Augmented Reality in the plastic piping industry How the combination of AR and visual AI can support, analyse and document the process of a plastic piping installation with an infrared welding machine

This research investigates how the combination of Augmented Reality (AR) and visual artificial intelligence (AI) can contribute to operational processes and analyse the quality of a piping installation with an infrared welding machine. The benefits of industrial AR have been discussed for training, maintenance, and human-robot interaction use cases. Furthermore, visual AI can be used through a cyber-physical system to inspect the quality in the manufacturing industry, connecting sensors to the internet and analysing the created data. However, a concrete proposal on combining AR and visual AI in the plastic piping industry for an infrared welding machine is currently missing. The conducted master thesis addresses this research gap.

Through digitalization, an increased amount of manufacturing processes are automated. However, it is not reasonable to fully automate all processes because humans, with their cognitive capabilities, can perform better in certain situations. Therefore, human-machine interaction is crucial, as the human factor is still the most important cause of accidents. Furthermore, due to market growth, the industry faces a shortage of skilled labour to conduct manual processes. For these reasons, the quality requirements of customers with critical applications, like rail, commercial vehicles, and the aerospace industry, are increasing. To improve an operator's skill level, AR can support by guiding through an operation process that is superimposed into reality. To monitor and document a process, visual AI is a suitable technology. The infrared welding technology for plastic piping systems is investigated during a literature review. Furthermore, the standards defining the quality requirements and other technologies which address this challenge are analysed. Additionally, the capabilities of industrial AR and the already used applications are highlighted. To evaluate how industrial AR combined with visual AI can support an infrared welding process and document the quality of the operation, an AR and visual AI prototype was created. The prototype was demonstrated to an IR-welding expert, a trained operator, and an untrained novice. During the evaluation, the participants conducted infrared welding with the support of the augmented instructions through a smart glass. Furthermore, the visual AI object detection algorithm analysed the defined quality criteria to evaluate the probability of using the correct tools. Finally, a semi-structured interview was conducted with all participants to receive valuable insights.

The AR guidance and visual AI feedback regarding the quality were perceived as helpful and supportive during a piping installation. Moreover, important factors which would limit the usage of the AR and visual AI experience during operations have been identified.