TY - JOUR
T1 - A new mechanical compound fault diagnosis method based on morphological filter and RLS
AU - Xue, Zhuo
AU - He, Dan
AU - Ni, Ze Xing
AU - Wang, Xiufeng
N1 - Publisher Copyright:
© 2023 IOP Publishing Ltd.
PY - 2023/11
Y1 - 2023/11
N2 - Mechanical compound fault diagnosis is a thorny issue in the industry. To overcome this problem, a method named morphological filter—recursive least squares (MF-RLS) is proposed in this paper. In the proposed method, MF-RLS is used to sequentially separate the different fault impulse features by decomposing the observed signal into a series of iterative morphological filtering components (IMFCs). First, the measured signal is decomposed into different scales by the multi-scale MF. Then, the product of kurtosis and envelope harmonic-to-noise ratio index is used to select the best IMFC. Finally, the IMFC is input to RLS to separate other fault features. After continuous iterations, the separation and extraction of the compound fault impulse features are achieved. The simulation and experiment of the mechanical compound fault have verified the effectiveness of the proposed method.
AB - Mechanical compound fault diagnosis is a thorny issue in the industry. To overcome this problem, a method named morphological filter—recursive least squares (MF-RLS) is proposed in this paper. In the proposed method, MF-RLS is used to sequentially separate the different fault impulse features by decomposing the observed signal into a series of iterative morphological filtering components (IMFCs). First, the measured signal is decomposed into different scales by the multi-scale MF. Then, the product of kurtosis and envelope harmonic-to-noise ratio index is used to select the best IMFC. Finally, the IMFC is input to RLS to separate other fault features. After continuous iterations, the separation and extraction of the compound fault impulse features are achieved. The simulation and experiment of the mechanical compound fault have verified the effectiveness of the proposed method.
KW - compound fault diagnosis
KW - feature extraction
KW - morphological filter
KW - recursive least squares
UR - https://www.scopus.com/pages/publications/85166484539
U2 - 10.1088/1361-6501/ace7e9
DO - 10.1088/1361-6501/ace7e9
M3 - 文章
AN - SCOPUS:85166484539
SN - 0957-0233
VL - 34
JO - Measurement Science and Technology
JF - Measurement Science and Technology
IS - 11
M1 - 115901
ER -