Proactive Control of Manufacturing Processes Using Historical Data

Today’s enterprises have complex manufacturing processes with several automation systems. These systems generate enormous amount of data in real-time representing feedbacks, positions, and alerts, among others. This data can be stored in relational databases as historical data which can be used for product tracking and genealogy, and so forth. However, historical data is not been utilized to proactively control the manufacturing processes. The current contribution proposes a novel methodology to overcome the aforementioned drawback. The methodology encompasses three process steps. First, offline identification of critical control-related parameters of manufacturing processes and defining a case base utilizing previously identified process parameters. Second, update the case base with real-time data acquired from automation systems during execution of manufacturing processes. Finally, employ similarity search algorithms to retrieve similar cases from the case base and adapt the retrieved cases to control the manufacturing processes proactively. The proposed methodology is validated to proactively control the manufacturing process of a molding machine.