Research on Gearbox Fault Detection and Diagnosis Based on Improved Spectral Kurtosis Algorithm

Abstract

A gearbox fault detection and diagnosis test table of wheeled armored vehicles is designed and established. Typical faults and vibration experiments can be preset. Aiming to the d that traditional spectral kurtosis algorithm cannot be applied to gearbox fault signal feature extraction under strong noise interference, the minimum entropy deconvolution theory is adopted, and a new kind of gearbox fault diagnosis method based on MED and FSK is proposed, which realizes the single fault diagnosis of gears and bearings at different speed conditions, and also achieves good result for composite fault diagnosis of rolling bearing inner and outer rings. Compared with the traditional method of wavelet analysis and EMD, the proposed method for gearbox vibration signal has better noise reduction result.