Gear Pitting Fault Diagnosis Based on Improved Variational Modal Decomposition

Abstract

Based on an improved variational mode decomposition（VMD） algorithm，a novel gear fitting fault diagnosis method is proposed to overcome the difficulties in extracting the features of gear surface faults. The number of components for VMD is determined by the energy ratios of the components in empirical mode decomposition（EMD）. The penalty factor is established with a genetic algorithm（GA） via a fitness function between the permutation entropy and minimum value of VMD components，which determines penalty factors of the VMD. The improved VMD algorithm is used to process the gear surface fault signal. We screen the intrinsic mode function（IMF） for envelope adjustment and analyze the characteristic information of the gear pitting fault through the envelope spectrum. Compared with the current methods，the result of the gear pitting signal we analyze shows the improved VMD algorithm can identify the gear pitting faults more accurately and provide practical application value in the fault diagnosis of transmission systems.

Keywords

• Gear fault;Variational modal decomposition;Energy ratio;Permutation entropy;Feature extraction