Modelling the laser cutting of carbon fibre reinforced plastics

Share this on social media:

Jan Keuntje and Peter Jäschke have developed a macroscopic finite element model to help establish laser cutting in serial CFRP component production

The demand for carbon fibre reinforced plastics (CFRPs) has continued to grow in recent years1CFRPs are particularly suitable for use in lightweight construction due to their high specific strength and low density. 

The choice of components and layer arrangements enables the mechanical properties of CFRPs to be adopted and optimised according to the intended load case.

The Airbus A350 is a good example of the increased usage of CFRPs as a construction material, with more than half of its components being made from them. However, the fabrication of CFRP parts is still challenging when looking to achieve highly automated series production.

Conventional manufacturing processes, such as milling, have disadvantages in that forces are introduced into the workpiece during machining. This leads to high tool wear due to the high hardness of the carbon fibres, and can further result in inadequate quality. 

As an alternative, laser-based cutting has shown promise. The laser operates without contact and is therefore practically force- and wear-free. In addition, laser machining offers a high degree of automation potential. However, the effects of laser cutting have not been adequately studied. Thus, strength and service life are only predictable to a small extent. 

The main challenge is the formation of a distinctive heat-affected zone (HAZ) during the cutting process caused by the absorbed laser radiation2. The heterogeneous build up of CFRPs, the different properties of the carbon fibres, and the matrix material are the reason for this phenomenon. Heat flows away from the cutting kerf along the carbon fibres, which have a significantly higher thermal conductivity than the matrix. At the same time, the decomposition temperature of the matrix is much lower than that of the fibres. This can lead to heat-affected areas in the matrix around the cutting zone, depending on the fibre orientation. The consequences of the HAZ in terms of strength loss can be determined experimentally by destructive methods3,4. To reduce time and resource consumption during pre-testing, the goal should be to predict the formation of these HAZ beforehand and adjust the process strategy accordingly, or to adapt the part design. 

A reliable prediction, which is not available yet, is therefore a basic requirement to establish laser cutting in serial CFRP component production.

Creating a macroscopic simulation

Consequently, a three-dimensional finite element model has been developed at the Laser Zentrum Hannover (LZH). The numerical simulation model aims to project the temporal and spatial evolution of temperature fields during the laser cutting process on a macroscopic scale. The simulated temperature distribution should allow a prediction of the expected HAZ.

Characteristic for composite materials, such as CFRP, are the anisotropic material properties, which have to be considered in developing the model. One way to handle this is with a heterogeneous modelling approach, separating realistically between the fibre and matrix. These kind of models are very useful to observe the formation of the HAZ itself and to analyse the exact ablation mechanisms. However, a mesh resolution in the micrometre range is necessary to distinguish between the two composite components. This is accompanied by a high effort in modelling and calculation. Hence, this approach is limited to small volumes.

For process development and analysis of the global temperature during laser processing, simulation on a larger scale is required. Therefore, a macroscopic simulation model has been used, which addresses the machining of larger volumes in the range of cubic centimetres. For this approach, a homogenised material model was used, with the material properties being implemented through the aid of the so-called laminate theory5. This takes into account the proportions of both materials according to the fibre volume fraction and the orientation of the individual layers in the composite. This way, the global material behaviour is represented without the detailed distinction between fibre and matrix. This approach is suitable for the spatially resolved representation of the generated heat and thus the HAZ in a higher order of magnitude.

Another challenge is the implementation of the laser beam as a heat source. Usually, a pulsed laser and a multi-pass cutting strategy are applied for the laser cutting of CFRP. In this specific case, a TruMicro 7050 nanosecond pulse laser from Trumpf is used with a pulse duration of 30ns and a maximum pulse energy of 80mJ. The implementation of each individual pulse in the simulation environment has decisive disadvantages in the context of consideration on a component scale. Depending on the cutting length and velocity, the number of resulting pulses is huge, because the laser operates in the kilohertz range. Additionally, the mesh size has to be set according to the ablation depth of a single pulse, which depends on the optical and thermal penetration depth of the material. This would, in turn, result in a microscopic approach and would lead to undesirable computational effort. Based on this – and due to the high frequency of 18.8kHz that is used – the process is simulated as a continuous-wave process. A moving heat flux is used as a boundary condition. 

The energy provided by the laser during a single pass is summed up and is then uniformly applied to the cutting area, step by step. Elements that exceed the sublimation temperature of CFRP, which is estimated to be 3,650°C6, are deactivated from the model after every calculation step to simulate the ablation process. To realistically determine both the resulting temperature distribution and the ablated volume, the element sizes and the computational step sizes were adjusted in an iterative process, comparing simulation results with experimental data. The longer the computational steps size, the longer heat can conduct from elements into the adjacent material until they get deactivated, even if they already exceed the corresponding temperature. Thus, the step size can be used as an iterative parameter, as well as the element size, which is mainly responsible for the ablated volume. It has to be taken into account that both parameters have an influence on the temperature distribution.

Figure 1: Comparison between measured and simulated temperature progression for different cutting velocities (blue: vcut = 3m/s, red: vcut= 1m/s) at a distance of s = 75μm away from the kerf

For experimental comparison, the process temperatures are measured with the help of thermographic imaging. The recorded data is heavily influenced by process emissions, so the comparison focuses on the cooling phase after a laser repetition. An example for the comparison of simulated and measured temperature is presented in figure 1.

The graphs show the temperature profile after a single laser passes over the specimen for two different scanning velocities (1 and 3m/s) at a 75µm distance next to the cutting edge. The zero point of the x-axis was set equal to the beginning of the cooling phase (laser off). The curves show a good agreement between experiment and simulation. In this case, unidirectional CFRP was used, which means all layers (and thus fibres) are orientated in the same direction. These cuts were performed along the fibre direction, so hardly any HAZ could be found. Hence, further cuts were carried out under a cutting angle of 90 degrees to the fibre orientation. The cutting velocity of 0.5m/s was deliberately chosen to create a visible HAZ. 

The right side of figure 2 shows a magnified image of the associated cutting result, where the HAZ can be easily observed in the form of exposed fibres and discolorations.

Figure 2: Top view of the cutting result after a single repetition with Vcut= 1m/s at 90° to the fibre orientation. Left: Simulated temperature plot with expected HAZ. Right: Microscopic image of the cut with visible HAZ

Next to this, the corresponding simulation result is shown as a temperature plot in the same scale. The 200°C-isotherm, representing the damage temperature of epoxy, is used as an indicator for the expected extension of the HAZ. While the current model likely overestimates the HAZ on the surface, it should be noted the visible HAZ corresponds to a minimum extension of the HAZ. Further damages are likely and could be visualised in a cross section.

Outlook

The evaluation of the model has to be further continued through experimental comparison. So far, the capability of the model is limited to a small number of laser beam passes. Increasing the number of passes should soon be achievable with programming effort.

Furthermore, the range of validity of the model needs to be further investigated for a wider range of parameters, like the cutting velocity and laser power. It needs to be elaborated if the iterative chosen parameters (especially the time step size), have to be adjusted for a wider range of experimental parameters. 

For further investigations we plan to add thermocouples on the top and bottom surfaces and even within the material to gain more data about the spatial temperature distribution within the compound structure. This is to go beyond a surface temperature measurement and to provide more knowledge about the heat flow, or rather the thermal conductivity, of the investigated material.

Jan Keuntje is a research associate in the field of laser processing of polymer and composites at LZH.

Peter Jäschke is head of the production and systems department at LZH.

The authors gratefully acknowledge the funding by the German Research Foundation (Deutsche Forschungs-gemeinschaft, DFG) of the project, ‘Characterization and modelling of the laser-based separation process and resulting damage mechanisms of carbon fibre-reinforced plastics under fatigue loading’ (project number 436398518).

The authors would also like to thank Trumpf for providing the TruMicro 7050 laser source.

References

[1] Sauer, M. Composites Market Report 2020 – The Global CF-Market 2020 – Composites United, Germany, 2020.
[2] Bluemel, S., Jaeschke, P., Wippo, V., Bastick, S., Stute, U., Kracht, D., and Haferkamp, H. ‘Laser Machining of CFRP Using a High Power Fiber Laser – Investigations on the Heat Affected Zone’, Proceedings of the 15th European Conference on Composite Materials, 2012.
[3] Herzog, D., Jaeschke, P., Meier, O., and Haferkamp, H. ‘Investigations on the Thermal Effect Caused by Laser Cutting with Respect to Static Strength of CFRP’, International Journal of Machine Tools and Manufacture, Vol. 48, Nos. 12–13, 2008, pp. 1464–1473.
[4] Jaeschke, P., Kern, M., Stute, U., Haferkamp, H., Peters, C., and Herrmann, A. S. ‘Investigation on Interlaminar Shear Strength Properties of Disc Laser Machined Consolidated CF-PPS Laminates’, Express Polymer Letters, Vol. 5, No. 3, 2011, pp. 238–245.
[5] Staab, G. H. Laminate Analysis, Laminar Composites, Elsevier, 2015, pp. 189-284.
[6] Freitag, C. ‘Energietransportmechanismen bei der gepulsten Laserbearbeitung Carbonfaser verstärkter Kunststoffe’ (Dissertation). Herbert Utz Verlag, Wissenschaft, München, 2017.

Navigation

Navigation

Navigation

Navigation

Navigation

Navigation

Richard Stähr and his colleagues are using laser drilling to produce perforated acoustic liners for insulating aircraft noise

18 February 2022

Lasers offer a non-contact, fast, flexible and cost-effective tool for cutting sheets across a range of industries. (Shutterstock/Parilov)

04 October 2022

Handheld laser cleaning devices are a popular choice in industry for delivering the beam to the workpiece. (Image: Shutterstock/Surasak_Photo)

15 September 2022

Richard Stähr and his colleagues are using laser drilling to produce perforated acoustic liners for insulating aircraft noise

18 February 2022

Dr Michael Jarwitz, of the University of Stuttgart’s IFSW, highlights the need for a single tool capable of highly versatile and adaptable manufacturing

31 August 2022

Multiphysics simulation software modeling the megahertz-level frequencies of Civan’s dynamic beam shaping lasers. (Image: TU Wien)

03 March 2022