基于物理信息神经网络的飞机气动参数辨识方法

付军泉; 钟伯文; 钟运琴; 刘赟

doi:10.7638/kqdlxxb-2022.0201

欢迎来到能源电力期刊网！

English Version 期刊登录

您当前的位置：

首页 >

文章列表页 >

基于物理信息神经网络的飞机气动参数辨识方法

研究论文 | 更新时间：2026-05-28

- 基于物理信息神经网络的飞机气动参数辨识方法
- A physics informed neural network based method for aircraft aerodynamic parameter identification
- 空气动力学学报 2023年41卷第9期页码：30-37
- 作者机构：
  
  1. 南昌航空大学飞行器工程学院,南昌,330063
  2. 江西省飞行器设计与气动仿真重点实验室,南昌,330063
  3. 中国科学院大学中国科学院大数据挖掘与知识管理重点实验室,北京,100190
- 作者简介：
  
  付军泉（1993-），男，江西抚州人，讲师，研究方向：空气动力学. E-mail：fujquan@nchu.edu.cn
  钟伯文*，研究员，研究方向：空气动力学. E-mail：zhongbowen@ nchu.edu.cn
- 基金信息：
  
  国家自然科学基金（12262023）;江西省重点研发计划项目（20223BBE51001）
- DOI：10.7638/kqdlxxb-2022.0201
  中图分类号： TP183;V212.1
- 纸质出版：2023
- 稿件说明：
移动端阅览
付军泉, 钟伯文, 钟运琴, 等. 基于物理信息神经网络的飞机气动参数辨识方法[J]. 空气动力学学报, 2023, 41(9): 30−37.

FU Junquan, ZHONG Bowen, ZHONG Yunqin, et al. A physics informed neural network based method for aircraft aerodynamic parameter identification[J]. 2023, 41(9): 30-37.
付军泉, 钟伯文, 钟运琴, 等. 基于物理信息神经网络的飞机气动参数辨识方法[J]. 空气动力学学报, 2023, 41(9): 30−37. DOI： 10.7638/kqdlxxb-2022.0201

FU Junquan, ZHONG Bowen, ZHONG Yunqin, et al. A physics informed neural network based method for aircraft aerodynamic parameter identification[J]. 2023, 41(9): 30-37. DOI： 10.7638/kqdlxxb-2022.0201.

摘要

在飞机设计与研制过程中，通过气动参数辨识建立可靠的飞行动力学模型非常重要。传统的气动参数辨识工程算法，诸如极大似然法，需要给出合理的飞行动力学模型以及待辨识参数的初值。基于传统神经网络的气动参数辨识可以避免飞行动力学建模过程，这种方法需要通过增量法、导数法间接地从神经网络提取气动参数。本文提出了一种基于物理信息神经网络的飞机气动参数辨识方法，可将含待辨识参数的飞行动力学模型作为正则项加入损失函数，直接辨识得到气动参数。该方法可以显著减少建模数据需求，也能提高建模精度。飞行仿真数据验证结果表明，该方法的无噪声、含2%噪声仿真数据，纵向飞行状态空间模型辨识最大相对误差分别为1.80%、4.64%，表明了基于物理信息神经网络的飞机气动参数辨识方法具有可行性，并对含噪声的飞行数据具有泛化性。

Abstract

It is of great importance to establish a reliable flight dynamic model based on aerodynamic parameter identification in the process of aircraft design. Traditional engineering methods for aerodynamic parameter identification

such as the maximum likelihood method etc.

require a reasonable dynamic model and initial values for the parameters to be identified. The identification of aerodynamic parameters based on traditional neural networks has no requirement for the flight dynamic model and parameter initialization. In this study

an aerodynamic parameter identification method based on physics informed neural network (PINN) is proposed

which can directly identify aerodynamic parameters by incorporating the flight dynamic model as a regularization term into the loss function. This method can significantly reduce the required training data and improve the modeling accuracy. Verification results for the flight simulation data show that the maximum relative errors of the PINN based method on the longitudinal flight state space model identification with no noise and 2% noise are 1.80% and 4.64%

respectively. It suggests that the PINN based aerodynamic parameter identification method is feasible and has the generalization capacity to noisy flight data.

关键词

Keywords

references

ZHANG M, LIU T J, MA T L, et al. High speed aerodynamic design of large civil transporter based on CFD method[J]. Acta Aeronautica et Astronautica Sinica, 2016, 37(1): 244-254. (in Chinese)

LIU B, LIU P Q, WANG L. A small trim-wing unmanned aircraft aerodynamic parameter estimation and analysis of wind tunnel test[J]. Aircraft Design, 2010, 30(3): 1-5, 22. (in Chinese)

CAI J S. Advances in identification of aircraft aerodynamic parameters[J]. Advances in Mechanics, 1987, 17(4): 467-478. (in Chinese)

WARNER E, NORTON F. Preliminary report on free flight tests[R ] . NACA-TR-70, 1920. https://ntrs.nasa.gov/api/citations/19930091132/downloads/19930091132.pdf

HARRY G. A survey of methods for determining stability parameters of an airplane from dynamic flight measurements[R ] . NACA-TN-2340, 1951. https://ntrs.nasa.gov/api/citations/19930082979/downloads/19930082979.pdf

LIN D M, MAKIS V. On-line parameter estimation for a partially observable system subject to random failure[J]. Naval Research Logistics, 2006, 53(5): 477-483. DOI: 10.1002/nav.20156

YANG Z S. Research on aerodynamic online identification and trajectory optimization based on integrated navigation for boost-glide missile[D]. Changsha: National University of Defense Technology, 2017 (in Chinese) .

CUI P Y. Research on the parameter identification method and application for vehicle[J]. Flight Dynamics, 1996, 14(3): 1-8. (in Chinese)

YANG G H, DU L F, LI H, et al. Parameter identification and adaptive control of aircraft based on BP neural network[J]. Aerospace Control, 2021, 39(5): 3-7. (in Chinese)

GHOSH A K, KHAN A S, RAISINGHANI S C. Estimation of longitudinal parameters from real-flight data using neural networks[J]. Journal of the Aeronautical Society of India, 2002, 54(1): 16-22.

KUTTIERI R A, SINHA M. Neural partial differentiation for parameter estimation of flexible aircraft dynamics[J]. Journal of Aerospace Sciences and Technologies, 2016, 68(4): 276-286.

VERMA H O, PEYADA N K. Aircraft parameter estimation using ELM network[J]. Aircraft Engineering and Aerospace Technology, 2020, 92(6): 895-907. DOI: 10.1108/aeat-01-2019-0003

LI Y, CHEN S C. Physics-informed neural networks: recent advances and prospects[J]. Computer Science, 2022, 49(4): 254-262(in Chinese).

PANG G F, LU L, KARNIADAKIS G E. fPINNs: fractional physics-informed neural networks[J]. SIAM Journal on Scientific Computing, 2019, 41(4): A2603-A2626. DOI: 10.1137/18m1229845

YANG L, MENG X H, KARNIADAKIS G E. B-PINNs: Bayesian physics-informed neural networks for forward and inverse PDE problems with noisy data[J]. Journal of Computational Physics, 2021, 425: 109913. DOI: 10.1016/j.jcp.2020.109913

COOK M V. Flight dynamics principles[M]. 2nd ed. Amsterdam: Butterworth-Heinemann, 2007.

LIU H, ZHANG L. Research on six degrees of freedom flight simulation based on JSBsim and flightgear[J]. Journal of Civil Aviation Flight University of China, 2018, 29(3): 44-48. (in Chinese)

LI F G, SHI J W, ZHANG W X. Identification of helicopter aerodynamic model parameters through frequency method using flight test data[J]. Aeronautical Science Technology, 2018, 29(9): 37-42. (in Chinese)

MCLEAN D. Automatic flight control systems[J]. Measurement and Control , 2003, 36 (6): 172-175. doi:10.1177/002029400303600602

WANG S F, TENG Y J, PERDIKARIS P. Understanding and mitigating gradient flow pathologies in physics-informed neural networks[J]. SIAM Journal on Scientific Computing, 2021, 43(5): A3055-A3081. DOI: 10.1137/20M1318043

TISCHLER M B, REMPLE R K. Aircraft and rotorcraft system identification: engineering methods with flight-test examples[M]. Reston, VA: American Institute of Aeronautics and Astronautics, 2006.

浏览量

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

基于U-Net的磁浮飞行风洞气动力干扰剥离算法

基于条件扩散模型的火箭气动外形快速设计方法

定常来流下钝体二维组合断面气动特性与流场的智能预测方法

基于物理约束深度学习的大跨柔性光伏阵列绕流场重构

基于深度神经网络的高速翼型流场降阶模型