We Get You on

                    Course

Better Injection Moulding with Globular Shaped Fibers?

Layout of injection molded parts with anisotropic local properties

Question
In order to improve the mechanical properties (especially the stiffness) of a part, glass fibers are often added to thermoplastic polymers. However, the thermo mechanical material properties of the so created compound are anisotropic (direction dependent) and not isotropic. There are two reasons for this behavior: First, the thermo mechanical material properties of a polymer matrix differ at least by one order of magnitude from the glass fiber's properties; second, the aspect ratios of the fibers are typically l/d = 20. Consequently, a fiber is not of globular shape but looks like a thin bar.

Let us start with the assumption that all fibers are randomly distributed in the component. In spite of the significantly different (microscopic) thermo mechanical properties, one will generally observe an isotropic part behavior (macroscopic). However, due to the shear flow during injection, the local alignment of the fibers is corresponding to the three dimensional flow situation. Hence, the assumption of  a random fiber distribution in the injection molded part is not valid, you will actually find a complex, three dimensional distribution of the fibers, Figure 1. The part shows (globally as well as locally) anisotropic thermo mechanical properties.

These local part properties are real challenges for the calculation engineers. First, the input of reliable thermo mechanical material data is necessary in order to be able to simulate static load case evaluation, fatigue life time, or life time prediction. The use of isotropic substitution values would at best lead to an overdimensioning of the part. In the worse case the part would fail completely. Second, the well-established programs for structure simulation only provide limited functionality for the efficient consideration of local fiber orientation.

Approach
SIGMASOFT simulates the local three dimensional fiber orientation following the 3D flow situation during the injection and packing phase. A suitable micromechanical modeling determines the local thermo mechanical properties of the injection molded component based on the temperature and pressure history of the polymer as well as on the fiber orientation. This data can not only be used for the internal shrinkage and warpage calculation in SIGMASOFT, but also for the mapping with the SIGMALink module, where these results can simply be transferred to FE meshes (shells and solids), e.g. for Abaqus, Ansys, Marc, Radios, etc. Having this is mind, the calculation engineer is now able to use the real local anisotropic component properties for his evaluations. Figure 2 shows the integrative simulation process.

Application

The deformation of a housing component made of PA GF50 that is caused by an internal pressure test at room temperature and tracked with a suitable optical measuring system delivers the green measuring values (Figure 3). The repetition of the test with FE simulation using the common isotropic substitution values returns the red values. It can be seen that the isotropic calculation returns a qualitatively accurate deformation image. However, the isotropic calculation significantly underestimates the deformation from the quantitative point of view (factor 2). The recalculation of the loading case - using the fiber orientations simulated with SIGMASOFT as well as the resulting local anisotropic stiffnesses for the structure simulation - will lead to the blue results. The anisotropic structure calculation shows a deformation image with qualitatively high accuracy. This is the basis for an efficient and secure component layout.

Conclusion
The integrative simulation, i.e. the consideration of local component properties provided by the injection molding simulation, significantly increases the reliability of FE analyses. What is more, the coupling of SIGMASOFT with FE analyses allows the directed setting of anisotropic part properties, i.e., to create a higher stiffness in one direction (high fiber orientation) and subsequently a lower stiffness in the traverse direction (fiber distribution at random). SIGMASOFT comprehensively supports this method. After all it is good that fibers are not of globular shape. Otherwise, we would completely miss out on this optimization potential.


Figure 1: Fiber orientation in clipping mode: x component of the orientation tensor


Figure 2: Integrative simulation: Export of SIGMASOFT results, e.g. to Ansys


Figure 3: Compared values