| For many years, corporations have poured millions of dollars into computers, robots, and other gee-whiz machines without fully understanding how these things make their companies more productive and profitable. A good example of this case is measurement systems in the automobile industry. As new technology emerges and old technology continues to be used, the utilization of several unique measuring systems can produce differing results, even when measuring the same part. Therefore, clarification of systematic analyzation among different measurement systems is required to effectively use these systems. Systematic methodologies which are proposed in this thesis extend the concept of conventional accuracy and R&R (repeatability & reproducibility) studies and also provides a new methodology for compatibility studies of more than two measurement systems. These systematic methodologies are applied to reduce the great amount of time needed to calibrate and analyze each measurement system.;A new systematic approach is represented for the improvement of the dimensional control process of an automotive body-in-white (margin, flushness, gap, etc.). By having availability of in-line 100% measurement data, the traditional assumption of independent variables frequently appears to be violated. The concept of time series has been used to remove the autocorrelation in the data for reducing false alarms when data is plotted on conventional control charts. Finally, this thesis also examines an application of multivariate quality control methods to the manufacturing of automotive body-in-white. |
