An Integrated Numerical Model for Predicting Flow and Compaction of 2D Textile Composite in the VARTM Process

Abstract

The vacuum assisted resin transfer molding (VARTM) technique is a low-cost alternative to the autoclave process suitable for manufacturing large-scale composite parts. During the VARTM process, carbon fiber sheets are covered by a vacuum bag, and a vacuum pump is utilized to assist the resin flow in the infiltration process. However, when dealing with carbon fiber fabrics with complicated microstructure such as 2D textile fabrics, the gaps between the fiber tows are sometimes not fully saturated by the resin flow, creating starvation areas that potentially compromise the mechanical performance of the composite part. In addition, during the resin infusion process, the cross-linking reaction is activated under the preheated condition, making the material properties of resin both temperature and degree of cure (DOC) dependent. Therefore, the key is to establish a multi-physics model that incorporate heat transfer analysis and cure kinetics into the flow compaction model. In this paper, an integrated flow compaction model is developed to predict the flow front advancement inside the textile fabrics during the VARTM process. The flow velocity and pressure distribution inside the resin flow are governed by the continuity equation and Darcy’s law. Due to the fact that the behavior of the resin flow is influenced by the local permeability, which in return is a function of fluid pressure, the flow model and the compaction model are coupled, and they are solved simultaneously to determine the flow behavior of resin inside the fiber fabrics. In the meantime, the viscosity model of the resin is incorporated into the flow compaction model through the material properties of the resin to account for the effects of infusion parameters. Using this flow compaction model, the permeability of the carbon fabrics used in this study is determined by comparing the flow front predicted by the flow compaction model to that observed in the radial injection experiment. Moreover, once the fluid distribution is determined, the local fiber volume fraction and thickness variation of