Research On 3D Iso-surface Construction Method Of Visualization In Scientific Computation

Large data sets are commonly obtained in models based on numerical computations in the field of science and engineering.Scientific computation visualization has become an indispensable technique with the aim of illustrating the characteristics of data sets and revealing their significance by using figures and graphics.The design of complex products in practical engineering integrates the computational analysis of various heterogeneous software in the multi-disciplinary field.Relevant scientific computation visualization is thus required to manipulate heterogeneous data sets in various states of different conditions.This multi-disciplinary visualization usually includes the visualization of finite element calculation results and the visualization of data sets that are provided by the measuring and scanning devices.The key point of scientific computation visualization lies in the visualization of three-dimensional data sets.In particular,three-dimensional iso-surface construction technology is required to provide various data sets with various solutions.In traditional three-dimensional surface construction methods,data sets that are obtained from finite element analysis possibly result in ambiguity.Besides,a higher-order or refined mesh can also yield inaccurate results.It is one of the most crucial problems that ensures the continuity among complex free-form surfaces when the data origins from a discrete data set.This thesis is based on the scientific calculation visualization in the design process of complex products.It investigates the characteristics of data that obtained from finite element analysis results and discrete data set respectively.It aims to generate three-dimensional surface effectively and efficiently with the'computing-building-storage-optimization' research strategy.The data distribution characteristics of tetrahedral and hexahedral meshes have been investigated respectively,and the complex surface construction of discrete data has been studied correspondingly.This thesis proposes a series of iso-surface construction technique for mesh elements(8-node first order hexahedron,10-node second order tetrahedron,20-node second order hexahedron)and the splicing technology for complex surfaces.These techniques solve the computational redundancy that is caused by various ambiguities in scientific computation,and achieve the splicing that guarantees the continuity among complex surfaces based on discrete data sets.The major contributions of this thesis are summarized as following:A method for iso-surface construction based on auxiliary points of diagonal plane has been proposed.This method calculates key geometry parameters that include equivalent points and key points using the shape function of finite elements.The equivalence points are tracked by involving gradient and geometrical significance,and auxiliary points of diagonal plane are used to avoid ambiguity of results.It controls the accuracy of iso-surfaces and the number of triangles by adjusting the step length and merging threshold.This proposed method has successfully solved the three dimensional iso-surface construction problem based on first order hexahedral unit.An Iso-surface construction method based on key point and compression index for two-order grid meshes has been proposed.This method simplifies the interpolation function of surfaces and edges for second order mesh elements,and the relationship matrix is proposed herein.Moreover,it avoids ambiguity by adding key points and solves the the problem of rotary Iso-surface exists in the same element.The triangulation process is implanted into the key point extraction process to get patches and calculation results at the same time.The design optimization repair strategy is presented to satisfy the needs of image accuracy.The geometric position and the logic relationship of the equivalence points extracted is analyzed,and the compressed index structure of triangles and related strategies of storage is designed for splitting and optimization to present low cost of storage and computation.A method that splices triangular Bernstein-Bezier surfaces with approximate geometric continuity has been proposed.This method achieves the splice that is capable of guaranteeing approximate G0,G1 geometrical continuity by adjusting existing control points and modifying existing surfaces.This newly proposed method is not constrained by factors that include the order,number and splicing positions of surfaces.It avoids complex operations(e.g.degree elevation etc.)in splicing process and guarantees the visual effect.A solution that aims to solve the heterogeneous data problem in pre-processing has been proposed.This method also achieves the intelligent analysis of heterogeneous data files.A visualization system with human-computer interaction system has been devised.An integrated multi-software platform that bases on parameter analysis has been constructed.This platform achieves the interaction and access among heterogeneous data files.Besides,it provides various costumed types of data input and output operation in the human-computer interface and the visualization of results.