Research on Feature Extraction Strategy Based on Improved Empirical Mode Decomposition Method

DOI: https://doi.org/10.62517/jes.202302306

Author(s)

Zhaohui Liu*

Affiliation(s)

Inner Mongolia Power (Group) Co., Ltd. Hohhot Power Supply Branch, Hohhot, Inner Mongolia, China *Corresponding Author.

Abstract

An improved empirical mode decomposition method is proposed to address the issues of modal aliasing and endpoint effects that affect the accuracy of fault diagnosis in traditional empirical mode decomposition methods. This method uses a median filter with a variable window when generating the intrinsic mode function (IMF). Compared with the traditional empirical mode decomposition method, the improved empirical mode decomposition method (IEMD) can reduce the mode aliasing and end effect, and improve the efficiency of feature extraction. In the IEMD method, the bearing vibration signals are decomposed by EMD, and the obtained IMF components are processed by a median filter with variable window values, in which the narrow window is used for the high frequency component and the wide window is used for the low frequency component. Then, the filtered internal model function is summed and subjected to a round of empirical mode decomposition to obtain an improved internal model function. The traditional EMD method and the IEMD method are compared by using the accelerated aging test equipment of induction motor bearings. The results show that the IEMD method can separate the characteristic frequency of the early fault of the motor and detect the early fault effectively.

Keywords

Empirical Mode Decomposition; Electric Machine; Early Fault Detection; Median Filter

References

[1]Xuegang N H J Y. An adaptive real-time modular tidal level prediction mechanism based on EMD and Lipschitz quotients method. Ocean Engineering, 2023, 289 (P1). [2]Sunitha R S. Assessment of Mode Shape in Power System Using TVF-EMD and Spectral Analysis. Electric Power Components and Systems, 2022, 50 (6-7): 349-358. [3]Li C X X. Vibration Diagnosis and Optimization of Industrial Robot Based on TPA and EMD Methods. Computer Modeling in Engineering & Sciences, 2022, 135 (3): 2425-2448. [4]Harrison T R I. Hybrid pixel based method for multimodal image fusion based on Integration of Pulse Coupled Neural Network (PCNN) and Genetic Algorithm (GA) using Empirical Mode Decomposition (EMD). Microprocessors and Microsystems, 2022, 94 [5]Xuyi Y L H Z C Z. Parameter-Adaptive TVF-EMD Feature Extraction Method Based on Improved GOA. Sensors, 2022, 22 (19): 7195-7195. [6]Zhou Z X J L. An Improved Arc Fault Location Method of DC Distribution System Based on EMD-SVD Decomposition. Applied Sciences, 2023, 13 (16). [7]Kumar P S I K B C S S. Fault detection technique using time-varying filter-EMD and differential-CUSUM for LVDC microgrid system. Electric Power Systems Research, 2023, 219.\ [8]Liu Z H W W D. Experimental study of cavitation noise characteristics in a centrifugal pump based on power spectral density and wavelet transform. Flow Measurement and Instrumentation, 2023, 94.