Experimental and Numerical Determination of the Thermal Conductivity Tensor for Composites Manufacturing Simulation

Abstract

The in-plane and through-the-thickness thermal conductivities are fundamental input parameters for cure simulations of composite manufacturing processes such as vacuum assisted transfer molding (VARTM). The modeling of these conductivities is challenging due to the inherent anisotropy of composite materials, difficulty of the measurements during the cure process, and availability of the equipment. In general, cure simulations consider the homogenized thermal conductivities at lamina level and hence, micromechanics is often used to connect the individual matrix and fiber properties with the homogenized lamina properties based on microstructure, fiber volume fraction, and void content. This work presents the experimental determination and multiscale modeling of the thermal conductivity tensor of a textile VARTM composite as a function of the thermoset cure. Mechanics of structure genome (MSG) is used to compute the equivalent thermal conductivity tensor of the lamina as a function of the thermoset degree of cure. Two-step homogenization is iteratively performed coupling the evolution of the resin thermal conductivity with SwiftCompTM, a general-purpose multiscale constitutive modeling code based on the MSG. The dynamic response of the thermoset resin is measured by means of differential scanning calorimetry to characterize the cure kinetics and thermal conductivity. In addition, an in-house measuring apparatus is used to determine the through-the-thickness thermal conductivities of the dry fabric and cured composite part. The predicted through-the-thickness thermal conductivities of the cured composite are compared against experimental data. The MSG proved to be an efficient method to predict the evolution of thermal conductivity during curing process, and hence, helped bridge the gap by providing data that could not be measured directly with the present experiment.