Simulation and validation of fibre alignment in fibre-reinforced plastics

Fibre-reinforced plastics are used in many injection moulded articles due to their excellent properties. However, they present themselves with one notable limitation: the material behaviour and thus also the shrinkage and warpage behaviour is clearly direction-dependent, i.e. anisotropic and not the same in all spatial directions, as is the case with isotropic metals, for example. For many users, the issue of anisotropy causes uncertainty, incorrect assumptions and thus suboptimal component and mould design, especially with regard to the article shrinkage to be assumed.

In the production of injection moulds, an average shrinkage allowance from the technical data sheet is often used in order to scale the component larger for the cavity and to achieve the desired dimensional accuracy in the injection-moulded article. This may work in some cases – especially for simple geometries – but when the geometry becomes more complex and fibre-reinforced materials are also involved, this simple assumption can deviate significantly from reality.

This is also the case in the following example: The connector developed, which functions both as a cover and a plug, has a much smaller dimension in the left-hand plug cage than desired – despite the shrinkage allowance applied in the mould. By means of injection moulding simulation, shrinkage and warpage tendencies can be predicted, the fibre alignment can be displayed in a meaningful way and exported as an anisotropic material model to the structure simulation for further development steps.

Evaluation options in the simulation

Basically, fibre orientations in the three spatial directions as well as in different planes can be evaluated individually. The sum of the three space-dependent individual values always results in
100 %. If there is 33% fibre orientation for each of the spatial directions x, y and z, an isotropic, i.e. direction-independent behaviour is shown.

For the use case of the connector, we will now restrict ourselves to the evaluation in the
longitudinal axis (z) through the article. Figure 2 shows the evaluation of the corresponding fibre orientation. It is clearly visible that the fibres in the affected left connector basket are mostly oriented in the longitudinal direction. The simulation shows orientation values of up to 80 % in the z-direction here. It is therefore possible to speak of a preferred direction in the left-hand connector basket, which is also noticeable mechanically.

This is exactly the reason for the “shrinkage problem” of the tool design discussed at the beginning: As a rule, shrinkage is significantly greater across the grain than in the grain direction. If a classic mean value is taken as the shrinkage measurement, this does not reflect reality. With the help of the simulation, conclusions can be drawn about the shrinkage allowance to be taken into account and the tool can be corrected accordingly. If this simulation had been carried out before the tool was built, this correction loop could have been avoided.

In order to validate our results and also take into account a feedback loop for continuous improvement of the simulation, we at BARLOG Plastics always carry out the comparison with the real component. Accordingly, the component was not only sent to the CT for dimensional inspection purposes – it is also possible to check the fibre alignment here! Particularly in the area of prototypes, it is a good idea to check the fibre alignment in the real article and draw conclusions regarding the component contour and injection position in order to further optimise this for the series article if necessary.

Fibre analysis using computed tomography

In recent years, industrial computed tomography (CT) has developed into an indispensable method for the non-destructive testing and characterisation of components. Starting with the complete initial sampling of components, through the visualisation of shape deviations, to non-destructive defect analysis, CT offers a multitude of application fields in the examination of plastic parts. One sub-discipline is the analysis of the fibre course. For this purpose, a sample is first prepared to ensure optimal visibility of the fibres. Taking a partial sample is necessary because this is the only way to achieve the resolution required to effectively evaluate the fibres in the CT.

In the CT, the sample is then transilluminated with X-rays and continuously rotated. From a large number of two-dimensional projection images, a high-resolution 3D volume model can be reconstructed with the help of a computer.
3D volume model can be reconstructed with computer support, which shows the fibre structure in detail. The fibres can now be segmented from the surrounding matrix via differences in grey value intensity using special image processing algorithms and analysed quantitatively with regard to various parameters. These include e.g. fibre volume & orientation, but also defects or inhomogeneities.

Simulation validation on the real component

Using the connector as an example, a sample can be taken from the side wall of the connector cage. As a result of the analysis, the diagram in Fig. 3 can be derived, which shows the fibre orientation in the different spatial directions over the sample thickness taken.

The red line in Fig. 3 shows the orientation of the fibres in the z-direction. As predicted in the simulation, the reality also shows a very high orientation of the edge layers of about 75% of the fibres in the flow direction. Another typical behaviour of fibre-reinforced plastic components can be seen in the diagram. In the core layer, the high orientation in the z-direction drops significantly, while the orientation in the x-direction (light blue line) increases equally strongly. This is due to the fact that the plastic melt solidifies directly in the boundary layer on contact with the mould wall and as a result a high velocity gradient is created between the solidified boundary layer and the still molten core layer (plastic core). This shear, together with the biaxial flow profile of the melt, ensures a longitudinal alignment of the fibre near the edge layer and a transverse alignment in the core layer. Looking at the whole sample, it can be seen that the highly oriented edge regions are strongly pronounced and the core layer only takes up about 20% of the total thickness. Thus, the CT analysis on the real component, analogous to the simulation, leads to the result that the fibres in the area of the connector cage are anisotropically aligned and thus a globally assumed shrinkage is not purposeful.

Do you have further questions or need help with the implementation of your project?
Our CAE department will be happy to help you under the contact options listed below. Feel free to contact us!

Your contact to the author for queries and technical discussions:

Tobias Haedecke, M. Eng.
Division Manager / Director Engineering
+49 160 / 271 43 60
tobias.haedecke@barlog.de
www.barlog.de/leistungen/cae-services/