Health state modeling for complex manufacturing system based on RQR chain and Hidden Markov Model

With the advent of Industry 4.0, the structure and function of the manufacturing system are becoming more and more complicated. Fault diagnosis and health management with predictive ability are the prerequisite for the effective and intelligent operation of the manufacturing system. Different from previous studies about the Prognostics and Health Management are always confined to the static modeling of the basic reliability of system components, a novel approach based on the big operational data and Hidden Markov Model is proposed in this paper. Firstly, the RQR chain is proposed to organize the big operational data, which includes the manufacturing system reliability (R) data, manufacturing process quality (Q) data and the produced product reliability (R) data. Secondly, a new concept of the health state of the manufacturing system is presented to highlight the operational performance of the production task. Thirdly, evolution of the key quality characteristics is adopted to fuse the big operational data, and the Hidden Markov Model (HMM) is used to model the health state of manufacturing systems. Finally, a case study of a manufacturing system for cylinder head is carried out to verify the effectiveness of the proposed approach. The final result shows that the proposed model has favorable dynamic modeling and prognostication capabilities.