No abstract

Differences in fracture patterns and properties at various moisture levels are experimentally and numerically evaluated and discussed. Experiments were performed on spruce, pine, oak and beech. The influence of moisture at 98%, 80%, 65% and 30% RH and the mechanisms involved were investigated for softwoods and hardwoods subjected to opening mode I fracture using in situ and ex situ real-time environmental scanning electron microscopy (ESEM). The wedge-splitting technique was employed. To quantify the effect of humidity, fracture toughness values were obtained from ex situ tests and finite element analysis of the contact problem in wedge-splitting. In addition, lattice fracture model simulations were performed for numerical investigation of the fracture mechanisms. Distinct changes in wood fracture behaviour were observed as a function of moisture content. Fracture toughness was highest at 30% RH for all species except for oak, and showed higher values in the radial-longitudinal than in the tangential-radial direction. In green wood, water droplets moved away from the cell lumens around the crack tip. Drying of wood promotes microcracking and crack bridging as toughening mechanisms. The findings reported may be useful for further research into the interaction between moisture transfer and stress gradients in wood accompanied by moisture-crack phenomena.

No abstract

ABSTRACT--In this paper, we explore ways to couple experimental measurements with the numerical simulations of the mechanical properties of wood. For our numerical simulations, we have adopted a lattice approach, where wood fibers or bundles of wood fibers are modeled as discrete structural elements connected by a lattice of spring elements. Element strength and stiffness properties are determined from bulk material properties. Damage is represented by broken lattice elements, which cause both stiffness and strength degradation. The modeling approach was applied to small specimens of spruce subjected to transverse uniaxial tension, and mode I transverse splitting. The model was found to be good at predicting the load-deformation response of both notched and unnotched specimens, including the post-peak softening response. In addition, the damage patterns predicted by the model are consistent with those observed in the experiments.

Challenges arise in finite element (FE) analyses that predict mechanical failure in wood-based materials because their structural complexity is difficult to mimic. When considered at the macro scale, wood and engineered wood composites can reasonably be assumed to behave as homogenous continua. However, accurate meso- and micro-scale representations require a different approach. Models employing discrete FEs are robust tools for detailed failure analysis, because the elements can be made to mimic the functions of morphological structures in the material. Hybrid models that meld continuum and discrete FEs also show good promise as generalised analysis tools, but as yet their development is in its infancy. In the future, beyond mechanical damage, other energy sinks also need to be included in models, and computational efficiency should be improved. In this overview, the advantages and limitations of alternative FE representations are demonstrated in terms of failure processes in wood-based materials via case analyses.