A metamaterials-augmented drone monitor for acoustics-based remote fault detection and diagnosis

Yubin Lin, Shiqing Huang, Rongfeng Deng, Minglei Wang, Zhexiang Zou, Fengshou Gu, Andrew D Ball

Research output: Contribution to journalArticlepeer-review

Abstract

Drone Acoustic Remote Sensing Technology (DARST) holds significant potential in the approach to promote condition monitoring performance and efficiency. However, due to high aerodynamic noises from the drone itself, DARST has to be equipped with large scale microphone arrays and embedded with complex algorithms in order to improve the signal-to-noise ratio (SNR). This complicated construction along with high costs, makes DARST systems hard to be deployed in industrial applications. To address this challenge, this paper proposes a compact acoustic metamaterial structure-based system to augment the drone acoustic sensing, which is shortened to Metamaterials-augmented Drone Monitor (MADM) approach for the sake of simplicity. The design and performance of the proposed MADM method were explored through numerical simulations and experiments, demonstrating superior acoustic signal enhancement, frequency-selective behaviour and directional sensitivity characteristics. Subsequently, a series of experiments with machinery fault signals under different SNRs were conducted, comprehensively assessing the fault detection capabilities of the proposed MADM in monitoring and diagnosing a machinery transmission system. The results show that the MADM method can efficiently suppress drone-induced noise and robustly extract diagnostic features from bearing and gear faults, even under strong noise influence at SNR as low as −20 dB, and the average SNR improvement surpasses 140 %. The simple structured system needs only a single-channel acoustic signal with commonly used FFT-based algorithms, which occupies merely 22 % of the drone’s total payload capacity, offering a new cost-efficient approach for DARST in online inspection and fault diagnosis of industrial machinery.
Original languageEnglish
Article number112346
Number of pages18
JournalMechanical Systems and Signal Processing
Volume226
Early online date18 Jan 2025
DOIs
Publication statusE-pub ahead of print - 18 Jan 2025

Cite this