Viscoelastic modeling of fiber preform compaction in vacuum infusion process

Abstract

A woven fabric's compaction was modeled by using five viscoelastic models - Maxwell, Kelvin-Voigt, Zener, Burgers, and Generalized Maxwell - to reveal the capabilities and limitations of the models. The model parameters were optimized by minimizing the deviation between the model results and experimental data collected in our previous material characterization study mimicking different compaction stages (loading, fiber settling, wetting, unloading, and fiber relaxation) that a fiber structure undergoes during vacuum infusion process. Although Burgers and Generalized Maxwell models have the highest performance due to their almost equal coefficient of determination values, they have diverse characteristics in terms of modeling different stages of compaction. Burgers model allowed modeling the permanent deformation in relaxation stage, but failed in modeling permanent deformation in settling stage. Generalized Maxwell model could do the opposite, i.e. failed in the former and could handle the latter. This study's major contribution is a holistic numerical approach and its conclusions by modeling all stages of the vacuum infusion process instead of one stage at a time, and thus optimizing only one set of model parameters (constants of springs and dampers) since they do not change with time. The numerical results of different models were fit to the results of a specially designed compaction characterization experiments conducted in our complementary study.