Approach to derive golden paths under time-varying machine performance in multistage manufacturing process

Abstract

In a multistage manufacturing process, identical machines are utilized at each process stage. According to such configuration, each product can go through a different process path (hereafter referred to as path) during production. In practice, however, machines at a process stage may have different operational performances, and thus the quality of the final product may vary depending on the path. This research proposes an approach to derive the golden paths (GPs), which are expected to produce products of higher quality than the user-defined quality level, when the machine performance degrades over time. To this end, the health indicator (HI) of a machine, which represents its performance at a certain time, is introduced to reflect the performance degradation. In deriving GPs, the proposed approach discovers a specific set of machine sequence patterns (MSPs), sequential combinations of machines in different process stages, satisfying several rules. Based on the discovered MSPs and the HI information, the proposed approach generates the GPs at a particular future point in time. The viability of the proposed approach is evaluated through simulated experiment. The derived GPs reflecting machine performance degradation over time are expected to contribute in increasing the production of superior quality products and reducing the re-production costs incurred by utilizing non-GPs.