Credit: DLR (CC BY-NC-ND 3.0).
It is estimated, that the application of artificial intelligence (AI) has high potential to increase shop floor productivity. These positive assessments are currently supported by only a small number of available options for quick implementation. Ready-to-use solutions in the field of mechanical and process engineering are only offered by manufacturers in combination with corresponding equipment. Therefore, complex AI projects have to be implemented as part of expensive unique operations. For SMEs, overcoming these obstacles represents a high financial effort, especially when the prospects for success are uncertain.
In this project, an easy-to-use yet adaptable framework for process support is to be developed by the project partner WOGRA AG (https://wogra.com/). The framework allows existing workshops to be upgraded cost-effectively by digitally mapping the production processes. Experienced and technologically skilled employees of small and medium-sized enterprises (SMEs) should be enabled to integrate AI into existing process chains on their own.
To achieve this, manufacturing data from components will be collected using low-cost mobile and embedded devices throughout the entire production process. The collected data will be used to evaluate individual process steps with regard to their automation capabilities for AI implementation. Based on the collected and processed data, if the framework determines that a process step can be automated (e.g. a quality control), it will make suggestion to the person responsible for the process.
During the project, AI will be used and evaluated in particular for sensor data of the continuous ultrasonic and resistance welding process of thermoplastic composites supplied by DLR. The goal here is to provide automatable and non-destructive quality assurance of the strength of the weld as well as support for process optimization. Finally, DLR will investigate the results obtained with regard to statistical significance and classification of the reliability of the AI's predictions.
The framework developed will be made available as open source in order to achieve wide dissemination and accelerate the use of AI in production.
Supported by: