基于FSDP图像和DCNN的无人机起落架关键件结构损伤智能检测方法

Intelligent detection approach for the critical part in the landing gear of UAV based on the FSDP figure and DCNN

随着无人机的应用越来越广泛,与之相适应的保障手段也需与时俱进.在人工智能快速发展的情况下,故障诊断与人工智能方法相结合为无人机保障关键技术的跨越式发展提供了重要契机,也为无人机的使用安全性与运行可靠性水平的提升提供了重要机遇.本文以结构健康监测中的超声信号分析方法为基础,针对无人机起落架关键结构件的损伤检测问题开展研究,结合深度学习模型,对结构件的损伤进行智能化检测,提高检测效率和检测精度.针对结构健康监测中损伤信号的非线性特点,提出了频谱对称点阵图案(frequency symmetrized dot pattern,FSDP)特征提取方法.在此基础上,提出了基于FSDP特征与深度卷积神经网络的损伤智能检测方法.在起落架T型构件实验环境中开展了结构损伤模拟与检测实验,通过实验数据验证了所提出的智能检测方法的有效性.

Considering the increasing use of UAVs,the corresponding means of support must keep up.With the large-scale development of artificial intelligence,the combination of fault diagnosis and artificial intelligence provides an important opportunity for developing key UAV support technologies and improving UAV safety and operational reliability.Using the ultrasonic signal analysis method in structural health monitoring,this study researches the damage detection of key structural components of UAV landing gear and combines the deep learning model to intelligently detect and improve the detection efficiency and accuracy of structural component damage.To address the characteristics of nonlinear damage signals in structural health monitoring,the feature extraction method of frequency symmetrized dot pattern(FSDP)is proposed.Consequently,we propose an intelligent damage detection method based on the FSDP feature and DCNN.The structural damage simulation and detection experiments are performed in the landing gear experimental environment,and the efficacy of the proposed intelligent detection method is validated using experimental data.