一种基于短时机动试飞数据的动力学辨识优化方法

An optimization method for dynamic identification based on short-duration maneuvering flight test data

飞行动力学辨识算法的一个关键问题是,如何通过简单的机动获取所关心频率范围的响应特性。短时倍脉冲是一种易于实施的激励信号,兼顾试飞安全性与经济性,但与频域辨识法通常使用的扫频输入激励相比,短时机动频谱范围窄、信噪比低,一般难以得到准确的辨识结果。对如何基于短时机动飞行试验数据,提高辨识结果准确性的问题进行了研究。首先分析了经典Welch谱估计进行时域-频域转换过程中,影响非参数模型辨识精度的主要因素,提出了削减窗函数边缘缩减效应的数据预处理方法,并结合多窗口综合技术,提高频域特性辨识结果的精度。在参数化模型辨识过程中,针对有限频谱范围,提出了利用相干函数和功率谱密度加权综合,确定等效拟配的频率范围和频率节点的自适应方法,使得低阶等效拟配与输入激励信号高度相关,提高参数化模型辨识的精度、一致性和适应性。通过不同类型飞机的大量短时机动和少量扫频飞行试验数据模型辨识的工程应用示例,验证了动力学辨识优化方法算法稳定、结果准确,可满足飞行品质模态特性评价等应用需求。

How to generate the accurate response at the frequency range of interest from the maneuvering is one of the significant problems for dynamic identification.Although short-duration doublet is commonly used in flight test activities and is easy to implement with a good consideration of both security and economy,it is difficult to get the accurate estimation due to its limited spectrum and lower signal to noise ratio.Thus,a method focusing on improving estimation accuracy based on the short-duration maneuvering flight test data is developed in this paper.The main factors that affect the accuracy of the nonparametric model identification in the classical Welch method of spectrum estimate are analyzed.A novel data pre-processing method which can taper the window function edge is presented,and the multi-window composite technology is integrated to improve the identification accuracy.Regarding the limited frequency spectrum,an adaptive low-order equivalent matching method is developed to select the frequency range and node based on a weighted function of coherence function and power spectral density.This method make the low-order equivalent matching highly correlated with the maneuvering input signal,and improve the accuracy,consistency and adaptability of the model parameter identification.The dynamic identification optimization technique is applied to a large amount of short-duration maneuvering and several sweeping flight test data for various types of aircrafts.The results meet the accuracy requirement for flight quality evaluation application,while the algorithm is stable and reliable.