Adaptive VMD and PSO-MOMEDA Algorithm based method for Rotor System Fault Diagnosis

Main Article Content

Fengfeng Bie, Ying Zhang, Fengxia Lyu, Xingting Miao, Yifan Wu


In this paper, a fault diagnosis method combining the variational mode decomposition (VMD) with adaptive parameters and multi-point optimal minimum entropy deconvolution adjusted (MOMEDA) optimized by partial swarm optimization algorithm (PSO) aiming at the problem that the fault information is hard to extract from the overwhelmed vibration signal when the rotor system works in strong noise setting is proposed. Firstly, the original signal is decomposed into intrinsic mode components (IMF) series by VMD, from which the specified IMF components are selected for signal reconstruction according to the correlation coefficient – kurtosis criterion, and then the reconstructed signal is processed with the MOMEDA model for the filtering. Finally, the rotor system fault type is identified through the signal envelope spectrum analysis. Simulation test and rotor fault experiments results show that the improved VMD model with PSO-MOMEDA method can primely remove the noise from the vibration signal of the rotor system and accomplish the fault mode recognition with avoiding the blind selection of parameters in the process.

Article Details

How to Cite
Xingting Miao, Yifan Wu, F. B. Y. Z. F. L. (2021). Adaptive VMD and PSO-MOMEDA Algorithm based method for Rotor System Fault Diagnosis. CONVERTER, 2021(8), 390 - 409. Retrieved from