When the rolling bearing of a helicopter swash plate works in a low speed environment, the characteristic frequency of faults completely submerged in various interferences, making the traditional diagnosis method depending on spectral peaks of characteristic frequencies ineffective. To solve the problem, a low speed rolling bearing fault diagnosis method based on frequency bands energy features is proposed. FFT and power spectrum of the fault signal are calculated. Energy distribution characteristics in the signals of four kinds of vibrations, i.e., normal, inner ring, outer ring and ball fault, are used to construct a fault feature vector. Finally, support vector machine (SVM), which can use small scale samples, is used to construct a classifer used for determining the type of fault. On an experimental platform, simulation test was carried out in a low speed working environment of rolling bearing of a helicopter swash plate. Analysis of the collected vibration signals shows that, compared with the traditional LMD and envelope spectrum characteristics method, the adaptability and superiority of the proposed method in low speed rolling bearing fault diagnoses are verifed.
