AbstractWhilst Part inspection and dimensional validation (PIDV) is a well-established practice for external and accessible features, the capture and analysis of internal features and the dimensions associated with them has always been an issue. As parts designed for additive manufacture become more widely used, this issue is compounded, since internal features and hollows can be easily introduced. For parts manufactured through additive manufacture or by traditionally cast methods, if the internal structure requires PIDV or an in-line validation check, the internal structure will need to be revealed so that a measurement can take place. Presently, the only available solutions to this issue are destructive measurement, or X-ray computed tomography (XCT) scanning. Neither of these solutions are universally practical, and in the latter case not readily available.
This thesis introduces an alternative method for measuring internal features. The method requires an intentionally induced temperature differential between the internal and external features. The resultant temperature distribution is measured on the surface using a thermographic camera.
Using this technique, in combination with standard multi-view projection, this method of data recovery can discern an object’s internal structure and provide inferred measurements for that structure. This result, combined with any industry-standard method for the measurement of the external features, can provide a complete 3D digital recreation of the object in question. This technique has the benefits of being non-destructive, not requiring extensive training or knowledge to operate, and being more affordable and more portable than XCT.
The aim of this investigation was to devise and evaluate the feasibility of this approach. This process began with a series of FEA simulations to prove that internal geometric measurements could be extracted from forcibly induced surface temperature profiles and that the spatial and temperature resolution required for this extraction were sufficient. For low conductive materials, with forcible induced internal temperatures or around 100°C, internal edge extraction was possible to within 1.5mm. From this initial validation, a series of physical experiments using 3D printed, and machined artefacts were performed to validate the computer simulations and understand the limitations when using a real thermographic camera, rather than picking ideal data from a simulation.
The experimental results showed that through the thermal manipulation of an object’s internal structure, geometric dimensions could be resolved to within +/-1.8mm with a repeatably of 0.6mm. When combined with external surface data from industry-standard capture techniques, this novel approach successfully resolved a complete Computer Aided Design (CAD) “solid model” encapsulating a holistic set of geometric measurements.
In addition to findings relating to spatial accuracy, and material limitation, this research has highlighted how this capture technique can discern differing internal structures based on temporally gathered temperature data. When the temperature data from the test artifacts is gathered temporally, the resultant data shows differing results for curved, angled, and straight internal faces respectively. This, going forward, would allow for the automatic categorisation and recognition of internal features based upon the recovered thermal response.
|Date of Award||2023|
|Supervisor||Simon Fletcher (Main Supervisor) & Andrew Longstaff (Co-Supervisor)|