Conversion of 3D scanned point cloud into a voxel-based representation for crankshaft mass balancing
Résumé
Controlling the part's balance of crankshafts is an important issue for automobile manufacturers. Unbalance measurement is carried out, in general, with high-precision mechanical balancing machines. Each model of crankshaft requires a specialized mechanical balancing machine. We propose a new solution for balancing crankshafts by replacing mechanical measuring systems with a non-contact digitizing system, which allows the acquisition of a representative 3D point cloud of the crankshaft surface. The Voxel2inertia method is proposed to convert 3D scanned point clouds into a voxel-based representation and to calculate the inertia matrix and the unbalance vectors. Theoretical CAD models are used to determinate the best setting for mesh and voxelization. Several levels of error are introduced on CAD models in order to estimate the necessary scanning quality, and thus define the appropriate scanning system for balancing measurements. The Voxel2inertia algorithm is validated on three different crankshaft models that were scanned with the most appropriate digitizing system. The results highlight that this new approach for determining unbalance vectors from scanned data is suitable for replacing the current mechanical measuring systems. This technology can make the production lines more flexible and more adaptable for future changes.
Origine : Fichiers produits par l'(les) auteur(s)
Loading...