Date of Award

12-2005

Level of Access

Campus-Only Thesis

Degree Name

Master of Science (MS)

Department

Civil Engineering

Advisor

Eric N. Landis

Second Committee Member

William G. Davids

Third Committee Member

Stephen M. Shaler

Abstract

To properly model the complex behavior and internal geometry of wood, traditional continuum-based approaches were abandoned and a lattice model was developed based directly on the morphology of wood. The lattice model is essentially a network of springs that have a variable strength and stiffness used to predict the response of a material. Generally lattice models have been used to model the response of concrete and ceramics. Recent research, however, has shown that a morphologically-based lattice model of wood can produce good predictions of bulk elastic constants, load-displacement response, and fracture patterns. Although previous lattice models of wood produced good response predictions they did not fully incorporate the natural morphology of wood. In developing of lattice model we established a reasonable lattice size, level of refinement, and included the natural variations found in wood. These natural variations tied the morphology seen in wood directly to the morphology in the lattice model. Grain perturbation and growth ring characteristics were measured in red spruce specimens and were directly incorporated into the lattice model. Grain angle was also included to mimic growth ring geometry. Experimental testing of small wood specimens subjected to three stress states, longitudinal tension, radial tension, and RL shear, was conducted. The experimental data was analyzed and the data prepared for use in the calibration of the lattice model element properties. Calibration of the lattice model was broken into two steps, stiffness calibration and strength calibration. Stiffness calibration consisted of performing elastic analyses in the three stress states tested and matching the bulk elastic constants to those found in the experimental testing. Strength calibration was completed by matching the full stress displacement response of the lattice model to the experimental responses. The goal of the morphological based lattice modeling is to not only capture response of wood, as other models have, but to also capture its variability. The fully calibrated model predicted strengths 8.9% low, 22.2% high, and 7.1 % low for longitudinal tension, radial tension, and RL shear respectively. Model and experimental strength variability was 12.9% and 13.4%, 7.6% and 27.7%, and 9.37% and 17.73% for longitudinal tension, radial tension, and RL shear respectively. Fracture in the model accurately resembled fracture in experimental testing. Verification of the lattice model in a different stress state was performed. The fracture path and stress-displacement response of the lattice model was consistent with the experimental testing. This verification indicates that the lattice model may be used to model stress states other than those that were used to calibrate the model. It is recommended that future work be performed with three dimensional lattice models to predict the behavior of wood. By adding a third dimension, lattice models may better capture the variation observed in real wood.

Share