Setup Operation Time Estimation System for the Adoption of Cyber-Physical Systems in High-Mix Low-Volume Production

The development and application of CPS (Cyber Physical System) to shop floor management has gained particular attention for its potential to revolutionize traditional manufacturing practices in recent years. However, applications in one-piece flow production or in low-volume, high-mix production, where it is necessary to deliver on time under high uncertainty in operations, has not yet made significant progress. Especially for SMEs (small and medium-sized enterprises) facing budget constraints, the investment for CPS implementation is challenging. Simplified performance management systems using devices like smartphones for manual input have been developed, and their adoption in the field is progressing. Such systems enable the identification of issues in work and production processes using collected data. However, improving accuracy of estimates for operation time and improving precision of production planning remains a challenge due to the limitations in accuracy of collected performance data. In this study, we propose a set-up time estimation system for the machining process of mold production, despite the limited amount of data and low data accuracy. It incorporates a cleansing algorithm based on both actual performance data and the expertise of experienced workers, along with an estimation model utilizing this cleansing algorithm. The effectiveness of the proposed system is examined through numerical experiments based on real-world data.