In the case of short fiber-reinforced plastics, it is not strictly correct to speak of “material properties”. Rather, we are dealing with part-specific properties, since the “material” itself is only formed during the part’s manufacturing process. Therefore, to realistically describe the part’s behavior, it is essential to incorporate the manufacturing process into the simulation chain. For short fiber-reinforced plastics, an injection molding simulation provides the local fiber distribution within the part, which forms the basis for determining the directional dependency of its mechanical behavior. The following sections explain the specific requirements this imposes not only on the material modeling but also on the part’s meshing. It then describes how the new 'Orientation Profiling' feature in Converse V 5.1 can be effectively used in this context.
Material and Part Properties
Fiber-reinforced materials are characterized by their layered structure. While this structure is intentionally designed into continuous fiber-reinforced parts, it naturally forms in short fiber-reinforced injection-molded parts due to flow processes during mold filling (Figure 1). The mechanical properties of the individual layers, overlaid across the part thickness, represent the part’s local properties. These are so-called homogenized properties, as no distinction is made between the components of the composite - fiber and matrix.
Orientation Tensor and Fiber Orientation
To determine such homogenized mechanical properties, knowledge of the local microstructure - specifically the fiber orientation - is necessary. Injection molding simulation can determine fiber orientation at every position and across the wall thickness of the part. However, the simulation only provides the fiber orientation distribution indirectly via the orientation tensor. The spatial distribution of fibers must be reconstructed from this tensor using mathematical methods (Figure 2). This reconstruction is not the focus here and is handled automatically in Converse in cooperation with MatScape.
To accurately represent the directional dependency of part behavior in an FEM model, special meshing requirements arise. In conventional isotropic FEM analyses, finer mesh resolution is required in regions with high stress gradients. Similarly, anisotropic simulations must resolve local fiber orientation, especially across the part thickness. Due to shear and elongational flow in the melt, a core layer and outer layers typically form (Figure 3). Fibers in the core tend to orient stochastically or transversely to the melt flow, while fibers in the outer layers align with the flow. These differences, as well as the thickness of the respective layers, significantly impact mechanical behavior.
Therefore, the injection molding simulation must resolve fiber orientations sufficiently, and the structural simulation’s mesh must match in resolution. Usually, these meshes differ, so Converse maps fiber orientations from the injection molding mesh to the structural mesh (Figure 3).
This means that, starting from the center of each finite element in the structural mesh, the corresponding fiber orientation at that coordinate in the injection molding mesh is identified and transferred. The structural model’s accuracy is thus limited by the injection molding model’s resolution. Often, the structural mesh has a coarser resolution, reducing the fidelity of the orientation mapping.
Use Case: Material Card Calibration
Various scenarios are exemplified below using the new profiling tool available in Converse V 5.1. A simple injection-molded plate geometry and tensile specimens with different mesh resolutions are used (Figure 4). This simulates the calibration process for creating an anisotropic material card. Tensile bars are cut from different directions in the plate (only 0° in-flow direction shown), and tensile tests are simulated via FEM. Material model parameters are iteratively adjusted until results match the experiments. Only the tensile test in flow direction is illustrated here for simplicity.
The injection molding simulation in this case used a 2.5D simulation (Cadmould, Simcon), capturing ten layers (five symmetric) over the plate thickness. This is considered adequate for representing fiber orientation. The tensile bar was modeled using ten linear hexahedrons and quadratic tetrahedrons with three and one element(s) over the thickness.
Figures 5 through 7 show orientation profiles derived from the injection molding simulation and mapped to the tensile bar models. “Source” indicates the injection molding simulation model; “Target” is the structural simulation model.
In the high-resolution hexahedron model (Figure 5), the orientations from the injection molding simulation are fully transferred, as both models share identical element counts per layer. Such resolution is rarely feasible in real parts; a more typical mesh might use three tetrahedrons (Figure 6). Here, the core layer is still resolved but slightly offset due to uneven element spacing. Nevertheless, this resolution is acceptable, particularly for stiffness predictions. In contrast, the one-element tetrahedron model (Figure 7) lacks core layer detail, representing only surface fiber orientations.
Figure 8 compares the mechanical response across different mesh. Comparative isotropic analyses confirmed that the differences are due to orientation transfer quality, not meshing alone. The hexahedron model is considered the reference, realistically representing the tensile bar's behavior. The three-element tetrahedron model shows nearly similar performance, while the single-element tetrahedron overestimates stiffness and differs in yield behavior. The effort involved in this case in an anisotropic simulation is hardly justified by the accuracy of the results achieved. The method’s full potential is lost here due to insufficient discretization in the structural model. If the effort for a holistic simulation is made, each step in the CAE chain should reflect this—especially via appropriate discretization.
Conclusion
The new profiling tool in Converse is a valuable resource for verifying mapping quality, particularly at critical part hot spots. It allows for easy validation of anisotropy representation. During material calibration, it also proves useful, especially in combination with MatScape during iterative material card calibration. The profiling tool is available starting with Converse version 5.1.
Author:
Dr. Wolfgang Korte is Managing Director at PART Engineering GmbH, Bergisch Gladbach
