A compound fault diagnosis method for rolling bearings based on the IPSO-MOMEDA and Teager energy operator

Abstract Different types of faults interact with each other and are easily overwhelmed by strong noise, which makes it challenging to identify and isolate single fault features in rolling bearing compound faults (RBCFs). To address this problem, a diagnosis method for RBCFs with improved particle swarm optimization multipoint optimal minimum entropy deconvolution adjusted (IPSO-MOMEDA) and a Teager energy operator is proposed. The optimal settings for the period and filter length are automatically determined by the excellent optimization capabilities of IPSO, which leads to quick and efficient identification of the global optimum. Moreover, the optimal deconvolution of various fault types is obtained by the optimized MOMEDA. The Teager energy operator is introduced to enhance the shock and periodicity components of the deconvolution signal, which can ensure that the accurate fault period can be selected by MOMEDA. Envelope analysis is employed for identification of compound fault characteristics. Both the simulation and experimental results demonstrate that inner ring and outer ring faults, inner ring and ball faults, outer ring and ball faults, and inner–outer ring and ball faults can be accurately diagnosed using the proposed method.