Characterisation of powder-filled defects in additive manufactured surfaces using X-ray CT

Assessing functional performance is the most important stage of any component verification. Mechanical properties can be evaluated by means of destructive testing which can be both expensive and lengthy in addition to loss of the original component under test. It is therefore advantageous where possible to utilise non-destructive techniques that can achieve the same or similar outcomes through collection of three-dimensional data that can then be used in simulation to determine functionality. Such non-destructive methods with 3D location ability are essentially density- and porosity-based testing methods. Additive manufacturing allows the creation of complex geometrical features that are often defined based on function.
Optimisation of AM component geometry based on functionality allows for the specification of components that have features that cannot be mapped efficiently to current GPS standards ISO 14638. In addition, the integrity of complex optimised AM structures that may lie on a critical stress or heat path must be assessed and any elements of unfused powder for example, must be detected. This seeks to investigate the ability of X-ray computer tomography to detect and characterised small scale empty and powder filled defects which may occur in AM manufactured parts. To achieve this, aim a Ti6AL4V artefact built using an Arcam Q10 electron beam-melting machine (EBM). Defects of between 50 and 1400 microns in diameter were machined into the surface of the artefact using a precision CNC machine equipped with micro-drills. Once this was achieved, the defects were characterised using focus variation microscope. Virgin Ti6AL4V powder was added to fill 50% of the defects and then the artefact was measured using a Nikon XTH225 industrial CT. This was used to analyse the relative size and volume of the defects and assess the capability of the inspection process to both assess the size of pores and to detect the powder-filled defects. To reduce the number of process variables, all the measurement process parameters, such as filament current, acceleration voltage and X-ray filtering material and thickness, were kept constant between the scans with hollow and powder filled defects. The acquired data processing, surface determination process and defect analysis was carried out using VgStudio Max (Volume Graphics, Germany). The focus of the study is on providing best practice regarding the selection of inspection parameters and identifying the capability of the process to detect unfused powder.