Computers and Industrial Engineering
| Hi-index | 0.00 |
To reduce process variability in complex manufacturing processes, a tremendous need exists to integrate monitoring process variability (PV) and identification of source of out-of-control signals (SOS). The advent of advanced measurement and information technology has provided promising opportunities to improve product quality. In this paper, one integrated system is explored for intelligent monitoring PV and identifying of SOS in multivariate manufacturing processes (MMPs). |S| control chart is used as the detector of abnormal signals and an improved particle swarm optimisation with simulated annealing-based selective neural network ensemble (PSOSAEN) is explored for identifying the SOS. The seamless integration of control chart and PSOSAEN provides abnormal warnings, reveals SOS and helps operators to take some necessary corrections and adjustments. A real application is illustrated to validate the usefulness and effectiveness of the developed integrated system. The analysis results indicate that the developed integrated system can perform effectively for monitoring and classifying variance increases. This study provides guidelines for developing integrated neural network ensemble-based multivariate statistical process control identification systems in MMPs.