Integrated predictive maintenance approach for multistate manufacturing system considering geometric and non-geometric defects of products

Abstract

The adverse consequences caused by product defects can intuitively characterize the abnormal performance state of manufacturing systems in advance. Defects can be divided into geometric and non-geometric defects according to the difference of expression characteristics. Compared with machine failure data, the various effects and economic losses caused by product defects are more conducive to guide the development and optimization of maintenance strategies, which are rarely discussed. Therefore, from the perspective of product defect information, a novel integrated predictive maintenance (PdM) strategy is proposed to improve the performance of manufacturing systems. First, the influences of product geometric and non-geometric defects on the performance of manufacturing system are expounded, and two types of reliability metrics and the corresponding connotation are proposed and applied to PdM activities. Second, the mechanism of product defects caused by machine degradation is analyzed, and the quantitative models of the reliability are established. Third, the expected maintenance investment and economic loss in the planning horizon are evaluated, and the optimal PdM strategy is proposed according to the principle of total cost minimization. Finally, an industrial example of PdM decision making for a ferrite phase shifting unit manufacturing system is given to illustrate the effectiveness of the proposed approach.