Fault Diagnosis of Wind Turbine Gearbox Based on Dual-tree Complex Wavelet and Information Entropy

Fault diagnosis for a wind turbine gearbox is of great significance for extending system reliability and reducing downtime. Gearbox vibration signals measured by sensors usually exhibit time-varying and non-stationary characteristics. Meanwhile, weak fault signals are always mixed with noise and buried in intensive vibration energy, so that it is difficult to find their early symptoms from a potential fault by traditional signal processing methods. This paper proposes a sensitive components selection algorithm (SCSA), which combines envelope spectrum entropy and correlation coefficient to highlight the fault characteristics. Then a novel time-frequency method combining information entropy (IE) with dual-tree complex wavelet transform (DT-CWT) is used for fault identification of planetary gearbox under varying-speed and noisy conditions. The vibration state of gearbox is monitored by time energy entropy, and the DT-CWT is used to filter the gearbox fault resonance features and extract the impulse features. Finally, a modulation signal and an actual application are used for testing. The results show that the proposed method is valid in diagnosis of weak faults for a planetary gearbox of wind turbine.