Non-isothermal direct bundle simulation of SMC compression molding with a non-Newtonian compressible matrix

TL;DR

This paper compares macroscale and mesoscale models for simulating fiber orientation and compression forces in SMC molding, demonstrating that simpler models can effectively predict key behaviors in complex, non-isothermal, non-Newtonian conditions.

Contribution

It shows that Jeffery's basic model can predict fiber orientations comparable to detailed mesoscale simulations in non-isothermal, non-Newtonian SMC compression molding.

Findings

Both models predict compression force contributions well for short flow paths.

Complex deformation mechanisms cause deviations in long flow paths and thick stacks.

Jeffery's model performs well for planar SMC flow, sometimes better than advanced models.

Abstract

Compression molding of Sheet Molding Compounds (SMC) is a manufacturing process in which a stack of discontinuous fiber-reinforced thermoset sheets is formed in a hot mold. The reorientation of fibers during this molding process can be either described by macroscale models based on Jeffery's equation or by direct mesoscale simulations of individual fiber bundles. In complex geometries and for long fibers, direct bundle simulations outperform the accuracy of state-of-the-art macroscale approaches in terms of fiber orientation and fiber volume fraction. However, it remains to be shown that they are able to predict the necessary compression forces considering non-isothermal, non-Newtonian and compaction behavior. In this contribution, both approaches are applied to the elongational flow in a press rheometer and compared to experiments with 23% glass fiber volume fraction. The results show that both models predict contributions to the total compression force and orientation reasonably well for short flow paths. For long flow paths and thick stacks, complex deformation mechanisms arise and potential origins for deviation between simulations models and experimental observations are discussed. Furthermore, Jeffery's basic model is able to predict orientations similar to the high-fidelity mesoscale model. For planar SMC flow, this basic model appears to be even better suited than the more advanced orientation models with diffusion terms developed for injection molding.