A quality-driven stability analysis framework based on state fluctuation space model for manufacturing process

Abstract

Industrial manufacturing processes often show multiple operating modes, where different modes present different regularities, so real-time monitor and analyzing the quality state stability is an important way to ensure product quality. This paper proposes a state-driven fluctuation space model for quality stability analysis for multimode manufacturing process. First, the whole process is divided into many sub-processes and the multimode formation mechanism is analyzed to form the stability analysis framework. Then each single-mode quality state fluctuation space model is built based on multi-kernel support vector data description method to determine the max effective fluctuation border of the process state. For the current process state, the deep neural network (DNN) is adopted to extract process state features automatically and recognize the mode type. Thus appropriate quality stable fluctuation space model is selected to monitor and analyze the process stability state. Finally, a case study is performed to evaluate the feasibility of proposed stability analysis method, and the result reveals that the method shows good effect for analyzing the process stability in manufacturing process.