Functional health prognosis approach based on extended quality stochastic flow network for intelligent manufacturing systems considering robustness

Abstract

The core function of intelligent manufacturing systems (IMS) is to produce products meet quality requirements stably. However, variations from 5M1E (Man, Machine, Material, Method, Measurement, and Environment) are fundamental factors interfering with operation of IMS, which poses challenge to ensure functional health. Existing studies have neglected impact of IMS robustness on functional health prognostic approach. Therefore, functional health prognosis approach for IMS considering robustness based on extended quality stochastic flow network (EQSFN) is proposed in this study. First, connotations of functional health and IMS robustness are defined. Second, EQSFN model is presented which describes relationship among process scheme, machine, work-in-process (WIP), and task requirements. Third, quantitative model of process parameters robustness is established, which takes into account effects of machine degradation, WIP variation and task requirement change on IMS robustness by combining IMS variation with signal-to-noise ratio and quality loss function. Fourth, integrated approach for functional health prognosis of IMS, which considers state of process scheme robustness, production machine, WIP quality, and task execution, is proposed. Finally, ferrite phase unit is used as case study to validate method. Results show proposed method reduces health overestimation by up to 21.6% and improves average prediction accuracy by 17.8%.