二维扩压叶栅流场的数据同化研究

Research of Data Assimilation on Two-dimensional Compressor Cascade Flow Field

为提高二维扩压叶栅流场的预测精度,构建了一套基于集合卡尔曼滤波方法的适用于扩压叶栅内流流场的数据同化方法,对MAN-GHH扩压叶栅不同工况流场进行了数据同化研究,通过校正流场数值模拟的来流边界条件和S-A湍流模型的系数,获得了与叶栅风洞测量数据高度相符的流场。结果显示:数据同化方法可以使预测偏差减小60%以上;在各种工况,边界条件校正在叶栅流场数值计算中十分必要;对于多数工况,S-A湍流模型预测出的叶片吸力面尾缘分流泡尺寸过大,湍流模型系数修正可以改善流动分离预测效果,提高预测精度;校正后的湍流模型系数呈现一定的规律性。

To predict the flow field of two-dimension compressor cascade accurately,a data assimilation framework for the compressor cascade inner flow field based on ensemble Kalman filter algorithm was performed.The framework has been applied on the flow fields of MAN-GHH compressor cascade at different working conditions.By correcting coming boundary conditions and coefficients of S-A turbulence models,the numerical simulation flow fields which are highly consistent with experiments measurements from the cascade wind tunnel were acquired.The results show that:the data assimilation can reduce the perdition error more than 60%;it is necessary to correct the coming boundary conditions;for most working conditions,the flow separation bubbles were over-predicted by the S-A turbulence model and coefficients correcting can improve the accuracy of flow separation prediction;the corrected turbulence model coefficients have some regularity.