Study of the elastic behaviour of wood–plastic composites at cold temperatures using artificial neural networks

Abstract

Using wood as natural reinforcement in composite materials, instead of mineral reinforcements, has several advantages such as low density, low cost, and less abrasive finish. The natural reinforcements also are non-toxic and recyclable. The general aim of this work is to provide a novel approach in material characterization and, in the future, to apply this method to investigate the elastic, hyperelastic, and viscoelastic behaviour of wood–plastic composites (WPCs) for potential applications in several industries. In this research, the bubble inflation technique is used to study the elastic behaviour of bio-composite materials at cold temperatures. The results of experiments with high-density polyethylene (HDPE) and WPC membranes with 30 and 60 wt% of wood fibre, under the combined effect of temperature and pressure, are presented. A unique approach that uses the results of computer simulations in Abaqus, a finite element package, combined with artificial neural networks is applied in order to find the optimum material constants for different materials at cold temperatures. The values of Young’s modulus for HDPE and WPCs with 30 and 60 wt% of wood fibre are reported at different temperatures.