Reverse Engineering of Geometrically Complex Automotive Structures Using X-Ray Computed Tomography and Digital Image Based Finite Element Methods 981193
Stress analyses of complex automotive components can be nearly impossible to achieve due to extreme difficulties in generating a realistic finite element model. A digital image based finite element approach was used to generate a 3-D finite element model from computed tomography (CT) scans of two automotive transmission cases. For the first case, original CT slices of 1024x1024x208 provided by ARACOR Inc. (Sunnyvale, CA) were used to generate a 3-D finite element model containing nearly 400,000 8-node brick elements. For the second case, 770x870x759 CT slices were used to generate a 3-D finite element model containing approximately 650,000 3-D elements. The mesh data generation from CT data for both cases took 6 minutes each on an engineering workstation. The resulting finite element meshes were analyzed using a specially designed finite element equation solver. Approximate solution time was 23 hours for the first case and 66 hours for the second case, again on a standard engineering workstation. Von Mises stress results for both cases under a wide open throttle reverse loading condition showed stress concentration in regions that correlated with experimental results. This analysis demonstrates the ability of a digital image based finite element approach to realistically analyze complicated components. When combined with CT data, the digital image based FE system can analyze already manufactured parts for comparison with product design data. The maximum total digital FE analysis turnaround time of 3 days including scan time represents a tremendous time savings over traditional finite element meshing techniques which, even if possible, often require an average turnaround time of over 90 days.
Citation: Hollister, S., Guldberg, R., Kikuchi, N., and McGee, R., "Reverse Engineering of Geometrically Complex Automotive Structures Using X-Ray Computed Tomography and Digital Image Based Finite Element Methods," SAE Technical Paper 981193, 1998, https://doi.org/10.4271/981193. Download Citation
Scott J. Hollister, Robert E. Guldberg, Noboru Kikuchi, Robert M. McGee
Psiphics Technologies, Inc., Ford Motor Company