选择离散型输入输出差分模型,运用最小二乘方法进行非定常气动力建模,并将辨识得到的降阶模型用于气动弹性的数值模拟。1个马赫数下的颤振临界点的计算仅需调用一次非定常流场求解器。计算精度保持与非定常欧拉方程计算方法相当的同时计算效率提高了1~2个量级。计算了跨声速具有S型颤振边界的气动弹性标准算例-Isogai wing和三维气动弹性标模算例AGARD445.6,辨识模型计算边界与非定常Euler方程计算结果吻合。证明非定常气动力辨识技术可以提供高效的高精度的气动弹性分析。
This paper aims at providing an accurate and efficient method for aeroelastic simulation. System identification is used to get the reduced order models of unsteady aerodynamics. LS(Least Square) method is used to estimate the coefficients of the input-output difference model. The reduced order models (ROM) are then used in place of the unsteady CFD codes for aeroelastic simulation. The efficiency is improved 1-2 orders while still retaining the accuracy of the complete unsteady CFD codes. Two standard aeroelastic computing examples (2D Isogai wing with S type flutter boundary and 3D transonic flutter example-AGARD445.6 wing) are computed. The computed flutter boundaries based on ROM are well compared with those of the directly coupling CFD/CSD method. The identification technology of unsteady aerodynamic loads shows a high efficient and accuracy approach for transonic aeroelastic solving.