Tolerance Simulation Optimization with Data Analysis

Abstract

This thesis study primarily focuses on simulation chain accuracy through data-driven optimization techniques within the BIW (Body In White) assembly. Geometry assurance plays an important role in maintaining the quality throughout the product’s lifecycle by managing the deviations that originate from manufacturing and assembly processes. In complex assembly scenarios, especially within BIW Assembly that involves mixed materials and varying structural rigidity, the simulation accuracy becomes challenging, and it requires a better understanding of existing models. The primary objective is to refine the simulation models that predict geometric variations to enhance the predictive capability. To achieve the objective, a detailed correlation study has been conducted, analyzing and comparing the real-world measurement data with rigid and non-rigid simulation models, assessing the effects, like process adjustments. The comparison study between the production data and tolerance simulation model data is used to derive root cause analysis within each sub-assembly. Based on this study, improvements were made within existing rigid simulation models and suggested optimized simulation guidelines. This research aspires to find the major contributors to the geometrical variations within the BIW Upper body assembly and also to bridge the gap between simulation models and real-world data, improving the robustness within the rigid body simulation.