航空学报 > 2025, Vol. 46 Issue (19): 531292-531292   doi: 10.7527/S1000-6893.2024.31292

基于飞行参数的结构关键部位载荷孪生技术

陈亮1,2, 黄蕾3, 顾宇轩1,2, 郭聪3, 林可欣1,2(), 管宇1,2, 宋健1,2   

  1. 1.航空工业 沈阳飞机设计研究所,沈阳 110000
    2.辽宁省飞行器结构强度数字孪生重点实验室,沈阳 110000
    3.大连理工大学 力学与航天航空学院,大连 116024
  • 收稿日期:2024-09-30 修回日期:2024-10-24 接受日期:2024-12-11 出版日期:2024-12-25 发布日期:2024-12-23
  • 通讯作者: 林可欣 E-mail:d20152303@163.com
  • 基金资助:
    国防基础科研计划(JCKY2021205B003)

Twinning technology of key part load based on flight parameters

Liang CHEN1,2, Lei HUANG3, Yuxuan GU1,2, Cong GUO3, Kexin LIN1,2(), Yu GUAN1,2, Jian SONG1,2   

  1. 1.AVIC Shenyang Aircraft Design and Research Institute,Shenyang 110000,China
    2.Key Laboratory of Digital Twin for Aircraft Structural Strength in Liaoning Province,Shenyang 110000,China
    3.School of Mechanics and Aerospace Engineering,Dalian University of Technology,Dalina 116024,China
  • Received:2024-09-30 Revised:2024-10-24 Accepted:2024-12-11 Online:2024-12-25 Published:2024-12-23
  • Contact: Kexin LIN E-mail:d20152303@163.com
  • Supported by:
    National Defense Basic Scientific Research Program(JCKY2021205B003)

摘要:

为了有效监控飞机的健康,针对飞机服役过程中所受的复杂载荷条件,提出了基于飞行参数的结构关键部位载荷孪生技术。首先,基于相邻值填补、一维/多维数据异常检测、小波包分解和贝叶斯阈值去噪等方法完成光纤传感器数据优化治理;其次,基于数据挖掘手段和线性回归法,建立了结构关键部位应变特征及相关参数提取方法;最后,针对影响飞机结构关键部位应变的相关飞行参数和特征,采用XGBoost模型训练相关飞行参数和特征到关键部位应变的映射关系,构建了飞行参数-应变的高精度孪生映射模型。高精度孪生映射预测模型以飞参和传感器数据作为原始输入,平均预测精度达到95%以上,能够高效准确对飞机结构健康状态进行监测。

关键词: 数字孪生, 飞行参数, 结构强度, 映射模型, 健康监控

Abstract:

To effectively monitor the health of the aircraft, a twinning technology for load of key parts of the structure is proposed based on flying parameters in view of the complex load conditions during the service of the aircraft. Firstly, fiber optic sensor data optimization governance is completed using the methods such as adjacent value filling, one-/multi-dimensional data anomaly detection, wavelet packet decomposition and Bayesian threshold denoising. Secondly, based on the data mining method and linear regression method, methods are developed to extract the strain feature and related parameter of the key parts of the structure. Finally, the XGBoost model is used to train the mapping relationship between the relevant flight parameters and features and the strain of key parts, and a high precision twin mapping model of the flight parameters and strain is constructed. Using the flight parameters and sensor data as the original input, the high precision twin mapping model has an average prediction accuracy of above 95%, and can effectively and accurately monitor the healthy state of aircraft structure.

Key words: digital twin, flight data, structural strength, mapping model, health monitoring

中图分类号: