基于POD和ARMA的阵风响应预测降阶模型研究

Study of Reduced Order Modeling Based on POD and ARMA for Prediction of Gust Response

针对阵风响应流场高效精确分析需求,基于非定常CFD数值模拟发展了一种快速流场预测的降阶模型。采用本征正交分解(POD)方法对高斯阵风激励的CFD非定常流场快照进行特征分析、并基于能量占比准则截取主要基模态来降维,采用自回归滑动平均(ARMA)方法建立各主要基模态系数的线性预测模型。采用NACA0012和DLR-F12的分析算例进行了验证,与CFD非定常计算结果对比分析表明,模型具有可靠的计算精度,适合于精确高效快速分析阵风响应,为阵风减缓设计提供技术支撑。

Aiming at the need for efficient and accurate analysis of the flow field in the gust response,a reduced order model for rapid flow prediction is developed based on unsteady numerical simulation of CFD.The proper orthogonal decomposition is used to analyze the characteristics of flow field snapshots calculated by CFD,and the main basis modes are intercepted based on the energy criterion to form the reduced order subspace.The linear prediction models of each basic modal coefficient are established by auto regressive moving average method,and the gust response analysis study on NACA0012 and DLR F12 is used to verify it.Comparative analysis with CFD unsteady calculation results and prediction results shows that the model has reliable calculation accuracy,is suitable for accuracy,efficient and rapid analysis of gust response,and can provide technical support for gust design.