Fault Diagnosis of Intershaft Bearing Using Variational Mode Decomposition with TAGA Optimization

Jing Tian,Cheng-Wei Fei,Shu-Guang Wang,Yu-Wei Zhang,Jie Zhou,Yan-Ting Ai,Radoslaw Zimroz

doi:10.1155/2021/8828317

Abstract

To efficiently extract the features of aeroengine intershaft bearing faults with weak signal, the variational mode decomposition (VMD) method based on the tolerant adaptive genetic algorithm (TAGA) (TAGA-VMD) is proposed by introducing the idea of tolerance into the traditional adaptive genetic algorithm in this paper. In this method, the tolerant genetic algorithm was adopted to find the optimum empirical parameters K and α of VMD. A fault simulation experiment system of intershaft bearings was built for the inner ring fault and outer ring fault of bearings to verify the proposed TAGA-VMD method. The results show that the proposed method can effectively extract the fault feature frequency of intershaft bearings, and the error between the extracted fault feature frequency and the theoretical value of fault frequency is less than 0.1%. The efforts of this study provide one promising fault feature extraction approach for aeroengine intershaft bearing fault diagnosis with weak signal.

Highlights

Rotor is a key system of aeroengine and directly affects the operating conditions of aeroengine
In the online monitoring of aeroengine, sensors are mainly installed on the outer surface of casings to monitor the health status of rolling bearing, so that too long signal propagation path induces vibration signals attenuation, low signal to noise ratio (SNR) of fault signals, and difficult extraction of fault features [1]. erefore, it is urgent to study the weak fault diagnosis of intershaft bearings [2,3,4]
T 1 σk μk where xk (t) is the kth decomposed intrinsic mode function (IMF) of the Variational mode decomposition (VMD); μk is the average value of xk (t); σk is the standard deviation of xk (t); and krk is the kurtosis of kth IMF component of the fault signal

Summary

Introduction

Rotor is a key system of aeroengine and directly affects the operating conditions of aeroengine. As a nonstationary and nonlinear signal processing method, empirical mode decomposition (EMD) [15] was widely used in bearing fault diagnosis [16, 17], because it could adaptively decompose fault signals into different modal components. To find the adaptive parameters more precisely, the improved genetic algorithm is utilized to synchronously optimize the number of modal components K and penalty factor α of VMD. The tolerance genetic algorithmbased parameter adaptive VMD is used to process simulation signals and experimental signals, respectively, to extract the fault characteristics of intershaft bearing and realize the monitoring and diagnosis of intershaft bearing conditions. E experimental results show that TAGA can greatly improve the global search ability of genetic algorithm and avoid being trapped in local optimum and TAGA-VMD can extract the feature frequency of intershaft bearing faults more efficiently than general VMD.

Variational Mode Decomposition Method

Variational Mode Decomposition Based on Tolerance Adaptive Genetic Algorithms

Experimental Verification