Modeling the behavior of wood-based composite sheathing under hygrothermal load
Lang, Elemer M.
MetadataShow full item record
In light-frame residential construction wood-based composite panels used externally or internally are exposed to relative humidity and/or temperature changes. The subsequent moisture content change of the panels will result in two types of deformations as follows: 1. elastic deformation of the panel due to the constraint, 2. warpage of the panel due to the unbalanced expansion of the layers. Such deformations can cause unacceptable serviceability problems in light-frame wood construction. A model was developed to predict quantitatively the global deflection of wood-based composite panels exposed to relative humidity changes. The model was based on the constitutive relationship of the Classical Lamination Theory and the thermal stress analogy in composites. As an alternative solution, the applicability of the eccentrically loaded column formula was evaluated. The developed models were experimentally validated for OSB and plywood sheathing. Test variables included the panel type, exposure (symmetric and non-symmetric MC gradient) and specimen configuration (single span, multiple span). The comparison of measured and predicted deflections are presented. The important elastic and hygroscopic material properties were acquired through testing. Statistical analyses of test results are discussed. The uncertainty analysis was used to make statistical inference comparing the means of measured deflection to the uncertainty interval of predictions. Good agreement between predicted and measured deflections was found for single span test structures. Also, for double span structures the models predicted the experimental response fairly veil. Uncertainties in Me measurements made the prediction less reliable when symmetric moisture content gradient developed during the exposure. Due to its lower variability in material properties, the response of OSB sheathing to moisture content changes is more predictable.
- Doctoral Dissertations