State entropy–based fluctuation analysis mechanism for quality state stability in data-driven manufacturing process

Abstract

Intelligent quality state analysis is a promising tool to deal with manufacturing big data due to its ability in efficiently processing state signals and providing accurate warning results. Inspired by the idea that uses the change of entropy flow to characterize the quality state change, this article proposes a fluctuation analysis mechanism for quality stability based on state entropy in data-driven manufacturing process. First, the multidimensional space cloud model with a three-tuple feature is constructed to describe quality state fluctuation in which the digital features of entropy and hyper-entropy represent the fluctuations’ uncertainty of quality state. Furthermore, in order to quantitatively analyze the fluctuation degree of process state, the entropy change mechanism is introduced into the manufacturing quality state to calculate the state fluctuation degree. The proposed method is validated by a fan blade machining process dataset, and the result shows that the approach could well monitor the quality state fluctuation and show good effect for process stability analysis, which will provide theoretical evidence for the real-time warning and evaluation for abnormal quality state in manufacturing process.