渦構造の抽出による非定常流れのデータ圧縮について

概要

論文の詳細を見る
Recent developments of numerical techniques and computer system have enabled us to obtain a highly accurate CFD results for unsteady flow fields. However, we have some difficulties to utilize such results in a design process due to the huge amount of data. Major problem is a handling of those data in comparing with other results, archiving in a database and exchanging among the researchers. In our study, a data compression method using a characteristic feature of a flow field will be considered. Firstly, the flow field is visualized using some display-based techniques such as a volume rendering method for a scalar field and LIC (Line Integral Convolution) or PEM (Pixel Exposure Method) for a vector field. And then, a vortical structure is interactively abstracted from the flow field. In a two-dimensional case, some point-vortices are substituted for the distribution of the vorticity. Vortex sticks will be utilized in three-dimensions. Those vortex elements represent the flow fields, and are stored for the comparisons, the archives and the exchanges. It will be expected that the size of data is extremely reduced without standard techniques of data compression in computer science. In this paper, we will propose a methodology of our data compression method, and demonstrate the method for a two-dimensional incompressible flow field.
社団法人日本船舶海洋工学会の論文