Data Acquisition Framework for spatio-temporal analysis of path-based welding applications

The use of digital technologies in industrial manufacturing reduces operational costs and improves production quality. The “Framework for Spatiotemporal Production Data Acquisition” (PathSense) aims to improve access and usability of production data by applying a common data ecosystem. Through integrating operational technology (OT) and information technology (IT), PathSense supports decision-making and process optimization. The framework uses the concept of digital shadows to collect critical information from sensors within the continuous manufacturing processes, e.g., welding or gluing lines. Spatial and temporal digital shadows are developed in accordance with human spatial cognition to support sophisticated-yet intuitive-human-computer interactions without the need for advanced data science skills. The backbone of the framework consists of robust data pipelines with a setup incorporating modern IT protocols, such as OPC UA and MQTT, to support the efficient acquisition and management of data. The paper addresses the challenges associated with the combination of IT and OT in cyber-physical systems, stressing modern complex data-intensive manufacturing as one exemplary domain to be tackled by scalable and secure data architectures. The paper identifies two major future directions: refining data integration processes and embedding advanced machine learning algorithms to enable automated data analysis and improve process quality monitoring. In summary, PathSense proposes a data-driven approach for quality inspection in manufacturing, which may eventually enhance industry practices and move towards data-driven decisions and increased operational flexibility.