A framework for process states structural interpretation of zero-defect manufacturing

Recent advances in quality manufacturing research have focused on the industrial landing of Zero-defect Manufacturing (ZDM) which is aiming toward a more precise, robust, and sustainable production paradigm. A systematic deployment platform for ZDM implementation need to take advantage of the various advanced technologies and integrate them. Cyber-physical system (CPS) is a critical framework and low-rank representation (LRR) is the method which has widely used in computer vision, signal processing and other research areas. This paper describes a novel framework based on the interdisciplinary integration of cyber-physical architecture and low-rank representation, which is named the CPS-ZDM-LRR framework. It transforms the quality control problem into the signal monitoring, to complete the process states interpretation and deal with the hidden defect problem in ZDM. Through the continuous monitoring of products and equipment' status during manufacturing process, the real-time raw data from different sources has been preprocessed to the time series features which are slide keyframe matrices, and LRR used to search the low-rank structure of slide keyframe matrices which can help us recognize the current status of manufacturing system deeply and give the preventive measures suggestion for quality assurance. Finally, an simulation experiment will validate our framework and show its performance in zero-defect manufacturing.