数字孪生机翼损伤模式快速识别与监测方法
Rapid identification and monitoring of digital twin wings damage patterns
王子一 1粟华 2龚春林 2蔡艳芳 3丁轩鹤 1杨予成1
作者信息
- 1. 西北工业大学航天学院,西安 710072
- 2. 西北工业大学航天学院,西安 710072;西北工业大学空天飞行技术研究所,西安 710072
- 3. 西安现代控制研究所,西安 710065
- 折叠
摘要
针对飞行器结构健康监测过程中存在的识别流程复杂、实时性较差问题,提出一种基于数字孪生技术的飞行器机翼损伤模式识别与监测方法.采用模块化技术构建飞行器机翼的数字孪生结构模型,基于概率神经网络建立了传感器数据在结构数字孪生模型中的映射方法,形成了通用的数字孪生飞行器结构损伤模式快速识别流程.以某无人机为例,基于此流程方法建立了其机翼的损伤模式快速识别模型并开展了对损伤的识别.结果表明:构建的飞行器结构数字孪生识别模型对损伤模式的识别准确率达到了 96％以上,能够实现动态航迹规划任务.
Abstract
To address the problems of complex recognition and poor real-time performance in the process of structural health monitoring of aircraft,a digital twin technology-based damage pattern recognition and prediction method for aircraft wings was proposed.The digital twin structural model of the aircraft wing was constructed using modular technology,and the mapping method of sensor data in the structural digital twin model was established based on probabilistic neural network,forming a fast monitoring process of general digital twin aircraft structural damage pattern.Based on an unmanned aerial vehicle,a rapid damage pattern recognition model of its wings was developed.The results showed that the damage pattern identification accuracy of the digital twin recognition model for aircraft structures reached over 96％,which could complete the dynamic trajectory planning task.
关键词
结构健康监测/数字孪生/损伤模式/模式识别/概率神经网络Key words
structural health monitoring/digital twin/damage classifications/pattern recognition/probabilistic neural network引用本文复制引用
基金项目
基础科研项目(JCKY2020204B016)
出版年
2024
NETL
NSTL
万方数据