Improved Temperature Monitoring and Control of Production Lines in Casting through BaSyx Framework and Edge Intelligence

Metal production is still permeated by brown-field environments, which are continuously becoming increasingly digitized. Therefore, smart and superordinate approaches facilitating autonomous decision-support systems and plant control, while accompanying the running production, are fundamental for the modernization progress. Cyber Physical Production Systems generating a digital envelope for modular communication and optimization offer these necessary prerequisites. Still, current platforms concentrate on mostly singular aspects ranging from cloud-based access, end-to-end security, connectivity, and interoperability. This paper is dedicated to shifting analytical capabilities to edge components, avoiding high-volume data transfers, comprising the main information to smart data, and reducing latency, allowing a real-time control under improved conditions. Here, a new consistent design is applied and tested within a brown-field environment of a centrifugal casting plant. Improving the highly energy-intensive setup process before casting is tackled by utilizing existing edge-controller hardware and designing an integrated approach based on the BaSyx-framework for deploying developed process models at plant level for online control. Results show how white-box models, including density of the coating and temperature behaviour of the mould, are integrated to generate and display suitable setup strategies for the casting process.