Ultra-Wideband Detection-Enabled Modularized Three-Axis Vibration Sensor for UAV Blade Damage Monitoring via Deep Learning

作者：Kuankuan Wang, Yixin Liu, Xiang Guan, Zhihong Wang, Zhipeng Pan, Yang Yu, Yongming Yao, Tianyu Li · 发表于：IEEE Transactions on Industrial Electronics · 年份：2025 · DOI：10.1109/tie.2025.3561854 · 被引用次数：8 · 研究领域：Structural Health Monitoring Techniques、Advanced Measurement and Detection Methods、Advanced Sensor and Control Systems

The blades, which act as the main propulsion components, are subject to external conditions and are susceptible to damage, thereby affecting unmanned aerial vehicle (UAV) safety and decreasing endurance. Vibration analysis, using vibration sensors, is an effective method for detecting mechanical faults. However, due to the compact nature of UAV structures, commercial sensors cannot be directly installed on motor sections for monitoring. This article presents a modular and expandable (ME) wideband vibration sensor that employs triboelectric nanogenerator (TENG). ME-TENG broadens its frequency detection range through a simple bouncing ball structure, adjustable materials for the bouncing ball, and realized vibration frequency range from 5 to 5000 Hz and allows for octave frequency monitoring. Experimental validation demonstrates ME-TENG's modular capability, allowing for three-axis vibration monitoring. Additionally, a data reading and storage system (UDRS) based on ME-TENGs is developed. UDRS incorporates a deep learning model for the accurate classification of damage types in UAV blades, achieving a high accuracy rate of 97.3%. The deep learning model is then deployed in the UDRS for real UAV flight tests. This research introduces an innovative method for real-time monitoring of structural faults in UAVs. The study also highlights the potential for applying TENGs in engineering contexts.