航空学报 > 2025, Vol. 46 Issue (7): 130920-130920   doi: 10.7527/S1000-6893.2024.30920

基于噪声特性估计的气动系数辨识方法

汪清1,2, 郑凤麒1,2(), 丁娣1,2, 岳茜2   

  1. 1.空天飞行空气动力科学与技术全国重点实验室,绵阳 621000
    2.中国空气动力研究与发展中心 计算空气动力研究所,绵阳 621000
  • 收稿日期:2024-07-09 修回日期:2024-07-29 接受日期:2024-09-24 出版日期:2024-10-24 发布日期:2024-10-23
  • 通讯作者: 郑凤麒 E-mail:zhengfq_mail@163.com
  • 基金资助:
    国家级项目

Aerodynamic coefficient identification method based on noise statistics estimation

Qing WANG1,2, Fengqi ZHENG1,2(), Di DING1,2, Xi YUE2   

  1. 1.State Key Laboratory of Aerodynamics,Mianyang 621000,China
    2.Computational Aerodynamics Institute,China Aerodynamics Research and Development Center,Mianyang 621000,China
  • Received:2024-07-09 Revised:2024-07-29 Accepted:2024-09-24 Online:2024-10-24 Published:2024-10-23
  • Contact: Fengqi ZHENG E-mail:zhengfq_mail@163.com
  • Supported by:
    National Level Project

摘要:

利用飞行试验数据验证和修正风洞气动力数据库,是飞行器设计与评估的一个重要环节。针对飞行试验普遍没有角加速度测量的情况,发展了一种新的气动系数辨识方法。首先将气动系数的时间导数建模为一阶Gauss-Markov过程,从而构建了气动系数辨识数学模型。然后,从似然函数最大化出发,通过理论推导给出了过程噪声和测量噪声协方差等未知统计量的解析表达式。采用平方根无迹Kalman滤波器(SRUKF)和无迹Rauch-Tung-Striebel平滑器(URTSS)进行状态估计。根据状态估计结果显式计算未知统计量并迭代修正,从而获得气动系数(作为增广状态变量)时间历程的辨识结果。2个飞机气动系数辨识算例演示了该方法的有效性。算例表明,该方法能够较好地估计未知统计量,给出合理的气动系数辨识结果。此外,该方法具有良好的收敛鲁棒性,不依赖于未知统计量的初始估计。

关键词: 气动系数辨识, 噪声协方差, 平方根无迹Kalman滤波器, 无迹Rauch-Tung-Striebel平滑器, 似然函数, 飞行试验, 气动参数估计

Abstract:

Using flight data to verify and update the wind-tunnel aerodynamic database is an important part of flight vehicle design and evaluation program. In response to the lack of angular acceleration measurement in flight tests, a novel aerodynamic coefficient identification method has been developed in this paper. Firstly, a mathematical model of aerodynamic coefficient identification was constructed by modeling the derivatives of aerodynamic coefficient respect to time as first-order Gauss-Markov process. Then, analytical expressions for the unknown statistics, such as the covariances of process and measurement noise, were derived theoretically by maximizing the likelihood function. The state estimation was conducted by using the Square Root Unscented Kalman Filter (SRUKF) associated with the Unscented Rauch-Tung-Striebel Smoother (URTSS). The unknown statistics were computed explicitly and updated iteratively, based on the state estimation results. Thereby, the time histories of aerodynamic coefficients, as the augmented state variables, were obtained. The effectiveness of the developed method was demonstrated by two examples of aircraft aerodynamic coefficient identification. The results showed that the unknown statistics and aerodynamic coefficients were estimated accurately. In addition, the method is of robust convergence with respect to the initial estimates of unknown statistics.

Key words: aerodynamic coefficient identification, noise covariance, square root unscented Kalman filter, unscented Rauch-Tung-Striebel smoother, likelihood function, flight test, aerodynamic parameter estimation

中图分类号: