Abstract
Shear strength is a wood property which is fundamental to the design of wood-based products and constructions. This property cannot be predicted at present for lack of sufficient knowledge, mainly because of the large number of timber species that occur in the Congo Basin. The main aim of this study was to provide a preliminary qualification of shearing in Congo Basin timber species, with consideration for its variability. For this purpose, we studied 12 timber species with very different properties, from the least dense to the densest. Their shear strength was determined experimentally using European standards specifications, on the scale of the wood material used. A statistical analysis was conducted. To reduce shear strength variability, the species were assigned to four distinct clusters defined according to FCBA Institute specifications. With a view to developing allowable design stresses to facilitate decision-making, we evaluated the relative goodness-of-fit of five probabilistic shear strength distributions (normal, lognormal, exponential, Weibull 2 parameters and Weibull 3 parameters) that are used in wood-related applications. The results of geometric regression (R2 = 0.81) show that shear strength is well correlated with density. Shear strength can be more reliably predicted with the three-parameter Weibull distribution than with the other distributions. The findings of this study open up new prospects to be considered for the design of wood-based products with regard to shear, when using tropical timber species from the Congo Basin.