High-resolution X-ray computed tomography simulations of synthetically-generated volume porosity in continuous carbon fibre-reinforced polymer samples

The fast and correct determination of porosity in carbon fibre–reinforced polymers and other polymeric material is an important application area of X-ray computed tomography (CT). In this CT simulation study, microstructures such as individual carbon fibres, micro-voids, as well as the polymer matrix including mesoscale voids, were taken into account to generate CT data synthetically with well-known porosity. It was found that the ratio of total surface area (SA) of the voids divided by the total volume (V) of the voids is suitable to describe and differentiate individual porosity samples. Our investigations revealed that the ratio SA/V can be used to estimate a minimal necessary voxel size for proper porosity segmentation by a simple ISO50 threshold. Under certain conditions, using an adapted ISOxx threshold value at a resolution of (10 µm)3 voxel size, the porosity in a specimen can be determined with an average measurement error below 10%. As long as the CT resolution is not high enough to completely resolve all void structures, using global threshold segmentation is always a compromise between over-segmentation of the macro-voids and under-segmentation of micro-voids.