Wind power gear box fault diagnosis method

The present invention provides a wind power gear box fault diagnosis method which aims to solve the problem in the prior art that the fault information is very difficult to detect at the operation of a gear box. When the gear box goes wrong, a vibration signal of the gear box has two characteristics: 1 a fault signal is weaker and is easy to inundated by other signals; 2 the fault signal generates a modulation phenomenon because of the fault impact. The wind power gear box fault diagnosis method of the present invention applies the ensemble empirical mode decomposition (EEMD) and a Hilbert square demodulation technology respectively to solve the above two problems. Firstly, by applying the EEMD, an original vibration signal is decomposed into a series of intrinsic mode functions (IMF), in the IMFs, the IMF having the maximum kurtosis value usually corresponds to a Fault feature component; and then the Hilbert square demodulation is utilized to demodulate the IMFc, and the spectral analysis is carried out on the demodulated signal, and finally, a spectral analysis result is compared with the rotation frequencies of the rotation parts of the gear box to determine a fault position.