EEMD-1.5 Dimension Spectrum Applied to Locomotive Gear Fault Diagnosis

The criterion of adding white noise in Ensemble Empirical Mode Decomposition (EEMD) method is established. EEMD, used for avoiding mode mixing in signal decomposition, is combined with 1.5 dimension spectrum, which is the bispectrum diagonal slice to eliminate white noise and extract nonlinear coupling feature. A new method of EEMD-1.5 dimension spectrum for fault feature extraction is proposed. Firstly, vibration signal is adaptively anti alias decomposed by EEMD to get Intrinsic Mode Functions (IMFs). Then, 1.5 dimension spectrum is adopted to process IMFs which contain fault feature. Finally, EEMD-1.5 dimension spectrum is introduced into monitoring diagnosis of the gear box of the locomotive running gear, the results show that it successfully extracts an early gear crack fault feature.