The objectives of this project were to develop and validate a finite element (FE) model of the hygroscopic warping of OSB panels and to suggest panel structures that ensure a higher level of stability during the storage, handling and use of panels.
A modification of the methodology developed at Université Laval for solid wood and already applied to MDF panels at Forintek Canada Corp. was used for the determination of diffusion coefficients in OSB. An existing finite element model developed jointly by Forintek Canada Corp. and Université Laval for the evaluation of MDF warping was adapted to the characteristic OSB structure. The finite element model was based on an unsteady-state moisture transfer equation, a mechanical equilibrium equation, and an elastic constitutive law. The experimental inputs were the mechanical properties E1, E2, E3, G12, G13 and G23 all as a function of moisture content, density and strand orientation; the expansion properties b1, b2 and b3 as a function of density and alignment; sorption isotherms and diffusion coefficient generated by producing a total of 50 laboratory OSB panels: 18 one-layer panels without density profile and strands oriented through the entire thickness (6 x 500 kg/m³, 6 x 625 kg/m³, 6 x 800 kg/m³), 24 three-layer panels with density profile (16 panels with density 625 kg/m³, aligned strands in surface, random strands in core, and 8 panels with density 625 kg/m³, different alignment in the two surface layers, random strands in core) and 8 panels with density 625 kg/m³, random strands. The panels had dimensions after trimming of 838 mm x 838 mm x 10.5 mm (33 in x 33 in x 7/16 in). To validate the model, warp was initiated and its dynamics was monitored by submitting 2 panels from each group to an 80% relative humidity.
The results showed that for all one-side sealed panels, the MC-increase in the zones close to the surface at the early exposure stages caused rapidly a convex deformation towards the exposed surface. When MC gradually homogenized across thickness, most of the panels returned close to their original flat shape. For panels with a flat density profile, the higher the average panel density, the higher the level of warp due to the effect of density on the expansion and swelling properties. Panels with oriented strands experienced higher strain differential and therefore developed stronger warp compared to panels with random strands. Panels with a one-layer structure experienced higher warp compared to panels with a three-layer structure. When the sealed surface layer was thicker than the exposed surface layer, or when the alignment in the sealed layer was higher than the alignment of the exposed layer, the panels continued to distort and their warp became negative, instead of stabilizing close to their original flat form.
The agreement between the experimental results and the finite element results confirmed the validity of the proposed model in the conditions and the OSB properties considered in this work. Simulations with the finite element method were performed corresponding to specific industrial applications and allowed the creation of a large database of results, which served for building the software package WarpExpert.
A group of 2x4 SPF samples was tested for bending stiffness in the Western laboratory of Forintek and then re-tested in the Eastern laboratory . Another group of 2x4 SPF samples was tested for bending stiffness in the Eastern laboratory and then re-tested in the Western laboratory. The bending stiffness tests were conducted on test machines set up in accordance with ASTM Standard D198-02. Additional bending tests were done according to ASTM D4761-02A using the “portable bending” machine in the Western laboratory and a modified Metriguard 312 bending machine in the Eastern laboratory.
Results from ASTM D198-02 bending stiffness tests showed a differences between the laboratories of 2.1% for the sample originating from the Western Laboratory and 1.5% for the sample originating from the Eastern Laboratory. The MOE bending test results were not adjusted to account for any increase or decrease in the moisture content of the specimens.
The wood industry is facing some serious challenges in how end users view the long-term reliability of wood construction systems. The 1990s have seen the industry hit with a series of high-profile wood product failures due to decay, for example in North Carolina and coastal British Columbia. There are several efforts underway in North America and around the world focused on developing predictive models for moisture conditions in exterior wall systems. All of these models can predict temperature and wood moisture content change over time, but the consequences of those conditions in terms of decay are not yet predictable. While it is known that wood below 20% moisture content will not decay and wood above 28% moisture content will decay, fungal response to conditions between 20 and 28% is not well documented, particularly for North American fungi and wood species. Forintek Canada Corp. has a project underway to determine the time required for wood products to suffer detectable strength loss under a variety of temperature and moisture conditions. The focus is on sheathing as it is the last place to dry out after wetting events. Since this project was initiated, other researchers have become involved in this issue and it is therefore timely to review the state of the knowledge in this area. There is a considerable volume of work published and a limited amount of work underway but little of this is directly relevant to developing damage functions for hygrothermal models. The work underway at Forintek needs to be completed to define the time to initiation of decay under constant moisture conditions. Further work needs to be done to define the time to initiation of decay under fluctuating conditions. Data on the initial rate of decay under limiting conditions should also be generated from this work.