基函数宽度对递归RBF神经网络气动力模型精度的影响研究  

Research on the Effects of Basis Function Widths of Aerodynamic Modeling Based on Recursive RBF Neural Network

作　　者：寇家庆 张伟伟[1] 

机构地区：[1]西北工业大学翼型叶栅空气动力学国防科技重点实验室,西安710072

出　　处：《航空工程进展》2015年第3期261-270,共10页Advances in Aeronautical Science and Engineering

基　　金：国家自然科学基金(11172237);新世纪优秀人才支持计划(NCET-13-0478)

摘　　要：由于非定常气动力的复杂性,通常所建立的气动力模型在稳定性、泛化能力和精度上均存在一定局限,采用递归RBF神经网络模型能够实现气动载荷的较准确预测。隐含层神经元的基函数宽度对该模型的精度及稳定性具有重要影响。首先通过数学分析和计算仿真研究训练过程中宽度与神经网络结构之间的关系,然后将NACA0012翼型俯仰运动作为算例,研究模型在不同训练信号、延迟阶数和流动状态下的性能,最后利用对随机俯仰运动样本的预测结果,验证宽度的最优选择范围。结果表明：基函数宽度对此类非定常气动力模型的稳定性及泛化能力影响较大;最优宽度的选择随训练及预测信号的变化有所不同;较多样本时,通常选择55-75的宽度能够保证非定常气动力模型具有较高的预测精度。Because of the complexity of unsteady aerodynamics, there are some limitations in stability, generalization capability and accuracy of the aerodynamic model which is usually established. Recursive radial basis function（RBF） neural network model is used to predict the aerodynamic load. The widths of basis function in hidden layer are one of important parameters to this aerodynamic model. To investigate the effects of widths for Recursive RBF neural network, mathematical analysis and simulations are executed at first, which shows the relationship between widths and framework of the model during training process. Then cases of NACA 0012 aerofoil with pitching maneuvers are simulated, to test the model＇s performance under different training maneuvers, delay orders and flow states. Finally, predicting results of random pitching maneuver verify the conclusion of best widths scale. Results show that the widths of basis function have much impact on the stability and generalization capability of this type of aerodynamic model. The best width varies with different training and testing maneuvers. With more samples, higher predicting accuracy of aerodynamic model is guaranteed with widths between 55-75.

关 键 词：径向基函数(RBF) 神经网络 非定常气动力 气动力建模 宽度 

分 类 号：V211[航空宇航科学与技术—航空宇航推进理论与工程]