Rybak + Hofmann rhv-Technik GmbH + Co KG
Contact at the AI Progress Center
Christof Nitsche
Use cases for artificial intelligence in the field of thermal spray processes and surface coating
Initial situation
Rhv-Technik is a medium-sized company based in Waiblingen and offers innovative and individual solutions in the field of surface technology, primarily by means of thermal spraying processes. Today, thermally sprayed components are used in almost all industries to adapt materials to individual applications, increase durability and reduce failures. Rhv-Technik develops individual coatings together with its customers. Various processes are used, which means that a wide range of machines are used at the same time. The batch sizes are comparatively small due to the highly specialized applications.
As part of the AI Explorer, the aim was to find out which use cases for artificial intelligence (AI) in the field of thermal spraying arise for a company like Rhv-Technik from the current state of research. In addition, awareness within the company for the topics of AI and digitalization was to be promoted.
Procedure
Three workshops were held to build up knowledge and identify and evaluate use cases. In the first workshop, the basic knowledge of AI and its potential in production was conveyed. In addition, the status within the company was recorded and potential fields of application identified. This was followed by the second workshop, in which the state of research was presented and four potential use cases were outlined. These were subjected to a benefit analysis in order to estimate added value and costs. Two of the four use cases were then fleshed out and discussed during the final workshop, and a recommendation for implementation was made.
Results
The following four use cases emerged from the workshops:
- Prediction of shift parameters based on public and internal data sets
- Audio-acoustic process monitoring by means of anomaly detection
- Quality control using a hand-guided system
- Predictive maintenance based on order and maintenance data using the Hidden Markov Model
The benefit analysis showed that use cases 1 and 4 could be particularly promising for medium-term, cost-effective implementation. The benefits of all use cases were almost identical, while 1 and 4 can be implemented with considerably less effort. Use case 4 is also preferable and should therefore be continued in a quick check if necessary.
