On the fracture behaviour of softwood : a combined experimental and computational study of the interaction between fracture behaviour and material anatomy

Sammanfattning: The propensity for fracture, i.e. failure through separation of the material, is a factor that must be regarded in the design of wooden structures, but also in wood machining operations and drying. In general, wood has a very complex fracture behaviour, and much of the complexity is due to the anisotropic and heterogeneous structure of the material. This thesis deals with the fracture behaviour of softwood, and in particular with the influence of the softwood fibre- and growth ring structures on fracture. The considered materials are spruce (Picea abies) and pine (Pinus sylvestris) in air-dry condition, and the fracture behaviour is studied both experimentally and analytically within the framework of linear elastic fracture mechanics. The fracture behaviour is first studied with respect to the softwood fibre structure and the anisotropy in elasticity and strength that it entails. Conditions of mixed mode I/II loading are considered, and the predictability of three mixed mode fracture criteria is tested on cracks oriented along the wood fibres (RL) and across the fibres (LR). A fracture criterion based on critical principal stress is found to be superior to criteria based on critical energy release rate and critical strain energy density. The criterion, which contains a single fracture parameter, works well for cracks oriented along the fibres. However, for cross-fibre cracks, the applicability is confined to cases in which the near tip T-stress, i.e. the non-singular stress acting parallel to the crack plane, is low. Secondly, the fracture behaviour is studied with respect to the softwood growth ring structure. The radial variation in stiffness from earlywood to latewood in the growth rings is found to strongly affect the behaviour of radially growing cracks (TR), loaded in mode I. Finite element analyses, in which the growth rings are represented by a repetitive stiffness gradient, show that the distribution of stress around the crack tip is very different from crack tip stress fields in homogeneous materials. The results indicate that fracture criteria formulated on the basis of critical stress intensity are unsuitable for cracks in the TR crack system, since the fracture toughness in this system cannot be experimentally determined without ambiguity. As a complement to the study on the effect of growth ring structure on fracture behaviour, a measurement of the radial stiffness variation across growth rings in air-dry spruce is presented. The radial and tangential deformations in a single growth ring, which is subjected to moderate tensile loading in its radial direction, are measured by use of digital speckle photography. From the measured displacement field, the transverse coefficients of elasticity are determined with respect to radial position in the growth ring by use of inverse finite element modelling. The tangential elastic modulus is found to be a factor 15 larger in latewood than in earlywood, whereas the corresponding variation for the radial modulus is merely a factor 3. The measured variations are in reasonable agreement with predictions from simple two-dimensional honeycomb models of the growth ring cellular structure.