Numerical Optimization of Vacuum Assisted Manufacturing of Aircraft Composite Parts Using the Predictive Assessment of Objectives

Abstract

The vacuum assisted technologies of composite parts manufacturing are becoming very popular owing to their relative simplicity and inexpensive equipment. They are implemented in the production of aircrafts, rotorcrafts, ships, cars and wind turbine blades. Unfortunately, in many cases this technology does not provide sufficient repeatability and quality due to very high process sensitivity to its layout (number, dimensions and positions of the resin gates and vacuum vents) and the temperature. Due to the fact that process testing is very expensive and time-consuming, the use of vacuum infusion process simulation is relevant. Such modeling tools must be able to correctly predict resin spreading in the preform and identify potential defects. In contrast to the well-known software tools, we propose to use the phase field equation as the leading one and providing a more accurate reconstruction of the front of the moving resin. To describe the pressure distribution in an incompletely saturated porous preform, we use the Richards equation, and the state of the moving resin is modeled by the convection-diffusion equation with a source controlled by the autocatalytic equation. The formulation of a forward coupled modeling problem also includes modified relationships for the dependence of the resin viscosity on the degree of cure and temperature, as well as for the diffusion coefficient of the degree of cure when mixing streams of liquid resin. The article presents the results of modeling the vacuum infusion of a 3D preform at varying positions of the vacuum vent, high-permeability medium (HPM) tape and temperature. It is shown that the indicators of the quality and performance of the process at the moment of stopping of the resin spreading due to beginning of its gelation can be predicted at an early stage of simulation, which is very important for using the developed CAE model for the process optimization. The validity and reliability of such predictive estimate using the objectives and their combination, which are determined at an early stage of simulation, are discussed.