航空学报 > 2023, Vol. 44 Issue (11): 227629-227629   doi: 10.7527/S1000-6893.2022.27629

基于数字孪生的飞机起落架健康管理技术

郭丞皓1, 于劲松1(), 宋悦1, 尹琦2, 李佳璇2   

  1. 1.北京航空航天大学 自动化科学与电气工程学院,北京 100191
    2.航空工业成都飞机工业(集团)有限责任公司,成都 610073
  • 收稿日期:2022-06-15 修回日期:2022-07-02 接受日期:2022-07-06 出版日期:2023-06-15 发布日期:2022-07-14
  • 通讯作者: 于劲松 E-mail:yujs@buaa.edu.cn
  • 基金资助:
    国家重点研发计划(2022YFB3304600);国家自然科学基金(51875018)

Application of digital twin⁃based aircraft landing gear health management technology

Chenghao GUO1, Jinsong YU1(), Yue SONG1, Qi YIN2, Jiaxuan LI2   

  1. 1.School of Automation Science and Electrical Engineering,Beihang University,Beijing 100191,China
    2.Avic Chengdu Aircraft Industrial (Group) Co. Ltd,Chengdu 610073,China
  • Received:2022-06-15 Revised:2022-07-02 Accepted:2022-07-06 Online:2023-06-15 Published:2022-07-14
  • Contact: Jinsong YU E-mail:yujs@buaa.edu.cn
  • Supported by:
    National Key R&D Program of China(2022YFB3304600);National Natural Science Foundation of China(51875018)

摘要:

针对飞机起落架系统,传统健康管理方式存在知识不完备、数据不平衡和模型固化等问题,开展了数字孪生驱动的健康管理技术研究,提出以自更新模型为基础实现故障诊断与预测的数字孪生健康管理框架。从起落架系统物理和行为2个维度建立孪生模型来实现真实系统的数字映射,依托强化学习算法实现数字孪生模型参数的更新,确保孪生模型实时跟踪和反映实体健康状态,并以此为基础设计基于事件的故障诊断和基于粒子滤波的故障预测方案。以起落架收放系统为例完成实验验证,与传统方法相比,本文方法在实时性、准确性和鲁棒性方面表现更优。

关键词: 起落架, 数字孪生, 健康管理, 基于事件的故障诊断, 模型跟踪, 故障预测

Abstract:

Traditional health management methods for aircraft landing gear systems face the problems of inadequate knowledge, unbalanced data, and rigidified models. This paper explores the application of the digital twin-driven health management technology, and proposes a digital twin health management framework based on self-updating models to reliably complete diagnostic and prediction tasks. The digital twin model is established from physical and behavior dimensions to realize the digital mapping of real systems. The reinforcement learning algorithm is used to update the parameters of the digital twin model to ensure real-time tracking and reflection of the entity health status by the twin model. Further, event-based fault diagnosis and particle filter scheme-based fault prediction are designed. In the validation experiment with the retraction/extension as an example, our method exhibits better performance in terms of real-time, accuracy and robustness than traditional methods.

Key words: landing gear, digital twin, health management, event-based fault diagnosis, model tracking, fault prediction

中图分类号: