.34 83 581.24 <7max (N.mm) = 80400.980 - 2753.602ZD 2 - 6735.88977? -159579.719G + 15159.49277? x G 0.1589 3.73 83 824501698.54 Ayld(mm) = 10.656 + 3.938Z7>2 +58.965G 2 -6.63977) + 0.29977? - 42.792G - 0.68477? x G 0.4353 9.76 82 3.73 Uyld(N.mm)-4239.853 + 941.504Z7)2 + 3.03 477?2 +21967.541G 2 -1562.13077) - 16339.548G -172.72177? x G 0.2627 4.51 82 646630.38 A„„ (mm) = 90.494 - 5.41977? - 169.668G - 0.84877) x 77? +13.22677? x G 0.2573 6.76 82 1105.59 Uull(N.mm) = 55296.840-24708.863Z7) 2 -162393.104G - 3086.59977) x 77? +151822.0947D x G + 4107.19677? x G 0.1978 3.70 80 1063186931.80 D2 =79.749- 27.1537D 2 -254.412G 2 -11.25 777? + 3.320773 x 77? +19.82877? x G 0.1842 3.30 78 1843.11 Note: All equations are significant at the 0.001 level, and all variables in the models are significant at the 0.10 level. It is indicated from Tables 4.12-4.16 that the nail connection properties under different situations are impacted by the various combinations of panel local density (LD), and/or 97 C H A P T E R 4. P H A S E II. N A I L C O N N E C T I O N T E S T S O N R O B O T - B A S E D F O R M E D P A N E L S panel to flake thickness ratio (TR) and/or lumber specific gravity (G) with different significance degrees. By examining the relationships of predicted and measured nail connection properties, initial stiffness and maximum load for example, it is found that there exist significant linear relationships at 0.001 significance level between predicted and measured initial stiffness and maximum load for both the OSB to lumber connection and the random flakeboard to lumber connection, and for different loading directions in the OSB to lumber connection (Figs. 4.4-4.13), which means the predicted models are mostly effective within the studied variable ranges, either for oriented and random flakeboard panels, or for different loading directions in OSB panels. Owing to the complicated relationships between nail connection properties and the three selected processing parameters amongst the regression models, the effects of the main member (SPF lumber) and the side member (OSB and/or random flakeboard panels) will be discussed separately by exemplifying general variable ranges in the practical application of flakeboard to lumber connections. Even though the regression models look very complex and hard to explain, they may be described if certain application conditions of the lumber/flakeboard connection were known or fixed. For example, we can discuss the effect of flakeboard characteristics on the nail connection performance if the lumber properties such as specific gravity (G) are known, and vice versa. 98 C H A P T E R 4. P H A S E II. N A I L C O N N E C T I O N T E S T S O N R O B O T - B A S E D F O R M E D P A N E L S Measured initial stiffness, N/mm Fig. 4.4 Relationship of measured and predicted initial stiffness in OSB to lumber connection 1700 -, Measured maximum load, N Fig. 4.5 Relationship of measured and predicted maximum load in OSB to lumber connection 99 C H A P T E R 4. P H A S E II. N A I L C O N N E C T I O N T E S T S O N R O B O T - B A S E D F O R M E D P A N E L S Fig. 4.6 Relationship of measured and predicted initial stiffness in random flakeboard to lumber connection X B -a + | * • y = 0.3684x + 752.46 R z = 0.339 700 900 1100 1300 Measured maximum load, N 1500 1700 Fig. 4.7 Relationship of measured and predicted maximum load in random flakeboard to lumber connection 100 C H A P T E R 4. P H A S E II. N A I L C O N N E C T I O N T E S T S O N R O B O T - B A S E D F O R M E D P A N E L S 2500 £ E Z 2000 CU £ 1500 •a s. 1000 500 0 45-degree line • . • |T y = 0.549x +467.63 R 2 = 0.5232 • *• • • • 500 1000 1500 2000 Measured init ial stiffness, N/mm 2500 3000 Fig. 4.8 Relationship of measured and predicted initial stiffness when loaded along face flake alignment in OSB to lumber connection 7S S5 O E •a O N 1600 1200 800 400 y = 0.7304x + 293.38 R z = 0.7293 45-degree line 400 800 1200 1600 Measured maximum load, N 2000 Fig. 4.9 Relationship of measured and predicted maximum load when loaded along face flake alignment in OSB to lumber connection 101 C H A P T E R 4. P H A S E II. N A I L C O N N E C T I O N T E S T S O N R O B O T - B A S E D F O R M E D P A N E L S 2500 E E Z 2000 1500 • | 1000 •a •a 500 y = 0.3538x + 771.85 R 2 = 0.3447 45-degree line 500 1000 1500 2000 Measured init ial stiffness, N/mm 2500 3000 Fig. 4.10 Relationship of measured and predicted initial stiffness when loaded at 45° to face flake alignment in OSB to lumber connection 600 800 1000 1200 1400 Measured max imum load, N 800 Fig. 4.11 Relationship of measured and predicted maximum load when loaded at 45 c to face flake alignment in OSB to lumber connection 102 C H A P T E R 4. P H A S E II. N A I L C O N N E C T I O N T E S T S O N R O B O T - B A S E D F O R M E D P A N E L S 2500 S E Z 2000 I 1500 • | 1000 •a o 500 45-degree line ~y = 0.427x + 809.87 R z = 0.4518 — * — s - ^ f 500 1000 1500 2000 2500 Measured init ial stiffness, N/mm 3000 3500 Fig. 4.12 Relationship of measured and predicted initial stiffness when loaded across face flake alignment in OSB to lumber connection 800 -! 1 T — — i 1 1 1 1 i 1 800 900 1000 1100 1200 1300 1400 1500 1600 1700 Measured maximum load, N Fig. 4.13 Relationship of measured and predicted maximum load when loaded across face flake alignment in OSB to lumber connection 103 C H A P T E R 4. P H A S E II. N A I L C O N N E C T I O N T E S T S O N R O B O T - B A S E D F O R M E D P A N E L S 4.3.2.3 Parametric Study of Regression Models OSB or random flakeboard sheathing material usually has a 9.5 mm thickness and a density of 625-640 kg/m3. Flake thickness used in the OSB manufacture is in the range of 0.50-0.90 mm. SPF lumber is one of the common connection framing materials in the practical joints. Combined these practical conditions with the parameters applied in this research project, we take SPF's average specific gravity as 0.430, OSB or random flakeboard's average panel to flake thickness ratio as 11.9 (9.5 mm panel and 0.80 mm flake) and average panel local density (15x15 mm zones) as 0.630 g/cm . Then we can separately investigate the impacts of main member (lumber), and side member (OSB or random flakeboard) characteristics on nail connection properties. As the most important nail-connection properties, initial stiffness and maximum load (connection strength) were selected for discussion in this section. 4.3.2.3.1 Impact of Lumber Specific Gravity When a typical 9.5 mm OSB with 11.9 thickness ratio (0.80 mm thick flake) and average panel local density of 0.630 g/cm3 were chosen, maximum load (also yield and ultimate loads) would enhance with the increase of lumber specific gravity both for OSB or random flakeboard, and for different loading directions in the OSB to Lumber connections, as shown in Table 4.17. For a connection with OSB as a side member, initial stiffness has a second-degree relationship with specific gravity, which increases with the enhancement of lumber specific gravity in the regression range of G. However, for the connection with random flakeboard as a side member, initial stiffness decreases linearly with increased G. For OSB to lumber connections loaded along and at 45° to the face flake alignment, initial 104 C H A P T E R 4. P H A S E II. N A I L C O N N E C T I O N T E S T S O N R O B O T - B A S E D F O R M E D P A N E L S stiffness has positive linear relationships with lumber specific gravity. But for the OSB to lumber connection loaded across the face flake alignment, initial stiffness has a second-degree relationship to specific gravity; it reaches its maximum when G gets close to 0.460 (Table 4.17). Table 4.17 Effect of lumber specific gravity (G) on initial stiffness and maximum load when TR=11.9 and LD=0.630 g/cm3 Oriented flakeboard (OSB) K(N / mm) =-\30\7G2 + 15420G - 3166.8 P m a x (TV) = 1324.7G + 588.6 Random flakeboard K(N / mm) =-1104.2G +1668.1 7m a x(A0 = 1927.8G + 375.9 OSB loaded along face flake orientation K(N/mm) = 5305.9G-1315.2 (7V) = 1316.7G +563.2 OSB loaded at 45 degrees to face flake orientation K(N / mm) = 1239.5G + 481.9 Pmm(N) = 1026.2G + 732.88 OSB loaded across face flake orientation K(Nlmm) = -141071.3G2 + 128902.4G - 28029.2 Pmax(7V) = 2250.5G + 21429.3 105 C H A P T E R 4. P H A S E II. N A I L C O N N E C T I O N T E S T S O N R O B O T - B A S E D F O R M E D P A N E L S 4.3.2.3.2 Influence of Panel Local Density Assume uniform lumber with G=0.430, and flakeboard panels with the panel thickness of 9.5 mm and the flake thickness of 0.80 mm (i.e. 11.9 for thickness ratio) were chosen, the influence of panel local density (density variation) on the connection properties could be discussed separately here (Table 4.18). Table 4.18 Impact of panel local density (LD) on initial stiffness and maximum load when TR=11.9 and G=0.430 Oriented flakeboard (OSB) K(N/mm) =-1935.7 LD2 + 5032.3Z7)-1345.5 pmm (TV) = -2069.3Z7)2 + 3647.8Z7) - 318.6 Random fla ieboard K(N 1 mm) = 2446.8Z7) - 348.2 pimx (TV) = -2043.7ZD2 + 3427.4Z7) -145.9 OSB loaded along face flake orientation K(N 1 mm) =-\2A2.\LD2 + 3621.5Z7)-822.2 P m a x (TV) = -2507.1ZD2 + 4127.4Z7) - 475.8 OSB loaded at 45 degrees to face flake orientation K(Nlmm) = 2300.5ZD- 434.4 pmm (TV) = -1898.3ZD2 + 3404.3ZD - 217.2 OSB loaded across face flake orientation K(N lmm) = -33\6.2LD2 +7100.911)-1842.4 — pmm (TV) = -1404.6ZD2 + 2584.87,7) + 111.1 106 C H A P T E R 4. P H A S E II. N A I L C O N N E C T I O N T E S T S O N R O B O T - B A S E D F O R M E D P A N E L S For random or oriented flakeboard and for all loading directions in OSB panels, maximum load has second-degree models with panel local density. Within regression ranges, both oriented and random flakeboards get to peak maximum loads when LD is around 0.800-0.850 g/cm3; the 0° and 45° loading directions have peak maximum loads when LD lies in the range of 0.850-0.900 g/cm3. For the OSB to lumber connection, initial stiffness has a second-degree relationship with panel local density and improves with the increase of LD within the regression range. For the random flakeboard to lumber connection and the OSB to lumber connection loaded at 45° to the face flake alignment, initial stiffness has positive linear relationships to panel local density. For the OSB to lumber connection loaded along and across the face flake alignment, initial stiffness has second-degree relationships to OSB local density. The 0° loading direction leads to larger initial stiffness with the increase of OSB local density for the whole regression range, but the 90° loading direction reaches a maximum initial stiffness when LD is around 1.00 g/cm . 4.3.2.3.3 Impact of Panel to Flake Thickness Ratio For the chosen SPF lumber (main member, average G=0.430), and the side member flakeboard with a 9.5 mm thickness and an average local density of 0.630 g/cm3, the impact of thickness ratio (flake thickness) on initial stiffness and maximum load can be individually examined (Table 4.19). For the OSB to lumber connection, initial stiffness linearly decreases with increased thickness ratio (i.e., decreased flake thickness when panel thickness is fixed at 9.5 mm); however, maximum load has a second-degree relationship with thickness ratio instead, 107 C H A P T E R 4. P H A S E II. N A I L C O N N E C T I O N T E S T S O N R O B O T - B A S E D F O R M E D P A N E L S and it reaches its peak value when TR is around 12.0-12.5. For the random flakeboard to lumber connection, maximum load is not affected by thickness ratio (flake thickness), but initial stiffness has a second-degree relationship to thickness ratio and it gets to its minimum value when TR is close to 11.0. Table 4.19 Impact of panel to flake thickness ratio (TR) on initial stiffness and maximum load when G=0.430 and LD=0.630g/cm3 Oriented flakeboard (OSB) K(N/mm) = -9.5TR +117.0 pmm (TV) = -2.977?2 + 72.677? + 705.3 Random flakeboard K{N/mm) = \56.3TR2 -3391.ITR +19411.3 Pmax, not affected by 77? OSB loaded along face flake orientation K, not affected by TR P m a x (TV) = 1.677?+ 1110.1 OSB loaded at 45 degrees to face flake orientation K{NI mm) = 6.277? + 940.6 Pmax(7V) = -7.977?2 +196.177?-40.2 OSB loaded across face flake orientation K'N/mm) = -28.677? +1655.1 PmW = 4.677? + 1127.4 For the OSB to lumber connection loaded along the face flake alignment, its initial stiffness is not influenced by thickness ratio, nevertheless, its maximum load linearly enhances with the increase of thickness ratio. When loaded at 45° to the face flake alignment, the OSB to lumber connection has initial stiffness improving with the increase 108 C H A P T E R 4. P H A S E II. N A I L C O N N E C T I O N T E S T S O N R O B O T - B A S E D F O R M E D P A N E L S of thickness ratio, and maximum load, which has a second-degree relationship to thickness ratio, reaches its peak when TR is around 12.0. When the OSB to lumber connection was loaded across the face flake alignment direction, initial stiffness linearly reduces with increased thickness ratio (decreased flake thickness for 9.5 mm panel), however, maximum load linearly becomes larger as thickness ratio increases, as shown in Table 4.19. 4.3.2.4 Prediction of Nail Connection Properties Using Hankinson's Formula Loading at an angle to lumber grain is one of the important considerations in wood structure design. The following equation, known as Hankinson's formula, is often used for computing the lateral resistance of wood connections (Canadian Wood Council, 1990): Nr = — (4.2) Prsm2 9 + Qrcos2 9 where: Nr: factored resistance at any angle Bio grain Pr: factored resistance parallel to grain Qr: factored resistance perpendicular to grain 9: angle between grain direction and direction of load The formula above is modified here for studying the relationship of loading directions in the OSB to lumber connections. Here, 9 is the angle between loading directions and the OSB face flake orientation. Based on the nail connection properties such as loads and 109 C H A P T E R 4. P H A S E II. N A I L C O N N E C T I O N T E S T S O N R O B O T - B A S E D F O R M E D P A N E L S initial stiffness along and across the face flake alignment, we may predict the nail performance at any loading angle using Equation 4.2. Table 4.20 Comparisons of measured and Hankinson's formula-predicted initial stiffness and maximum load Na i l propert ies and fa i lure modes Measu red value at 0° Measu red value at 90° Measured value at 45 ° Predicted value at 45° In i t ia l stiffness K, N/mm W h o l e data set 1016 1429 1192 1188 Pul l - th rough 536 1043 729 708 Pul l -out 1195 1601 1328 1369 M a x i m u m load, N Who le data set 1085 1223 1159 1150 Pul l - th rough 875 1152 950 995 Pul l -out 1164 1254 1221 1207 Table 4.20 summarizes predicted and measured initial stiffness and maximum load data at the 45° loading direction. Obviously, average initial stiffness and maximum load at the 45° loading direction can be well predicted based on the average nail properties at 0° and 90° loading angles (see Tables 4.5-4.7). The nail properties at any loading angle may also be predicted using Equation 4.2. Figures 4.14 and 4.15 show the relationship of a loading angle to initial stiffness and maximum load based on the average properties at 0° 110 C H A P T E R 4. P H A S E II. N A I L C O N N E C T I O N T E S T S O N R O B O T - B A S E D F O R M E D P A N E L S and 90° loading angles. More loading directions (angles) should be set up in the future study to determine the difference between predicted and measured nail properties. ••—Whole data set » Pull-through - _- Pull-out 0 30 45 60 75 90 Load ing angle relat ive to face a l ignment , degrees Fig. 4.14 Relationship of initial stiffness with loading angle in OSB to lumber connection -•—Whole data set --a—Pull-through - Pull-out 1400 Z 800 -! r — 1 1 1 ' 0 30 45 60 75 90 Loading angle relative to face a l ignment , degrees Fig. 4.15 Relationship of maximum load with loading angle in OSB to lumber connection 111 C H A P T E R 4. P H A S E II. N A I L C O N N E C T I O N T E S T S O N R O B O T - B A S E D F O R M E D P A N E L S 4.4 C O N C L U S I O N S According to T-tests and the regression analysis on the main experimental data, the following conclusions can be made: • Significant differences existed in most nail-connection characteristics between two failure modes: pull-through and pull-out, and the specimens that failed in the pull-out mode always had higher initial stiffness and connection strength (loads) than those in the pull-through mode. • Compared to OSB panels, random panels had higher connection strength (loads) for the pull-through mode, bigger maximum displacement for the pull-out mode, and larger maximum strain energy and ultimate strain energy and larger ultimate displacement for both failure modes. • There was not much difference in nail properties between 0° and 45° loading angles under the pull-through mode, but connection strength at the 45° loading direction was significantly higher than that at the 0° loading angle under the pull-out mode. • The 90° loading direction in OSB panels led to significantly different nail properties for the two failure modes, especially higher initial stiffness and connection strength (loads), compared with 0° and 45° loading directions. • Most of the nail connection properties were influenced by different combinations of panel local density (LD), board to flake thickness ratio (overlaps, TR), and 112 C H A P T E R 4. P H A S E II. N A I L C O N N E C T I O N T E S T S O N R O B O T - B A S E D F O R M E D P A N E L S lumber specific gravity (G) for the OSB to lumber connection and the random flakeboard to lumber connection, and for different loading directions in the OSB to lumber connection. • A parametric study was conducted to illustrate the potential use of the information developed in this paper. In general, higher lumber specific gravity and panel local density mostly lead to better initial stiffness and connection strength (loads) within the regression ranges and the certain lumber and panel performance. But the effect of panel to flake thickness ratio is comparatively complex. Different types of connection or loading conditions may produce opposite changing tendencies. • Hankinson's equation has predicted very close initial stiffness and maximum load to measured values at 45° loading angle based on nail properties along and across OSB face flake alignment. 4.5 REFERENCES Canadian Wood Council. 1990. Wood Design Manual: The Complete Reference for Wood Design in Canada. Ottawa, Ontario, Canada. Dai, C. and P.R. Steiner. 1994a. Spatial structure of wood composites in relation to simulation of a randomly formed flake layer network. Part 2. Modeling and simulation of a randomly formed flake layer network. Wood Sci. and Technol., 28(2): 135-146 113 C H A P T E R 4. P H A S E II. N A I L C O N N E C T I O N T E S T S O N R O B O T - B A S E D F O R M E D P A N E L S Dai, C. and P.R. Steiner. 1994b. Spatial structure of wood composites in relation to processing and performance characteristics. Part 3. Modeling the formation of multi-layered random flake mats. Wood Sci. and Technol., 28(3): 229-239 Dai, C. and P.R. Steiner. 1994c. Analysis and implication of structure in short fiber wood composites. Second Pacific Rim Bio-Based Composites Symposium, Nov. 6-9, Vancouver, BC, Canada, p 17-24 He, M.H. 1997. A Study of Wood Based Shear Walls Sheathed with Oversize Oriented Strand Board Panels. M.Sc. thesis. Department of Wood Science, UBC, Vancouver, BC, Canada Hicks, C R . 1993. Fundamental concepts in the design of experiments. Saunders College Publishing, New York, USA Kamke, F.A. and M.P. Wolcott. 1991. Fundamentals of flakeboard manufacture: wood-moisture relationships. Wood Sci. and Technol., 25(1): 57-71 Lu, C. 1999. Organization of Wood Elements in Partially Oriented Flakeboard Mats. Ph.D. thesis. Department of Wood Science, UBC, Vancouver, BC, Canada Lu, C. and F. Lam. 1999. Study on the x-ray calibration and overlap measurements in robot formed flakeboard mats. Wood Sci. and Technol., 33(2): 85-95 Lu, C ; Steiner, P.R. and F. Lam. 1998. Simulation study of wood-flake composite mat structures. Forest Prod. J., 48(5): 89-93 McLain, T.E. 1975. Curvilinear load-slip relations in laterally loaded nailed joints. Ph.D. dissertation. Colorado State Univ., Ft. Collins, Colorado, USA 114 C H A P T E R 4. P H A S E II. N A I L C O N N E C T I O N T E S T S O N R O B O T - B A S E D F O R M E D P A N E L S Mohammad, M . A. H. and I. Smith. 1996. Effects of Multi-Phase Moisture Conditioning on Stiffness of Nailed OSB-to-Lumber Connections. Forest Prod. J. , 46(4): 37-44 Mohammad, M. .A.H. and I. Smith. 1994. Stiffness of nailed OSB-to-lumber connections. Forest Prod. J. 44(11/12): 37-44 Rawlings, J. O. 1988. Applied Regression Analysis: A Research Tool, Wadsworth & Brooks/Cole Advanced Books & Software. Pacific Grove, California, USA Rogerson, D.E. 1998. Corner and edge nailing test results on OSB and hemlock lumber with 2.5" common nails. MB Research Memo. MacMillan Bloedel Ltd., Vancouver, BC, Canada SAS. 1996. SAS® System for Regression, SAS Institute Inc. Sieber, D., Lam, F. and H. Prion. 1997. The Behaviour of Nailed Sheathing-to-Frame Connections under Static and Cycle Load. Research Report. Department of Wood Science, UBC, Vancouver, BC, Canada Suchsland, O. and H. Xu. 1989. A simulation of the horizontal density distribution in a flakeboard. Forest Prod. J. , 39(5): 29-33 Wang, K. and F. Lam. 1998. Robot-based research on three-layer oriented flakeboard. Wood Fiber Sci., 30(4): 339-347 Wang, K. and F. Lam. 1999. Quadratic RSM Models of Processing Parameters for Three-Layer Oriented Flakeboards. Wood Fiber Sci., 31(2): 173-186 115 C H A P T E R 5. C O N C L U S I O N S C H A P T E R 5. C O N C L U S I O N S From the study on flakeboard-lumber joints and their properties, the following conclusions can be drawn: 1. The test setup method and the new jig applied in this research were effective to acquire more data from small nailing and testing areas in the lab-based panel (240x240 mm) to 38x89 mm SPF lumber joints. 2. Simulation panel local density data around a nailing position (15x15 mm) can be easily obtained from the simulation program WinMat®, which has been used to represent real local density data in the final regression analysis due to small limited errors between simulated and measured values. 3. Significant differences in most nail properties existed between two main failure modes: pull-through and pull-out. The specimens that failed in the pull-out mode always had higher initial stiffness and connection strength (loads) than those in the pull-through mode. 4. Compared to OSB panels, random panels had higher connection strength for the pull-through mode, bigger maximum displacement for the pull-out mode, and higher maximum strain energy and ultimate strain energy and bigger ultimate displacement for both failure modes. 116 C H A P T E R 5. C O N C L U S I O N S 5. There was not much difference in nail properties between 0° and 45° loading directions in the pull-through mode, but connection strength at the 45° loading angle was significantly higher than that at the 0° loading angle in the pull-out mode. 6. The 90° loading direction in OSB panels led to significantly different nail properties for two failure modes, especially higher initial stiffness and connection strength, compared to 0° and 45° loading directions. 7. Most of the nail connection properties were influenced by different combinations of panel local density (LD), board to flake thickness ratio (77?), and lumber specific gravity (G) for the OSB to lumber connection and the random flakeboard to lumber connection, and for different loading directions in the OSB to lumber connection. 8. A parametric study was carried out to demonstrate the use of the regression model results to investigate the influence of individual changes of G, LD and TR on loads and initial stiffness. Larger lumber specific gravity and panel local density mostly lead to better initial stiffness and connection strength (loads) within the regression ranges and chosen lumber and panel properties. But the impact of panel to flake thickness ratio is comparatively complicated. Different types of connection or loading conditions may produce conflicted changing tendencies. 9. Hankinson's equation has predicted very close initial stiffness and maximum load to measured values at 45° loading angle based on nail properties along and across OSB face flake alignment, and may be have good predictions for any loading angle, which needs to be verified in the further study. 117 Some suggestions are derived from this study: 1. Further research is needed to investigate the effect of the variability of the nails on nail-connection performance. In our research, the uniform nail characteristics were assumed. 2. The effects of varied environmental conditions such as humidity and temperature on nail connection performance need to be further studied. 3. The cyclic behavior of flakeboard to lumber joints should be investigated to reflect the connection load-carrying characteristics in shear walls subjected to seismic loading. 4. Combined the X-ray density-scanning method and the small-size panel nail testing method used in this research, the nail-connection properties of commercial panels may be efficiently tested, and the relationships between nail properties, local density (or density variation), and other processing parameters may be easily set up. With a representative database on nail connection property tests, OSB nail performance in varied applications may be obtained in the near future. 5. A study of the relationship between flake orientation and loading directions is worthwhile to select an optimum product structure which may depend on the combination of the number of layers, the layer percentage, the layer alignment, etc. 6. From the regression analyses, partly strengthening the local areas around nailing positions such as adding more flakes along four sides of a panel may be one of ways to improve nail-connection performance. From the manufacture's point of view, this kind of products may be applicable in the OSB/flakeboard production. 118 C H A P T E R 5. C O N C L U S I O N S 7. Choosing uniform and/or a little high specific gravity framing lumber may be another option to enhance the nail connection properties, especially stiffness and strength. 8. Based on this research, lower nail properties under the pull-through mode may be critical to the establishment of the building and structure codes when OSB to lumber connections are involved. 119 

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:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0075171/manifest

Comment

Related Items