Hierarchical modelling of softwood hygro-elastic properties
Thesis DisciplineMechanical Engineering
Degree GrantorUniversity of Canterbury
Degree NameDoctor of Philosophy
The hygro-elastic behaviour of wood under load or when subjected to environmental changes is of considerable practical interest. This behaviour can be determined by exhaustive experimentation, but such an approach makes explaining its origin, which from some perspectives can be more important than its prediction, problematic. This thesis attempts to establish a model (actually a hierarchical set of models) that goes some way toward both predicting and explaining the mechanics of wood. Attention is focused on radiata pine because of the commercial importance of this species in New Zealand, but much of the modelling is applicable to other softwoods and, to a lesser extent, hardwoods. Wood can be looked on as a hierarchical material, that is as a material possessing structure at multiple scales. For many problems involving such materials the heterogeneous structure at a particular scale can be replaced by a homogeneous one possessing similar properties. Homogenization theory defines what is meant by similar and also provides the means for determining these effective properties. In this thesis wood structure is treated at three different scales: namely the supramolecular or nanostructural, the cell-wall or ultrastructural and the cellular or microstructural scales. Homogenization across these levels is performed either analytically or numerically, using the finite element method. At the smallest scale, the constituent phases are treated as homogeneous continua. Models for the hygro-elastic phase properties, as functions of temperature and moisture content are developed based on available experimental data. The models devised to describe wood at each of the above mentioned scales introduce a large number of structural parameters, such as constituent mass fractions and cell-wall layer volume fractions. In the abscence of specific data, estimates for these parameters are developed based on data from the literature. Together with these auxiliary models, the main sequence of structural models can then be used to obtain estimates for the material properties of small domains within macrostuctural models.