ROLLING BEARING WEAK FAULT FEATURE EXTRACTION METHOD WITH ALIF⁃NLM

Aiming at the problem that the early weak fault feature was difficult to extract of rolling bearing under the strong noise background,combined with the advantages of adaptive local iterative filter(ALIF)and non⁃local means(NLM)method,an ALIF⁃NLM bearing weak fault feature extraction method was proposed.Firstly,a weighted kurtosis⁃energy ratio criterion was constructed to filter the intrinsic mode function(IMF)components of the ALIF decomposition and reconstruct the signal.Secondly,the minimum energy entropy⁃kurtosis ratio index was constructed by combining the sensitivity of kurtosis to the impact signal with the evaluation performance of energy entropy to the uniformity and complexity of signal energy distribution,and using this index as the fitness function,the adaptive selection of parameter combinations in NLM method was realized by particle swarm optimization(PSO)algorithm.Finally,the fault feature of the reconstructed signal was extracted with the adaptive NLM.The simulation and experimental results show that this method can effectively extract the weak fault feature information of rolling bearing under the strong noise background.