Approximating Material Interfaces During Data Simplification.

Signeti, D. E. ; Gregorski, B. F. ; Amrosiano, J. ; Graham, G.

Technical Information Center Oak Ridge Tennessee

关键词: Computational grids;    Interfaces;    Topological cells;    Algorithms;    Data;   

RP-ID  :  DE200515013473

学科分类：工程和技术（综合）

美国|英语

来源: National Technical Reports Library

【 摘 要 】

We have presented a simplification method for scientific data sets that explicitly represents material interfaces in mesh cells. Our algorithm constructs an approximation that can be used in place of the original data set for visualization purposes. Explicitly representing the material and implicit field discontinuities allows us to use multiple field representations to better approximate the field within each cell. The use of the tetrahedral subdivision allows us to generalize our algorithm to a wide variety of data sets and to support interactive level-of-detail exploration and view-dependent simplification. Future work will extend our error calculations to support complex input cell types such as tetrahedra and curvilinear hexahedra. Our current ghost value computation assumes that the field is constant on the other side of the interface. Higher-order extrapolation methods should be investigated for ghost value computation to determine if a superior field approximation can be obtained. Similarly, material interfaces are defined by approximations based on linear functions. The tradeoff between cell count and higherorder approximation methods should be investigated to determine if a better approximation can be obtained without a great increase in computational complexity. Finally, we plan to apply our algorithm to more complex unstructured data sets.