An Optimized Approach for Assembly Models in Non-Assembly 3D Printing

Printing an assembly model embedded with kinematic semantics in a high-quality and non-assembly manner is more and more popular. However, achieving this printing manner automatically and efficiently is still difficult, since it needs to carefully consider many problems, such as the removability of support structures, the consumption of material, and the quality of the surfaces concerning kinematic semantics. To account for this, an optimized non-assembly printing approach for assembly models is proposed.Firstly, the input assembly model is voxelized based on bounding box division and voxel association relationship. Secondly, the Bresenham algorithm is improved to rapidly generate voxel paths which are adopted to efficiently analyze the removability of support structures. Thirdly, an incremental method is presented to update voxel association relationships, which can avoid the time-consuming revoxelization for the input model when changing its parts' layouts. Finally, to determine the optimized printing orientation of the input model as well as its optimized parts' layouts, an optimization method is designed based on particle swarm optimization. Furthermore, the objective of the designed method is to simultaneously make all the support structures removable, minimize the material consumption in the support structure, and maximize the quality of the surfaces concerning kinematic semantics. Experiments are also carried out on several assembly models holding typical kinematic semantics. The experimental results show that printing each of the adopted models with the orientation and parts' layout provided by the above approach can ensure that less material will be consumed, all the support structures can be removed, and the roughnesses of the surfaces concerning kinematic semantics are also reduced greatly. Besides, compared with the method analyzing support structure removability based on octree and the method printing mechanism in a non-assembly manner by adjusting parts' layout, the proposed approach has outstanding applicability and advantage for printing the assembly models holding kinematic semantics in a high-quality and non-assembly manner.