Tool Wear Feature Extraction Based on Hilbert Marginal Spectrum

International Conference on Materials Sciences and Nanomaterials

材料科学;物理学

Guan, Shan^1 ; Song, Weijie^1 ; Pang, Hongyang^1

School of Mechanical Engineering, Northeast Electric Power University, Jilin

132012, China^1

关键词: Correlation coefficient;    Empirical Mode Decomposition;    Feature extraction methods;    Hilbert marginal spectrum;    Hilbert time-frequency spectrums;    Intrinsic Mode functions;    Metal cutting process;    Recognition features;   

Others  :  https://iopscience.iop.org/article/10.1088/1757-899X/230/1/012049/pdf
DOI  :  10.1088/1757-899X/230/1/012049

学科分类：材料科学（综合）

来源: IOP

【 摘 要 】

In the metal cutting process, the signal contains a wealth of tool wear state information. A tool wear signal's analysis and feature extraction method based on Hilbert marginal spectrum is proposed. Firstly, the tool wear signal was decomposed by empirical mode decomposition algorithm and the intrinsic mode functions including the main information were screened out by the correlation coefficient and the variance contribution rate. Secondly, Hilbert transform was performed on the main intrinsic mode functions. Hilbert time-frequency spectrum and Hilbert marginal spectrum were obtained by Hilbert transform. Finally, Amplitude domain indexes were extracted on the basis of the Hilbert marginal spectrum and they structured recognition feature vector of tool wear state. The research results show that the extracted features can effectively characterize the different wear state of the tool, which provides a basis for monitoring tool wear condition.