基于随机森林算法的非定常气动力建模研究

针对现有的非定常气动力建模方法对气动弹性预测的准确性和效率问题,将随机森林算法引入非定常气动力建模研究领域,构建了基于随机森林算法的非定常气动力降阶模型。将所得模型用于预测气动弹性,选择二维NACA0012翼型进行颤振边界的预测,选用NACA64A010翼型预测LCO特性,并说明了该降阶模型建模的详细过程,将其计算结果与CFD/CSD耦合计算结果及试验结果进行了对比。研究结果表明,该模型可行、高效且精确,可以快速准确地预测飞行器气动弹性特性。

Research on unsteady aerodynamic modeling based on random forest algorithm

Aiming at the accuracy and efficiency of the existing unsteady aerodynamic modeling method for the prediction of aerodynamic elasticity,the random forest (RF) algorithm is introduced into the field of unsteady aerodynamic modeling,and the unsteady aeroelasticity reduction based on random forest algorithm is constructed,and the resulting model is used to predict the aerodynamic elasticity.The two-dimensional NACA0012 airfoil is used to predict the flutter boundary and the LCO characteristics of NACA64A010 airfoil.The results of the RF/ROM method are compared with the CFD/CSD coupling calculation results and the test results.It is found that the RF/ROM is feasible,efficient and accurate.The method can predict the aerodynamic elastic properties of the aircraft quickly and accurately.