基于数据合成的飞行器结构损伤状态快速识别方法  

Rapid identification method for aircraft structural damage patterns based on data synthesis

作　　者：王浩渊 粟华[1,2] 李鹏 龚春林 WANG Haoyuan;SU Hua;LI Peng;GONG Chunlin(School of Astronautics,Northwestern Polytechnical University,Xi’an 710072,China;Shaanxi Aerospace Flight Vehicle Design Key Laboratory,Xi’an 710072,China;Xi’an Institute of Modern Control Technology,Xi’an 710065,China)

机构地区：[1]西北工业大学航天学院,陕西西安710072 [2]陕西省空天飞行器设计重点实验室,陕西西安710072 [3]西安现代控制技术研究所,陕西西安710065

出　　处：《系统工程与电子技术》2024年第11期3774-3783,共10页Systems Engineering and Electronics

基　　金：西安市科技计划项目(21RGZN0012,23ZDCYTSGG0026-2022);咸阳市重点研发计划项目(2021ZDYF-GY-0031)资助课题。

摘　　要：针对当前飞行器结构健康监测过程中存在的识别流程复杂、识别准确度低的问题,提出一种基于数据合成的飞行器结构损伤状态快速识别方法。建立数字孪生结构损伤快速识别模型构建流程,构建飞行器结构数字模型。基于数据合成思想,提出一种对传感器数据的可信度评价方法,建立具有可解释性的生成-判别模型,解决了因样本数据不足导致的学习准确率低的问题。引入分类边界模糊化方法,使用判别模型确定模糊化区域以提升神经网络识别的稳定性。最后,以某无人机为例,对所提识别方法进行验证。结果表明,该方法能够高效构建结构损伤状态数据库并提升识别泛化能力和稳定性,对损伤状态的识别准确率超过99%。In response to the complexity of the identification process and the low accuracy of identification in the current health monitoring process of aircraft structures,a rapid identification method for aircraft structural damage patterns based on data synthesis is proposed.A digital twin structure damage rapid identification model construction process is established to construct a digital model of aircraft structure.Based on the idea of data synthesis,a credibility evaluation method for sensor data is proposed,and an interpretable generative-discriminative model is established to solve the problem of low learning accuracy due to insufficient sample data.The method of fuzzy classification boundary is introduced,and the discriminative model is used to determine the fuzzy area to improve the stability of neural network identification.Finally,the proposed identification method is verified with a certain drone as an example.The results show that this method can efficiently build a structural damage pattern database and improve the generalization ability and stability of identification,with an identification accuracy rate of over 99% for damage patterns.

关 键 词：数字孪生 结构健康监测 损伤检测 数据合成 神经网络 

分 类 号：V11[航空宇航科学与技术—人机与环境工程]