直升机飞行力学线性变参数建模方法研究

Research on linear parameter varying modeling method for helicopter flight dynamics

针对直升机飞行力学建模问题,分别基于模型缝合及模糊逼近提出了两种线性变参数(LPV)建模方法。在模型缝合方法中,基于经过滤波加权的飞行速度进行气动导数和操纵导数的插值,完成气动力的实时计算,结合非线性欧拉运动方程完成LPV模型的建立。在模型逼近方法中,建立了基于T-S模糊模型的非线性状态空间逼近模型,采用anfis结构完成了模糊模型训练。利用UH-60直升机非线性模型生成了不同飞行速度的线性状态空间模型,并分别通过两种方法建立了两套LPV模型。最后,基于频率响应和泰尔不等系数对典型LPV建模参数对模型精度的影响特性进行了分析。

Aiming at the problem of helicopter flight dynamics modeling,two LPV modeling methods have been proposed based on model stitching and fuzzy approximation separately.In the model stitching method,aerodynamic and control derivatives are interpolated according to current flight speed weighted by a low-pass filter,then the aerodynamic forces and moments are calculated in real time.A LPV model is obtained based on nonlinear Euler equations.In the fuzzy approximation method,a nonlinear state space approximation model is established according to T-S fuzzy model,the fuzzy model is trained by using the anfis structure.Several linear state space models in different flight speed are generated based on the nonlinear flight dynamics model of a UH-60 helicopter.Two sets of LPV models are established by two methods.Finally,the influences characteristics of typical LPV modeling parameters to model accuracy have been investigated based on the Theil inequality coefficient and frequency responses.