WANG Peng fei, HU Jian zhong

（School of Mechanical Engineering，Southeast University，Nanjing 211189，China）

 

Abstract: Aiming at the problems that most gearboxes fault vibration signals were always muti component amplitude modulated and frequency modulated (AM FM) signals and the selection of parameters usually depend on experience in conventional envelope analysis method. Decomposition of gearbox vibration signal, envelope analysis and extracting the characteristic value were researched and a gearbox fault diagnosis approach of envelope spectrum characteristic based on LMD (local mean decomposition) was proposed. Firstly, the vibration signal of gearbox could be decomposed into a set of product functions (PF) through LMD method, which coverd different frequency characteristics. Secondly, the first stage PF component including dominant fault information was analyzed by the envelope spectrum and the characteristics frequency was obtained, according to this method working condition and fault patterns of gearboxes were classified. The LMD method was verified by three types of gearbox vibration signals including the normal state, the local damage and the abrasion. The results indicate that the proposed method can be used to classify working condition and fault patterns of gearboxes.

 

Key words: gearboxes; local mean de composition (LMD); envelop spectrum characteristics; fault diagnosis