The strength of bonded wood strand composites

A method for modelling the strength of bonded wood strands which are oriented principally in one direction is proposed. The hypothesis tested was that strand grain orientation data, fitted to a von Mises probability distribution, could be used in an analysis for estimating the potential tensile strength of an ideally bonded composite. The strand strength, resolved at any loading angle with respect to the principal composite strand orientation axis, was multiplied by the distribution probability at that angle. When integrated over all angles, this product yielded the mathematical expectation of strength for the composite. The model predicted composite strength at off-orientation axis angles and represented the material in two dimensions in an orthotropic fashion. A feature of this research is the use of a parametrically quantified strand orientation level in an algorithm developed to estimate composite strength. A practical number of strand angle readings (100) were taken to characterize each composite. These angle readings defined orientation in terms of a parameter which described composites ranging from random to highly oriented. The model input also required microtensile strength means from samples of strands tested in the longitudinal and radial or tangential directions. Comparisons between the model and actual specific strengths were made at five equally spaced-composite principal axis load angles from 0 to 90 degrees. Both tensile and flexural tests were performed to evaluate the model. The evaluations were designated in terms of resin content, distribution, and droplet size. These variables were studied using colorimetry and computerized image analysis. Composite density profiles through the specimens' thickness were obtained from direct reading x-ray densitometry. Composites made of juvenile trembling aspen, red alder, red cedar, mature lodgepole pine and yellow birch were studied. Assumptions concerning wood shear strength and strand length/thickness ratio were discussed in the interpretation of an overlapping strand stress-transfer model. This led to the definition of failure criteria based on stress transfer. A trial of orientation modelling in elasticity estimation was made and a random function model of composite elasticity based on laminated plate theory is outlined in a supplementary proposal for further research. The simplified algorithm for the strength of aligned.wood strand composites provides design targets for reconstituted high strength strand lumber and panel products of the future.Forestry, Faculty ofGraduat