Tension Perpendicular to Grain Strength of Wood, Laminated Veneer Lumber, and a Wood Plastic Composite. Tracy Hummer, Research Assistant J. Daniel Dolan, Professor Michael Wolcott, Professor Wood Materials and Engineering Laboratory, Washington State University Pullman, Washington Summary A study to quantify the tension perpendicular to grain (or extrusion direction) strength of dimensional lumber, laminated veneer lumber (LVL), and wood plastic composite (WPC) structural members has been completed. Wood, LVL and WPCs display orthotropic properties where decreased properties are noted perpendicular to the primary direction of the member (member axis or extrusion direction). The study was undertaken after significant evidence that tension perpendicular failures were controlling the failure mechanisms for diaphragms with framing members using LVL related products and shear walls using WPC for sill plates. The tension perpendicular to grain study followed the ASTM D143 test standards for fabricating the test specimens and conducting the tests. The results indicate that LVL does have lower tension perpendicular to grain strength than solid wood and that WPC has higher strength than wood, but is adversely affected when employing certain die technologies for manufacture. 1. Introduction 1.1 Background Currently, most wood connection design follows the basic yield theory, which considers yielding of the fastener and member material, thereby creating permanent deformation (plastic deformation). The yield theory applies to a variety of connection designs and member materials including solid wood, laminated veneer lumber (LVL), composite strand lumber (CSL), and potentially, woodplastic composites (WPC). Recent tests of two different structural building systems have highlighted performance governed by the strength perpendicular to grain properties. Specifically, these systems tests included (1) diaphragms with framing members using I-joists with laminated veneer lumber (LVL) flanges and (2) shear walls using wood plastic composites (WPC) as sill plates, that the. In both cases, strength was affected by nail densities, nail spacing, and row spacing. For instance, laminated veneer lumber (LVL) has shown signs of premature splitting of flanges in I joists and rim joists. Interpretation of these structural tests prompted us to further investigate strength properties by testing the materials loaded perpendicular-to-grain. Part of the results are directly tied to multiplebolt connection tests results and will be used in the future to develop a relationship between the connection capacity and the tension perpendicular to grain strength. 1.2 Objective The purpose of this study was to quantify and compare the strength of solid sawn lumber, LVL, and WPC perpendicular to principal member axis. This direction corresponds with the axis perpendicular-to-extrusion for the WPC members and perpendicular-to-grain for solid wood and LVL. To develop a connection design methodology (directly related to most building assembly designs) based on capacity rather than yield theory, the tension perpendicular to grain strength values need to be quantified since splitting of the member is the most common failure mechanism
for wood-based material connections. Once material properties are accurately quantified, proper guidelines can be set based on capacity. 2. Methods and Materials 2.1 Specimen Preparation All specimens were prepared for testing at in accordance with the ASTM.D143-94 (2000) standard for specimen preparation and testing. Specimens were cut into the dog bone shape with dimensions Illustrated in Figure 1 and varying thicknesses depending on the inherent thickness of the particular product. Specimens were constructed from Douglas-fir lumber, LVL, and WPCs. The wood specimens were cut from connection test specimens from a previous study by Knudson (2006). The LVL specimens were cut from LVL obtained from local sources and manufactured by three different manufacturers. The WPC specimens was extruded from poly(propylene)-wood formulations detailed elsewhere (DuChateau 2005). Despite the standard Figure 1 Specimen size requirements specimen thickness of 51-mm, the actual thickness of the specimen varied depending on dimensions of the material thickness. Specimens cut from dimension lumber were either 38-mm or 51-mm thick depending on the raw material dimensions (e.g. 38-mm or 89-mm, respectively). LVL specimens were limited 46-mm. WPCs specimens were produced with an average thickness of 17mm. A total of 278 wood specimens were cut and tested for tension perpendicular-to-grain, including 146 specimens from 38x lumber and 132 specimens from 89x lumber. The wood used for testing was Douglas-Fir Larch dimensional lumber (38x89, 38x140, 38x190, 38x254, 89x89, and 89x140) that had previously been tested at the WMEL for multiple-bolt timber connection strength study by Knudson (2006). All specimens for this study were cut from undamaged portions of the connection test specimens. LVL specimens were sampled from multiple manufacturers. A total of 69 specimens were prepared and tested from commercially available boards of lengths varying from approximately 2.4 to 4.3 meters. Specimens were cut from the center of the board; 25mm away from any edge or end. At least three specimens were cut from each board at random locations within the board. WPC specimens were cut from beams that were extruded at the WMEL. The specimens were cut from the outer edges of the beams. These particular beams were produced using die equipped with a stranding plate (Laver 1996), which produces oriented wood particle in the extruded form. A total of 33 specimens were prepared and tested for tension perpendicular-to-extrusion direction. All specimens (wood, LVL, and WPCs) were first cut into strips 51mm high. If the material exceeded 51mm thickness, they were also cut to meet that 51mm depth requirement at this time. Then, hole centers were located for subsequent drilling using a 25mm diameter Forstner drill bit mounted in a drill press and supported by a piece of wood underneath the specimen. This technique was found to minimize the weakening the hole edges during drilling. Once the both holes were drilled completely though the material, specimens were trimmed to fit the 64mm width requirement. 2.2 Testing Procedure The solid wood, LVL, and WPCs were tested using the Instron electro-mechanical universal test machine equipped with a 8.9-kN load cell. As specified by ASTM D143-94, the specimens were
placed securely in tensile grips, as shown below in Figure 2, and loaded at a constant rate of 2.5mm/min until failure. Only maximum load was recorded. 2.3 Failures or Fractures Failures were classified by visually examining the fracture pattern of the failed specimen. Common failures are shown in Figure 3. Figure 3a illustrates a clean fracture of the specimen and was the most common failure mode for wood specimens. A less common wood specimen failure where the fracture surface runs perpendicular to the growth rings is illustrated in Figure 3b. A typical failure for LVL is illustrated in Figure 3c, where the fracture is jagged and is located at different positions in each veneer. Two typical failures for WPCs are illustrated in Figures 3d and 3e. The failure in Figure 3d follows the stranding pattern of the composite and the failure in Figure 3e initiates more in the edges of the holes for the dog bone shape after the specimen had deformed in the grips. Failure was judged to occur when a steady decrease in load for at least five seconds was observed and splitting occurred (with or without complete fracture). Fracture surfaces could be classified according to the material being tested. Typical fracture surfaces for solid wood either followed the path of the tree rings or cut cross them fairly perpendicular to the rings. LVL fracture surfaces typically occurred through the center of the specimen following a jagged pathway as illustrated in Figure 3c. Typical WPC specimens also failed through the center of the specimen at the hole quadrants as illustrated in Figure 3d. However it was not uncommon to find vertical fractures forming simultaneously from the opposite upper or lower corner as illustrated in Figure 3e. This latter failure mechanism often resulted from the specimen being deformed by the test so that the holes no longer were circular and the result is a localized strain gradient in the location of the grips causing the specimen to rotate slightly during the test. (a) Wood Parallel to Growth Rings (b) Wood Perpendicular to Growth Rings
(c) LVL Perpendicular to Veneers (d) WPC Parallel to Extrusion Direction (e) WPC After Deformation in Grips Figure 3 Typical Failures Observed During Tests for Wood, LVL, and WPC 2.4 Data Recorded or Calculated In addition to the maximum load, the width and thickness of the specimen were measured and recorded at the minimum tensile area. Once the specimens were tested, the moisture content was determined. A small section was cut from each specimen approximately 25x25mm by the thickness of the specimen, similar to the illustration in Figure 4. The moisture content was determined following ASTM D 4442-92. The dry section was then given a thin coating of wax and the volume was determined following ASTM D 2395-93. The moisture content and density values reported are based on the oven-dry weights and volume. The moisture content of the wood-plastic composite was not determined for these tests. 3. Results Summary statistics for the strength, moisture content, and specific gravity of the four material types are presented in Table 1. As shown, the LVL strength is considerably lower (ca. 42-47%) than that for Douglas-fir Larch lumber. One possible source of the lower strength of the LVL is that the veneers used in the manufacture of the product possess lath checks, which allow the veneer to split easily. While the veneers Figure 4. Section for Determining Moisture Content and Specific Gravity.
might distribute the different grain patterns and other flaws throughout the material, the lathe checks are frequent enough to result in the slightly jagged failure pattern observed in the tests. The WPC was the by far the strongest material, and had a strength 249-274% of the value for Douglas-fir Larch lumber. The differences in the strengths of the two samples of wood are likely due to the differences in the grading rules for the two different szes. Table 1 Mean, and coefficient of variation for wood, LVL, and WPCs Material Number of Specimens Tested Wood (38x) 146 Wood (89x) 132 LVL 69 WPC 33 Mean Tensile Strength, μ (kpa) [COV, %] 2074 [31.3] 1878 [32.8] 874 [31.6] 5155 [16.7] MC (%) [COV, %] SG [COV, %] 14.4 [20.8] 0.46 [11.5] 14.9 0.45 [5.4] [14.7] 12.5 0.55 [13.5] [6.2] N/A 1.0 [0.5] Despite the high specific gravity, or density, of the LVL specimens when compared to that of the wood specimens, a decrased tension strength was observed. This fact illustrates why the nailing of diaphragms and other structural members using LVL should become a critical concern of designers. The potential for splitting of the material is significantly higher than wood and therefore, should have requirements determined for the minimum nail spacing before staggered patterns are required. This determination would be similar to the requirement that 64 mm lumber be used for nail schedules that require 64 mm or closer nail spacing for shear walls and diaphragms. The high value for WPCs also provides the information necessary to understand why braced wall panels tested by DuChateau (2005) were up to three times as strong as braced wall panels using traditional wood framing for the sill plate. The uplift of the end studs produce tension perpendicular to grain stresses in the sill plate, and the higher strength of the WPC was able to be utilized in increasing the racking strength of the wall assembly. 4. Conclusion The results of the study confirm that tension perpendicular to grain in wood, LVL, and WPCs were not equal. These differences could be responsible for the differences in the structural response of building components. When compared to the traditional benchmark of solid sawn lumber, LVL is weaker and poly(propylene)-based WPCs are stronger. This information should be incorporated into the design methodology for connections made with these materials. The minimum spacing requirements for fasteners in LVL should be increased over those required for lumber and decreased for fasteners in WPCs when splitting of the material is of concern. The strength values should also be incorporated directly into a design methodology that is based on connection capacity, rather than yield theory, so that the designer is aware of the true overstrength of the structure, and the process becomes transparent to the designer. Finally, these results indicate that engineered products need to be evaluated on all of their mechanical properties to insure that all aspects of their performance are designed in a safe manner. An example is the modification of the ICC acceptance criteria AC14 (ICC, 2004) requiring diaphragm tests for I-joists with thin flange material. However, if the base connection design were
to be changed to account for all of the effects of the mechanical properties, the expensive tests associated with full scale structural testing could be minimized. References [1] American Society of Testing and Materials (2000). D143-94 Standard Test Methods for Small Clear Specimens of Timber. ASTM Annual Book of Standards Vol. 4.10. ASTM, West Conshohocken, PA. [2] American Society of Testing and Materials (2000). D2395-93 Standard Test Methods for Specific Gravity of Wood and Wood-Based Materials. ASTM Annual Book of Standards Vol. 4.10. ASTM, West Conshohocken, PA. [3] American Society of Testing and Materials (2000). D4442-92 Standard Test Methods for Direct Moisture Content Measurement of Wood and Wood-Based Materials. ASTM Annual Book of Standards Vol. 4.10. ASTM, West Conshohocken, PA. [2] DuChateau, K. (2005). Structural Design and Performance of Composite Wall-Foundation Connector Elements. Thesis submitted in partial fulfillment of requirements of Master of Science degree in Civil Engineering at Washington State University, Pullman, WA. [3] International Code Council Evaluation Service, 2004. AC 14: Acceptance Criteria for Prefabricated Wood J-Joists. ICC-ES, Whittier, CA, http://icces.org/criteria/pdf_files/ac14.pdf. [4] Kundson, C. (2006). Effect of Row Spacing on the Cyclic Performance of Multiple-Bolt Connections in Wood. Thesis to be submitted in partial fulfillment of requirements of Master of Science degree in Civil Engineering at Washington State University, Pullman, WA. [5] Laver, T.C. (1996). Extruded Synthetic Wood Composition and Methods for Making Same. US Patent Number 5,516,472.