At the Institute of Joining and Welding Technology a palm-sized test specimen for the characterization of casting materials is currently being researched. This consists of a metal insert which is completely surrounded by a polymeric casting material. In order to ensure the comparability of the test specimens produced and thus the results obtained in material screenings, the position of the metal insert in particular must be ensured within certain tolerances. Since the potting polymer is usually not transparent, a purely visual inspection to determine the positional tolerance of the insert is not possible. Common methods such as the CT scan work, but are not suitable for the rapid inspection of a large number of test specimens. The development of a new combination of methods would therefore appear to make sense. In preliminary tests, a combination of inductive heating and the recording of a thermal image due to the different thermal conductivities has emerged as a possible method.
Within the scope of the student thesis, a system for automated position determination of metal inserts in test specimens is to be designed, implemented and evaluated. Thereby the focus is to be put on the process combination of inductive heating and recording a thermographic image. The position of the metal insert is to be evaluated from this image using image processing and/or machine learning algorithms and an OK/nOK evaluation of each test specimen examined is to be made based on specified tolerance values.
