Skip to Main Content
In discrete manufacturing processes such as stamping, assembly, or machining processes, product quality, often defined in terms of the dimensional integrity of work pieces, is jointly affected by multiple process variables. During the production phase, the states of tooling components, which are measured by adjustable process variables, are subject to possible random continuous drifts in their means & variances. These drifts of component states may significantly deteriorate product quality during production. Therefore, maintenance of the tooling components with consideration of both their continuous state drifts as well as catastrophic failures is crucial in assuring desired product quality & productivity. In contrast to traditional maintenance models where product quality has not been well addressed, especially for discrete manufacturing processes, a general quality-oriented-maintenance methodology is proposed in this paper to minimize the overall production costs. In this research, the total production cost includes product quality loss due to process drifts, productivity loss due to catastrophic failures, and maintenance costs. The quality-oriented-maintenance model is built based on a response model linking process variables with multidimensional product quality. It can be obtained either from engineering analysis for specific processes, or from statistical design of experiments. Three typical multi-component maintenance models are investigated under the general quality-oriented-maintenance framework. A case study for a sheet-metal stamping process is presented to demonstrate the effectiveness of the proposed methodology.