Electron Beam Melting (EBM) is an Additive Manufacturing (AM) process which can be used to fabricate complex 3D metal components. Ti-6Al-4V is one of the high value aerospace materials that has been widely used in the EBM process. However, defects such as gas pores and lack of fusion are often found in the as-built components, which can make a significant difference to the fatigue resistance of the parts. The Manufacturing Technology Centre (MTC) uses X-ray Computed Tomography (XCT) and 3D image segmentation techniques to quantify the size, morphology and distribution of defects in EBM samples. This has allowed further process parameter optimisation to manufacture component of better quality.
XCT data sets were post-processed using Avizo 9.3, an advanced 3D visualisation software from Thermo Fisher Scientific. A Python script was produced to automatically analyse the data and extract statistical information used to compare each sample based on the process parameters. Three types of image segmentation were used to automatically identify potential defects within each sample based on the highest peak in the histogram defined as "material" (Interactive Thresholding), the Otsu thresholding (Auto Thresholding) and the low greyscale values that could be potential voids or low density regions (Adaptive Thresholding). The results of each were merged together to avoid any duplication of data, and then analysed using Label Analysis. The resolution was approximately 10 microns for each scan the results were therefore filtered to ignore any defect smaller than 30 microns (three voxels) in all directions. The main characteristics that this study focused on were the sphericity of potential defects (derived from Shape_VA3d), the biggest Feret diameter (Length3d) and the total porosity from the sample. The porosity value will be biased depending on what settings are used during the image processing. Consequently, the automatic XCT data analysis script was used only for comparing results between samples that were consistently acquired.
Quantification of potential defects in Ti-6Al-4V samples using Electron Beam Melting.