Wind power turntable bearing fault feature extraction method

The invention relates to the technical field of electromechanical equipment state monitoring, in particular to a wind power turntable bearing fault feature extraction method. Collecting a fault signal of the wind power turntable bearing; performing sample filtering processing on each fault signal to obtain a plurality of components, and selecting the components for analysis; calculating a mean square error sigma, kurtosis K and energy E of the selected component; calculating a kurtosis mean value, an energy mean value and a mean square error mean value of the selected components in each signal; establishing a fuzzy BP neural network; respectively taking a kurtosis mean value, an energy mean value and a mean square error mean value of the selected components in each signal as parameters of a fuzzy BP neural network; according to the method, the fuzzy BP neural network is trained and fault identification is carried out on the fuzzy BP neural network, so that the fault special diagnosis signal can be accurately