Abstract
In the future, one of the key essential functionality of smart factory is reconfigurable manufacturing. Collecting, analyzing, and monitoring large amounts of sensor data are becoming a key enable technology to implement reconfigurable manufacturing. These day, the manufacturing industry is in the midst of a data-driven revolutionary, which means sensor data has become more accessible, ubiquitous and in the end generating large amounts of stored sensor data. This phenomenon would make it challenging for manufacturing SMEs to store and analyze them into useful and actionable information. However, it is not easy for Small and Medium-sized Enterprises (SMEs) to implement such actions due to their limited budgets and lack of ICT infrastructure. Thus, in this paper, we developed a distributed processing system for real-time sensor data, which is affordable to such manufacturing SMEs. The proposed framework incorporates machinelearning algorithms that could learn from historical sensor data and generate the classifiers which will be used to predict the product quality based on the sensor data. An injection-molding process is used to verify the proposed framework.
Published in: Proceedings of KSPE 2017 Spring Conference
Link: http://www.dbpia.co.kr/Journal/ArticleDetail/NODE07205708