Remaining useful life prediction of manufacturing system based on fuzzy Quality State Task Network

Accurate prediction of operational state and remaining useful life (RUL) is the premise for PHM (Prognostic and Health Management) implementation and predictive maintenance decision of intelligent manufacturing system. The quality state of the produced WIP (work-in-process) is also a core indicator of the RUL of the running manufacturing system, which can have an impact on the degradation of the downstream machine. However, few studies have fused the quality data of WIP(work-in-process) into the RUL prediction considering the fuzzy operational mechanism of the manufacturing system. Therefore, an RUL prediction method of intelligent manufacturing systems based on the fuzzy Quality State Task Network (QSTN) model is proposed. First, the principles of the RUL prediction of manufacturing system is expounded by taking fuzzy mission reliability as the indicator of system health state. Second, a fuzzy QSTN model that can simultaneously characterize the performance degradation of manufacturing equipment and the quality degradation of WIP is established. Third, a dynamic mission reliability and remaining useful life prediction method of the manufacturing system based on dynamic Bayesian networks is proposed. Finally, the effectiveness of the method is verified by taking the continuous stamping process of a bogie manufacturing system as an example.

Automated summary

This study proposes an RUL prediction method for intelligent manufacturing systems based on the fuzzy Quality State Task Network (QSTN) model. By considering the fuzzy mission reliability indicator and the degradation of manufacturing equipment and WIP quality simultaneously, the accuracy of RUL prediction is improved.