Fuzzy-inference-based decision-making method for the systematization of statistical process capability control

Abstract

This paper proposes a decision-making method based on fuzzy inference to facilitate process capability analysis based on the knowledge and experience of experts, and implement systematized statistical process capability control. Data screening is implemented in the form of a rule-based decision-making tree to perform normality testing, R- or s-control chart testing, and x-bar control chart testing on process data to determine whether a process is in a state of statistical control. After setting the improvement direction of the process using a four-block diagram, the processes with a high probability of defect leakage due to large dispersion compared to the specification are reexamined after the fundamental improvement is completed by reinforcing the technology. Additionally, an optimal process capability index is selected using fuzzy inference by considering the degree of bias in a distribution and the differences between short- and long-term process capabilities. The feasibility of the proposed method was verified by applying it to a process for manufacturing home gas boilers. The method enables process control engineers to examine the results of the statistical analysis and the priority of the process to be improved, which are visualized in real time using a dashboard. These results are subsequently used for decision-making.