Theory and numerical implementation for modelling viscoelastic properties of wood including effects of temperature, moisture content, and mechanosorption
Please login to view abstract download link
Viscoelastic creep and stress relaxation in wood are affected not only by temperature and moisture content but also by rate of change in moisture content during natural exposure and experiments. This moisture-change-rate effect is commonly referred to as mechanosorption. Experiments indicate that creep and relaxation accelerate whether the moisture content is increasing or decreasing. This effect has previously been modelled using analogues of elementary time-dependent viscoelastic creep, with time variable substituted by cumulative moisture change, to be solved in parallel with the time-dependent effects (e.g. [1-5]). This talk presents an alternate approach with numerical implementation for extending superposition methods in linear viscoelasticity to include the mechanosorptive effects. The relaxation time is modelled using a total shift factor a_tot = a_T a_c a_dc that is a product of three component shift factors: a_T, commonly used for time-temperature superposition methods in polymer viscoelasticity modelling, a_c is sometimes proposed to account for moisture content in superposition analyses for wood, and a_dc, is specific to this new approach, proposed to derive time-temperature-moisture-moisture-change-rate superposition tools for modeling viscoelasticity of wood including mechanosorption effects. The new approach with combined shift factors was implemented in a material point method (MPM) model for anisotropic, viscoelastic materials like wood. Here, wood was considered transversely isotropic to capture differences in creep (or relaxation) parallel or perpendicular to the grain. Extension to complete orthotropic formulation is straight forward, but requires more material parameters for the second transverse direction, not available from current experimental results. The MPM model was compared to experimental results [6]. The a_tot factors shift creep (or relaxation) curves horizontally along the time axis, but that was not enough to fit experiments. To account for temperature and moisture effects on initial mechanical properties, the modeled was extended to include vertical shift factors. To capture details of experiments, the viscoelastic properties needed to include more than one exponential relaxation time.
