X-ray computed tomography (CT) has emerged as innovative measuring technique for dimensional metrology in industry over the last years. Thanks to unique capabilities, CT provides several advantages in comparison with other well-established coordinate measuring systems (CMSs). In particular, CT allows obtaining a holistic three-dimensional model of the scanned workpiece and performing non-destructive and non-contact measurements of outer as well as inner features and geometries difficult to access. However, important drawbacks limit a wider acceptance of this technology in industry. One of the most critical problems is the complexity of metrological traceability establishment due to difficulties in evaluating the task-specific uncertainty, as well as specification and determination of metrological performances of CT systems. In fact, a dedicated international standard for CT acceptance test and performance verification is still under development.
In this thesis, experimental activities were mainly oriented at evaluating and improving CT metrological performances. The material influence on length measurement errors was studied by investigations based on two reference objects: aluminium hole plate, with significant material effect and ruby ball plate, with negligible material influence. The obtained results contributed to the test survey organized on this topic by the ISO working group that is developing the future ISO standard for CT.
The image quality of the 2D projections acquired by CT is fundamental for achieving a good reconstruction quality. It is directly connected to the image blurring content and, consequently, to the focal spot quality. In this thesis, new methods developed to assess the focal spot drift and size are presented.
Considering the entire CT measurement procedure, an important metrological characteristic to be evaluated is the metrological structural resolution (MSR). Despite several methods for MSR evaluation have been proposed, a standard test to be included in the ISO standard for CT has still to be defined. In this thesis, the ‘Hourglass’ method –developed by the University of Padova– was selected to evaluate the MSR. The method was improved by measuring a high number of distances and by applying a definition of MSR similar to the one proposed by the guideline VDI/VDE 2617-13. Moreover, the method concept was compared with concepts of other proposed methods. Finally, the main influence quantities affecting the method were identified and evaluated.
Since CT is a multi-purpose measuring technique, it has become attractive for many industrial applications. However, the accuracy of CT measurement results is often unknown. For this reason, the evaluation of CT accuracy for specific measurement tasks as well as the comparison with other well-established evaluation methods is crucial for the acceptance of CT in industry. In this thesis, two industrial case studies were addressed: (i) porosity analysis for metal additive manufactured parts and (ii) wear evaluation of polymeric prosthetic components. In the first study, CT was compared with other inspection techniques such as Archimedes method, gas pycnometry and microscopic analysis of cross-sections. A multisensory CMM was used as well in order to get reliable reference area values for pores lying on selected cut sections. In the second study, the proposed CT-based method was validated through comparison with the gravimetric method, which is the current reference method used for wear assessment.