Litcius/Paper detail

Bearing Fault Diagnosis Based on the Maximum Squared-Enveloped Multipoint Kurtosis Morphological Deconvolution

Mingjun Tang, Yuhe Liao, Rongkai Duan, Jiutao Xue, Xining Zhang

2022IEEE Transactions on Instrumentation and Measurement18 citationsDOI

Abstract

Extracting the fault-related repetitive impulses from the observed signal disturbed by deterministic components, random impulses and background noise are a critical problem in diagnosing the bearing fault. This article proposes a maximum squared-enveloped multipoint kurtosis morphological deconvolution (MSEMKMD) method to tackle this issue. First, an unbiased-autocorrelation double-scale morphological filter (MF) is designed to preserve the repetitive-impulse characteristic of the observed signal and filter out these noises from different sources. The minimum entropy deconvolution (MED) method is then used to enhance recovering the repetitive impulses disturbed by these noises. In this method, the squared-enveloped multipoint kurtosis (SEMK) indicator is developed to select the optimal structural element (SE) scale and filtered result. Finally, the results of analyses with simulation and experimental data show the superiority of the MSEMKMD in fault-related impulse recovery for bearing fault diagnosis.

Topics & Concepts

KurtosisDeconvolutionBlind deconvolutionAutocorrelationAlgorithmImpulse (physics)Finite impulse responseComputer scienceEngineeringElectronic engineeringMathematicsStatisticsPhysicsQuantum mechanicsMachine Fault Diagnosis TechniquesGear and Bearing Dynamics AnalysisAdvanced machining processes and optimization