Identification, monitoring, and diagnosis for dimensional control of automobile body assembly.
Roan, Chinmo
1993
Abstract
Because of the lack of diagnostic support, many attempts of statistical process control (SPC) fail to produce significant results. In this thesis, systematic methodologies are developed to improve the dimensional quality of automobile bodies by combining statistical techniques, decision-making rules, and the knowledge of manufacturing processes. These methodologies are used to utilize the information contained in the real-time 100% measurement data of automobile bodies for process identification, process monitoring, and root cause diagnosis. Process identification is used to determine measurement points with sustained large variations, group them based on their correlation coefficients and characteristic locations, and find the variation patterns for each group. It is developed based on clustering methods, correlation analysis, the hierarchical structure of the bodies, and principal component analysis. The variation patterns of each group are valuable information for root cause diagnosis because specific discrepancies in certain assembly stations will cause specific variation patterns. Process monitoring is used to detect sudden process changes and to monitor the dimensional relationships of the measurement points. Not only are the absolute positions of the measurement points critical to the dimensional quality of the bodies but also their dimensional relationships. An algorithm for univariate monitoring is developed based on statistical indices and decision-making rules to detect and classify sudden process changes and to group the points for root cause diagnosis. In order to monitor the dimensional relationships between points, index selection, simultaneous confidence interval, principal component analysis, and a classification technique are employed to establish algorithms for multivariate monitoring. Root cause diagnosis can be executed systematically based on the results of process identification and monitoring, knowledge of solved dimensional problems, and knowledge of the structure and of the assembly process. The results and knowledge are used as the essential information to localize problems and suggest the root causes of dimensional variations. The last part of the thesis develops interpretation charts relating the changing pattern of measurement data and that of the residuals after an adequate Autoregressive Moving Average (ARMA) model is used to filter the data because of the existence of autocorrelation. The interpretation charts can reduce misinterpretations for problem identification and root cause diagnosis by comparing the pattern of the residuals and that of the original data if a sudden process change has occurred (e.g. a mean shift or a sporadic jump).Subjects
Assembly Automobile Body Diagnosis Dimensional Identification Monitoring Process Control
Types
Thesis
Metadata
Show full item recordCollections
Remediation of Harmful Language
The University of Michigan Library aims to describe library materials in a way that respects the people and communities who create, use, and are represented in our collections. Report harmful or offensive language in catalog records, finding aids, or elsewhere in our collections anonymously through our metadata feedback form. More information at Remediation of Harmful Language.
Accessibility
If you are unable to use this file in its current format, please select the Contact Us link and we can modify it to make it more accessible to you.