| Automobile bodies must be built to have high dimensional integrity at competitive costs in the current environment of increasing competitions. To assure the dimensional integrity of the automobile body, design engineers always assign tight tolerances, which will cause higher cost. Manufacturing engineers, on the contrary, prefer loose tolerances, which make production easier and less expensive. Since excessive cost or poorer quality will eventually result in a market share loss, a balance between a good quality as well as a sensible production cost has become a major task for researchers to work on. Currently, sheet metal parts are assumed to be rigid bodies when conducting tolerances stack-up analysis. And the addition theory of variance is applied in accumulating the tolerances of components. The common experience found in auto body industries, though, does not agree with the results from the tolerance analysis based on rigid bodies. In fact, the tolerance analysis results always over-estimate the tolerance stack-up for compliant sheet metal assemblies, while they always under-estimate that for closure panel settings.; The objectives of this research are to develop a methodology for (i) modeling compliant sheet metal assemblies and panel settings more accurately, (ii) predicting tolerances of assemblies more effectively, and (iii) improving robustness of panel setting processes. A systematic approach has been developed to conducting tolerance analysis for sheet metal assemblies and panel settings in auto body assembly by considering (i) the rigidity evolution of compliant sheet metal assembly and (ii) the kinematic effect in panel settings. The major accomplishments are summarized as the followings: (1) Characterization of dimensional variability and rigidity evolution for auto body assembly, (2) A methodology is developed for tolerance analysis modeling of compliant sheet metal assemblies, (3) A 3-D tolerance analysis modeling technique is developed for rigid body assemblies, and (4) A framework is proposed for robust process design. |
