An Optimization Framework for Fixture Layout Design for Nonrigid Parts: An Automotive Perspective

摘要

The inspection process of non-rigid parts during manufacturing and assembly is inherently challenging. This is exacerbated by the need for accurate real-time part data in the digital age. Although many ad hoc techniques exist, there are no rigorous methods to evaluate the quality of a fixture layout before final parts and gauges are available. This typically happens so late in the manufacturing process that errors found can scarcely be remedied. Additionally, the modifications to the gauge are usually costly and can result in significant delays, when performed this late in the process. This article proposes an optimization-driven mathematical approach tailored toward non-rigid parts to identify the best locator layout, early in the part design phase. A metric is proposed using robotic grasping theory to quantify the quality of the locating scheme and serves as the objective of optimization. The proposed method is implemented using a tolerancing software that performs finite element analysis (FEA) on the parts to predict its state given the force and torque inputs, including the effect of gravity. An evolutionary algorithm is used that optimizes the performance of the fixture layout. We demonstrate a significant improvement in gauge repeatability when compared with an arbitrary layout scheme on two design problems. The first problem is a representative design problem using a sheet metal part, while the second one is an actual automotive production part. General recommendations regarding fixture layout design for non-rigid parts are made, as well as directions for future work are provided.