A Signal Generation Fault Diagnosis Method for Planetary Reducer of High‐Pressure Grinding Roll DOI
Ronghua Chen,

Yiliu Gu,

Guangqi Qiu

et al.

Quality and Reliability Engineering International, Journal Year: 2024, Volume and Issue: 41(2), P. 704 - 718

Published: Nov. 25, 2024

ABSTRACT This paper proposes a signal generation fault diagnosis method for the challenge of insufficient training samples in planetary reducer high‐pressure grinding roll (HPGR). The expression vibration response is derived, and model established HPGR. To generate expanded samples, an adaptive fox optimization (AFO) algorithm employed optimizing parameters model, so that simulated matches measured signal. Before that, HPGR preprocessed by tunable Q‐factor wavelet transform (TQWT). Finally, residual convolution neural long short‐term memory network with attention mechanism (ResCNN‐LSTM‐ATT) utilized feature extraction from collected signals, hierarchical classifier developed classification. experimental results show AFO has better performance, signals generated optimized are good agreement signals. classification accuracy normal reaches 99.70%, which shows both known unknown identification. Compared to other methods, proposed diagnostic accuracy.

Language: Английский

A performance-interpretable intelligent fusion of sound and vibration signals for bearing fault diagnosis via dynamic CAME DOI
You Keshun,

Lian Zengwei,

Yingkui Gu

et al.

Nonlinear Dynamics, Journal Year: 2024, Volume and Issue: 112(23), P. 20903 - 20940

Published: Aug. 24, 2024

Language: Английский

Citations

13
A Signal Generation Fault Diagnosis Method for Planetary Reducer of High‐Pressure Grinding Roll DOI
Ronghua Chen,

Yiliu Gu,

Guangqi Qiu

et al.

Quality and Reliability Engineering International, Journal Year: 2024, Volume and Issue: 41(2), P. 704 - 718

Published: Nov. 25, 2024

ABSTRACT This paper proposes a signal generation fault diagnosis method for the challenge of insufficient training samples in planetary reducer high‐pressure grinding roll (HPGR). The expression vibration response is derived, and model established HPGR. To generate expanded samples, an adaptive fox optimization (AFO) algorithm employed optimizing parameters model, so that simulated matches measured signal. Before that, HPGR preprocessed by tunable Q‐factor wavelet transform (TQWT). Finally, residual convolution neural long short‐term memory network with attention mechanism (ResCNN‐LSTM‐ATT) utilized feature extraction from collected signals, hierarchical classifier developed classification. experimental results show AFO has better performance, signals generated optimized are good agreement signals. classification accuracy normal reaches 99.70%, which shows both known unknown identification. Compared to other methods, proposed diagnostic accuracy.

Language: Английский

Citations

0