Dimensional analysis and process control of body-in-white processes.

The significance of the body-in-white in U.S. automotive manufacturing is profound. Stamping dies and assembly tooling can represent half the cost of developing a new car. Significant quality-related costs can be traced to rework, scrap, and idle equipment. In engineering and manufacturing processes, improved statistical methods for product specifications, dimensional analysis, and process control can reduce costs while enhancing quality. The objectives of this research were threefold: to identify and examine the systems that contribute to dimensional control; to quantitatively develop, using a total-system perspective, areas lacking in numerical analysis; and to recommend a general systems approach for achieving dimensional process control. The first research step benchmarked dimensional control systems at companies in Europe, Japan, and the U.S. Second, part measurement data was collected and analyzed, and statistical tools developed. Third, recommendations for dimensional process control were made. The benchmarking process noted several differences between companies that concerned design philosophy, die tryout, process preparation, and process control. These are reflected in the recommendations. A statistical analysis of stamped part measurements estimates the mean-shift and part-to-part variation. The two standard deviations are approximately equal, with an average of 0.2 mm. Natural part tolerances range from 0.3 to 1.5 mm. Out-of-control processes have variation dominated by mean shifts. A tolerance analysis model was developed for stamping and assembly, combining a mean-shift model with regression equations. The model can be used to evaluate process capability prior to production. Several system recommendations are made. One is to use the functional-build approach where possible because it increases the average design capability index 133%. This philosophy defines quality based on the functionality of the assembly, rather than requiring that each individual part meet specification. Another is to match the process mean with the specification nominal to increase interchangeability. Also, establish an operating envelope for process variables to improve process control. Finally, follow a root-cause problem correction approach to improve long-term process stability. This requires a system infrastructure that includes a supporting information system, standardized procedures, and inventory control disciplines.PhDIndustrial and Operations EngineeringUniversity of Michigan, Horace H. Rackham School of Graduate Studieshttp://deepblue.lib.umich.edu/bitstream/2027.42/103181/1/9308270.pdfDescription of 9308270.pdf : Restricted to UM users only