Efficient Data Processing Technology Of Large-scale Complex Models For Additive Manufacturing

With the continuous application of additive manufacturing(AM)in aerospace,military and automotive fields,additive manufacturing parts tend to be large and complex.Large-scale complex mesh models often have a data scale of tens of millions of triangular patches,and the corresponding mesh file can easily reach the size of Giga Byte,resulting in the inefficiency or inability of data processing in AM software?These problems and shortcomings are more severe in engineering applications,and has become one of the bottlenecks in the development of big area additive manufacturing.In order to solve the problems,the basic data structure models,tree support generation method,segmentation slice contour generation approach,etc.are studied in this dissertation to improve the data processing efficiency of large-scale complex model,covering the model preprocessing,support generation,model slicing and other contents in AM data processing.These research contents were systematically integrated in virtue of the proposed service-oriented architecture of AM data processing software.The research work of this dissertation mainly includes the following aspects.The basic data structure models are proposed and established to improve the data processing efficiency of large-scale complex mesh model,which contains mesh-volume hierarchy model and vertices-mesh hybrid model.The basic data structure models are constructed based on the idea of voxel model and divide&conquer strategy,which can realize the rapid retrieval of local patch shape information and topology information from complex model under the premise of low spatial complexity.In virtue of reducing the data scale of triangular patches in mesh model processing,the basic data structure models bring us good property of high efficiency in searching procedure of feature triangles and feature vertices.A local-barycenter based tree support generation approach is proposed to improve the efficiency of tree support generation for large-scale complex mesh models and reduce support material consumption.The idea of divide&conquer strategy is adopted to construct tree support structure hierarchically,which has a time complexity of(9)log9)),and it is better than the violent searching tree support method with(9)~2).By analyzing the stability characteristics of tree support structure,a new tree node generation method based on local-barycenter is proposed,which improves the stability of the tree support structure and reduces the consumption of support materials.An umbrella intersection method based on mesh-volume hierarchy model is proposed to improve the efficiency of intersection vertice between tree structure and mesh model,and it has a time complexity of(1),avoiding the traditional searching process base on GPU(Graphics Processing Unit)deep cache method.In the test of large-scale complex industrial propeller model,the execution efficiency of the proposed method is290 times higher than that of meshmixer software that adopts violence search tree support generation method.The feasibility of the proposed tree support structure generation approach in Fused Deposition Modeling(FDM)industrial big area AM was demonstrated.A region-based segmentation slicing contour generation method is proposed to improve the efficiency of segmentation slice contour generation for large-scale complex models,and is suitable for big area AM with multi lasers.The idea of divide&conquer strategy and the proposed vertices-mesh hybrid model are adopted to guarantee acquiring the segmentation slice contours results from complex models directly and efficiently,and it has a time complexity of(9)log9)).With regard to the slicing results of these incomplete meshes in local segmentation grid units,a trimming-mapping strategy is proposed to solve the ordered-pairing problem of the end-points of slicing contours located on the square edges.A square based open-contour closing method is proposed to overcome the closure ambiguity of trimmed contours.In the test of large-scale complex mesh models,the execution efficiency of the proposed method is about7?10 times than that of the conventional first-slicing-then-decomposition method,and the time saving is about 86.35%?90.56%.The feasibility of the proposed method in Selective Laser Sintering(SLS)industrial big area AM was verified.Finally,in virtue of the proposed service-oriented architecture of AM data processing software,these research contents of software modules are integrated into a basically complete AM data processing software platform,which is applied to the hardware platforms of Huake 3D's full series AM equipments and self-developed big area AM equipment,and supports the mainstream AM processes,such as FDM,SLS and selective laser melting(SLM),etc.The above research results have practical significance for improving the efficiency of AM data processing for large-scale complex models.