Research On The Key Technologies In Fusion Modeling And Optimization Of Complicated Models For 3D Printing

In recent years,3D printing has become an important manufacturing method with its rapid development.It has signification impact on the modern manufacturing industry.3D printing can achieve rapid product prototyping which meets the requirements of customized product design and manufacture.Compared with the traditional manufacturing,3D printing almost has no limitation for the shape of the model and makes it possible to manufacture the complicated models.However,more and more requirements arise for the modeling technique of 3D printing.It requires rapid design without considering the requirement of the shape for traditional manufacturing.However,these designed models may not be printed directly,and need further processing and optimization.Model construction and optimization for 3D printing have been a research hotspot for 3D printing in recent years and are valuable to research.In this paper,some key technologies in fusion modeling and optimization of complicated models for 3D printing are studied.Some new algorithm is proposed and a relatively complete process for 3D printing is formed.These key technologies include 3D model construction,model partition printing,model light-weighting and support structure generation.The main research of the dissertation is as follows:1.Two kinds of model construction methods are studied: mesh fusion technique based on component and mesh Boolean technique.For the mesh fusion technique,an improved Discrete Exponential Map parameterization method is adopted to parameterize the local meth.A Laplacian deformation technique based on the level weight is presented to reconstruct the component model.For the mesh Boolean technique,floating point arithmetic errors and singularity of intersections are analyzed to guarantee the unique intersection between a segment and a face,and the continuity of intersections,which can guarantee the robustness of the Boolean.2.For the problem that the size of model is larger than the printer,a model partition method which is based on the skeleton is presented.The method can guarantee the integrity of the meaningful parts of a model.A mesh contraction method is used to extract the skeleton of a 3D model.Then,the 3D model is partitioned into many smaller parts using the space sweep method and the minima rule.The preliminary partition is optimized using the greedy algorithm.During the partition process,printability,symmetry,over-constraint and assemblability are considered.Finally,a method based on the harmonic distance field is proposed to design the connectors.3.For the problem of light-weighting problem,a novel algorithm which is based on the weakly balanced octree technique is proposed.The Euler-Bernoulli assumption is used to analyze the structure of 3D model and detect the weak parts and the stress distribution of the object.An adaptive filling method is presented to generate different filling density for different places of the object,which matches the stress distribution.For the adaptive filling,the Possion-sample which is based on the Monte Carlo theory is used to implement the adaptive sampling on the surface of an object in terms of the stress distribution.Then the weakly balanced octree is used to partition these sampled points to realize the filling result.4.For the problem of support structure generation,this paper presents a smart support structure generation algorithm which is based on the principle of minimum volume.Some approximate directions of minimum volume are calculated using a spherical coordinates system.The final direction is obtained by the derivative-free optimization method which optimizes these approximate directions.After the best printing direction is obtained,a tree-like shape support structure is constructed,and a stable support strut is designed.