Manufacturing systems typically contain processing and assembly stages whose output quality is significantly affected by the output quality of preceding stages. The deficiencies of using standard statistical process-monitoring procedures in such systems have been highlighted in the literature. This article proposes a procedure to monitor process and product quality in multistage systems. By accounting for the quality of the input to each stage, the procedure not only detects the presence of out-of-control conditions but also helps to identify the stages responsible for such departures. We extend previous research to the common case where the process parameters are unknown. An extensive performance study shows that the procedure is effective in detecting out-of-control conditions and that it convincingly outperforms existing methods. We illustrate the use of the procedure using production line data from a major electronics manufacturer.
†Deceased
Acknowledgements
The authors thank the anonymous referees and the Department Editor for several constructive comments and suggestions that have improved this article. This research is partially supported by grants from the Dauch Center for the Management of Manufacturing Enterprises at Purdue University, the Basil Sidney Turner Foundation, and the Purdue Research Foundation.
Notes
†Deceased
*Statistically significant at the 5% level.
**Statistically significant at the 1% level.
a Computed using the ANYARL program of CitationHawkins and Olwell (1998).