UNIMAT: An enhanced forming simulation model of prepreg woven fabrics, with application to process optimization for wrinkle mitigation

Abstract

Processing simulation of prepreg fabrics requires considering multiple interactive deformation mechanisms to reliably predict the response of a formed part. However, these mechanisms, especially the evolving fabric properties and their interactions, are often overlooked. In this study, through integration of a series of user-defined subroutines, a unified (enhanced) numerical model (called UNIMAT) is developed to simulate the forming of a plain-weave fiberglass/polypropylene prepreg. The model specifically involves simultaneous incorporation of the fabric nonlinear in-plane and out-of-plane behaviours (including local bending/reverse bending effect with hysteresis), the ply viscous behavior at room temperature, and inter-ply anisotropic friction as a function of the ply orientation, pressure, and slippage. UNIMAT is first validated with a benchmark hemisphere forming test, and is then used for process optimization to minimize wrinkle formation over a complex shape tool under a vacuum bagging process. The model accurately predicted the forming outcomes including the part topology, punch force, shear angles, and wrinkles’ overall severity. The optimization search, through a Convolutional Matrix Adaptation Evolution Strategy (CMA-ES) algorithm, demonstrated that the wrinkles state can be diminished by up to 30% if local constraints on the fabric boundaries are optimally applied using pressure risers (modifiers), prior to the start of the vacuum.