On 16 April 2025, Patrick Timm defended his master's thesis on ‘AI-based quality control in shipbuilding’. Patrick is a Master's student on the Business Informatics programme and wrote his thesis in cooperation with the Fraunhofer Institute for Large Structures in Production Engineering IGP. He was supervised by Anne Gutschmidt, a member of staff at the Chair of Information Systems, and Christian Scharr, a member of staff at the Fraunhofer IGP.
Abstract
This master’s thesis is dedicated to the development of an AI-based system for the automated quality control of padeye adhesives in the shipbuilding industry. The transition to an automated bonding process renders intelligent monitoring technologies suitable for detecting critical bonding defects, such as air inclusions or incomplete seals.
As part of this work, the requirements for such a system were first determined through semi-structured interviews with industry and scientific specialists. This was followed by a structured literature analysis to obtain a well-founded overview of existing visual quality control methods. The findings of this study guided the selection of the Faster R-CNN for a feasibility study, implementation and evaluation.
The results obtained from this study indicate that the Faster R-CNN model is generally suitable for detecting bonding defects. However, limitations with regard to model generalisation and image quality were identified. The study thus provides the first reliable findings on the technical feasibility of AI-based quality control and shows potential for further research and development work.