Internal stresses develop in wood during moisture conditioning because of its hygroscopic properties and unequal shrinkage behaviour in the three grain directions. However, such stresses can be inherently difficult to detect. External wood distortion usually indicates the presence of internal stress but all stress may have been relieved through such distortion. Conversely, wood, which shows no external distortion, may still exhibit an internal stress imbalance which can lead to distortion after sawing. Stress effects associated with kiln drying can result in significant sawn lumber product degrade as a result of warping and checking.

Two methods exist to assess internal stress in wood as a result of moisture change. The first method is direct measurement, as with prong and slices tests, and the second method is to model internal stress. Direct measurement assessments are usually unable to be performed during moisture conditioning and are invasive in that they lead to sample destruction. Conversely stress modeling is predictive and requires fundamental wood material properties and complex software to solve the mechanical equations in space and time.

A work programme was initiated at Scion, New Zealand, with the overall goal of collecting high temperature (70 to 150ºC) material properties for use in modeling the stress developed in radiata pine during high temperature kiln drying. A key focus was also to design and develop equipment to accurately measure fundamental material properties required for such modeling. The study was designed to extend an earlier three dimensional distortion model developed by Ormarsson (2000) to predict drying stress in wood up to a temperature of 70°C where the total strain equation was assumed to consist of the summation of elastic, moisture-induced shrinkage, mechanosorptive and creep deformations. Each component of the total strain equation was studied separately with respect to the variables stress, time, temperature and moisture content using purpose designed equipment and the results combined in a finite element stress model.

Experimental material property results for tangential grain, tensile, radiata pine (Pinus radiata D.Don) sapwood samples were combined into an instantaneous distortion model to predict stress or distortion as a function of temperature, moisture content and either external load or distortion. All samples were taken from one representative plantation tree from which wood quality information was obtained for a covariate analysis of the results. This included basic wood density, spiral grain, conic angle, and cardinal point mean microfibril angle which was obtained using SilviScan. Poissons' ratios were estimated from digital speckle photography experiments perfomed using an Aramis system at 20°C whilst the remaining compliance matrix variables such as shear moduli and moduli of elasticity for longitudinal and radial grain directions were estimated from published ratios.

The distortion model architecture involved Python code for use as an input into Abaqus software with Fortran subroutines. For each study a moisture field output file, based upon the chosen constant temperature, was obtained for each geometric sample element and the results used as an input into the stress model. Stress model results were compared with representative experimental results from instantaneous, free shrinkage and mechanosorption studies.

Results showed good agreement between the model and instantaneous elastic, free shrinkage and mechanosorption experimental results for tensile sample geometries, provided a 'thermo-sorptive-aging' effect was used to compensate for reduced hygroscopicity, if wood was alternately wet and re-wet in a cyclic fashion. A full board distortion model was then used to test the resultant stresses associated with wood that was dried to 5% moisture content followed by steaming to simulate final stress relief as implemented by industry. As expected, full board model results revealed the average stress in the longitudinal, radial and tangential directions to decrease after steaming, compared to drying with no steaming. Even though internal stresses from modelling were higher after steaming compared to only drying, total stress was improved. This theoretically caused less external distortion in a simple geometry like a board, although creep was not included in the simulations and may help to equalise internal stress further over time. Stress result profiles within the board agreed with earlier studies for two dimensional distortion by Chen et al (1997b).

The three dimensional high temperature distortion model that was developed as part of this work is a valuable tool for distortion assessment compared to one- or two-dimensional models that are limited to a single plane. Future work will seek to optimise moisture conditioning schedules and thermo-hygro-mechanical treatments using the full board distortion model with full mechano-sorptive-creep analysis.