论文部分内容阅读
在标准径向基函数(RBF)神经网络模型的基础上发展了带输出反馈的RBF神经网络。将计算流体力学(CFD)方法计算的时域气动载荷作为输入信号,建立跨声速非定常非线性气动力模型,并进一步运用CFD方法验证模型的精度。算例表明带输出反馈的RBF神经网络较标准RBF神经网络精度更高,能更准确描述跨声速激波大幅振荡时的非线性和非定常特性,并可推广用于多自由度运动的动态非线性气动力建模。用多级信号训练,预测简谐信号输入下的气动力算例表明带输出反馈的RBF神经网络能够预测不同振幅、不同频率的信号激励下的非线性气动力。
Based on the standard radial basis function (RBF) neural network model, RBF neural network with output feedback is developed. The time-domain aerodynamic loads calculated by computational fluid dynamics (CFD) method are taken as input signals, and a transonic unsteady nonlinear aerodynamic model is established. The CFD method is used to verify the accuracy of the model. The results show that the RBF neural network with output feedback has higher precision than the standard RBF neural network, which can more accurately describe the nonlinear and unsteady characteristics of the transaxial shock wave with large oscillation and can be used to promote dynamic non-stationary Linear aerodynamic modeling. The results show that the RBF neural network with output feedback can predict the nonlinear aerodynamic forces under the excitation of signals of different amplitudes and frequencies.