Fiber Bragg grating sensing and compound fault diagnosis for key component of wind turbines

Abstract

Wind power is an important part of the energy transition, and the wind turbine gearbox is a core transmission component of wind power generation systems. Effective detection and identification of gearbox compound faults are therefore important for ensuring the stable and safe operation of wind turbines. This paper proposes an intelligent compound fault diagnosis method for wind turbine gearboxes based on fiber Bragg grating (FBG) vibration sensing and a ResNet model with efficient channel attention (ResNet-ECA). First, a flexure-hinged FBG accelerometer is designed and calibrated. The proposed sensor exhibits a flat sensitivity response from 5 Hz to 1000 Hz, a nominal sensitivity of 31.49 pm g−1, and linearity exceeding R2 = 0.999. Subsequently, vibration signals under eight different health states are collected using the FBG accelerometer. A Dragonfly algorithm-optimized variational mode decomposition method is applied to denoise the FBG wavelength-shift signals and obtain high-quality one-dimensional (1D) vibration data. The denoised 1D signals are then transformed into two-dimensional image representations using the Gramian angular summation field, and the resulting images are input into the ResNet-ECA model for compound fault identification. The results show that the proposed method achieves an exact-match accuracy of 95.83% and a macro-average F1-score of 0.9583. These results demonstrate the effectiveness of the proposed FBG sensing and intelligent diagnosis framework for compound fault diagnosis of wind turbine gearboxes.