Statistical Error Analysis for Dimensional Control in Automotive Body Assembly Process

In mechanical assemblies, the performance, quality, cost and assemblability of the product are significantly affected by the geometric errors of the parts. This paper develops the statistical error analysis approach for dimensional control in automotive body multi-station assembly process. In this method, the homogeneous transformation matrices are used to describe the location and orientation of part and assembly features and the small homogeneous transformation matrices are used to model the errors. In this approach, the effective errors in automotive body assembly process are classified in three categories: manufacturing errors (dimensional and geometric tolerances), locating errors (fixture errors) and process errors (joining errors). In a mechanical assembly, small variations due the errors propagate according to a complex mechanism that in this approach it formulated in error analysis procedure. The propagation chain of geometric errors is described based on CAD models. The estimation of the error accumulation and the percent contributions of individual errors are based on the statistical model (root-sum-square method). The application of the proposed method is illustrated through presenting an example problem.