Characterizing wood-plastic composites via data-driven methodologies

Abstract

The recent increase of wood-plastic composite materials in various application areas has underlined the need for an efficient and robust methodology to characterize their nonlinear anisotropic constitutive behavior. In addition, the multiplicity of various loading conditions in structures utilizing these materials further increases the need for a characterization methodology that accounts for this complexity. In responding to these needs, this paper presents a data-driven methodology that utilizes multi-degree of freedom mechatronic loading systems to systematically collect behavioral data of specimens constructed from these materials. These data represent force, stiffness, and dissipated energy versus displacement data for each specimen tested. The highly nonlinear behavior of these specimens was confirmed and is discussed in the context ofthe associated damage strain-induced micro-cracking. This methodology further utilizes the acquired data to construct an analytically defined model of the observed behavior in terms of identifying the total (recoverable and irrecoverable) energy density spent for deforming and damaging these materials. A constrained optimization formulation leads to the determination of the unknown parameters defining this methodology. Subsequently, the identified constitutive behavior in terms of the dissipated energy density is utilized to predict load-displacement behavior of specimens tested outside the context of the automated loading frames utilized for the characterization. Validation discussion between the predicted behavior and the experimentally established behavior conclude this work.

Citation

Michopoulos, John G.; Hermanson, John C.; Badaliance, Robert. 2007. Characterizing wood-plastic composites via data-driven methodologies. 9th International Conference on Wood & Biofiber Plastic Composites : May 21-23, 2007 ... Madison, Wisconsin, USA. Madison, WI : Forest Products Society, 2007: Pages 145-151