基于CCA与POD的气动光学畸变分析

摘要

The methods of canonical correlation analysis（CCA） and proper orthogonal decomposition（POD） were used to study the 3D density field data computed by computational fluid dynamics（CFD）.2D-CCA was adopted to extract the canonical correction features between two cross sections along the spanwise,and then POD was adopted to eliminate the redundancy information between two cross sections.In this way,the features containing the domain information were extracted from the original data.The extracted feartures are then used to reconstruct the 3D density field,and the simulated corresponding aero-optical effect distortions gives the evidences that light transmission in reconstructed field is similar to the original field.%利用典型相关分析和本征正交分解方法,对计算流体力学得到的导弹红外窗口绕流三维密度场数据进行研究.取三维密度场沿展向两个相邻截面的数据先进行典型相关分析提取出两个截面的相关性最强部分,再利用本征正交分解消除两个截面之间数据的冗余信息,获得了能够包含两个截面密度场主要的信息部分.利用提取出的数据重构流场,并进行气动光学畸变仿真分析,仿真结果能较好地近似原始流场的情况.