Frequency slice graph spectrum model and its application in bearing fault feature extraction

Mechanical Systems and Signal Processing, Journal Year: 2025, Volume and Issue: 226, P. 112383 - 112383

Published: Jan. 22, 2025

Language: Английский

---

Fault diagnosis of rotating machinery with high-dimensional imbalance samples based on wavelet random forest

Measurement, Journal Year: 2025, Volume and Issue: unknown, P. 116936 - 116936

Published: Feb. 1, 2025

Language: Английский

---

A noise reduction method for rolling bearing based on improved Wiener filtering

Review of Scientific Instruments, Journal Year: 2025, Volume and Issue: 96(2)

Published: Feb. 1, 2025

To accurately identify compound faults of bearings, a new noise reduction method is presented. With the method, input signals and order Wiener filtering are adaptively determined according to feature mode decomposition (FMD), signal evaluation index, Euclidean distance. First, effectively separate frequency components from vibration signals, decomposed into modal based on FMD algorithm; second, kurtosis, root mean square, variance, which sensitive fault information, selected build vectors. Third, distance between vectors component original calculated represent correlation among signals. By acquiring two that have greatest least an actual mixed required by can be determined. Furthermore, with maximum kurtosis as criterion. Finally, features extracted through spectral analysis after type judged that. demonstrate accuracy effectiveness proposed compared classical method. The result comparison shows presented restrict more determine complex bearings accurately.

Language: Английский

---

Bearing Fault Diagnosis Grounded in the Multi-Modal Fusion and Attention Mechanism

Applied Sciences, Journal Year: 2025, Volume and Issue: 15(3), P. 1531 - 1531

Published: Feb. 3, 2025

This paper proposes a novel method called Fusion Attention Network for Bearing Diagnosis (FAN-BD) to address the challenges in effectively extracting and fusing key information from current vibration signals traditional methods. The research is validated using public dataset Vibration, Acoustic, Temperature, Motor Current Dataset of Rotating Machines under Varying Operating Conditions Fault Diagnosis. first converts into two-dimensional grayscale images, extracts local features through multi-layer convolutional neural networks, captures global self-attention mechanism Vision Transformer (ViT). Furthermore, it innovatively introduces Channel-Based Multi-Head (CBMA) efficient fusion different modalities, maximizing complementarity between signals. experimental results show that compared mainstream algorithms such as Transformer, Swin ConvNeXt, achieves higher accuracy robustness fault diagnosis tasks, providing an reliable solution bearing diagnosis.The proposed model outperforms ViT, CBMA-ViT terms classification accuracy, achieving 97.5%. comparative clearly demonstrate yields significant improvements outcomes.

Language: Английский

---

Markov model based circular frequency feature extraction method for electronic communication signal anti-jamming

Applied Mathematics and Nonlinear Sciences, Journal Year: 2025, Volume and Issue: 10(1)

Published: Jan. 1, 2025

Abstract This paper proposes a wavelet packet thresholding noise reduction algorithm for the problem of large signal frequency deviation and timing error caused by interference electronic communication signals, at same time researches spectral overlapping separation based on cyclic smoothness signal. The sets up shift filter, including portion conjugate receiving end, which facilitates using temporal correlation In order to solve feature extraction problem, discrete hidden Markov model structure is used combining characteristic intervals waveform. Simulation experiments confirm that in this has advantage lower signal-to-noise ratios, when ratio as low −15 dB, monitoring success rate matched filter about six times higher than traditional filter. comparison with SVM method BP method, completeness results paper’s reaches 95% average, accuracy always stays above 91%, extracted features have high accuracy. It shows excellent performance designed can make problems such be improved some extent.

Language: Английский

---

Memory-Augmented Prototypical Meta-Learning Method for Bearing Fault Identification under Few-Sample Conditions

Neurocomputing, Journal Year: 2025, Volume and Issue: unknown, P. 129996 - 129996

Published: March 1, 2025

Language: Английский

---

An adaptive informative band extraction method with an improved DTCWPT and impulsiveness-to-sparsity evaluation ratio for bearing diagnostics

Mechanical Systems and Signal Processing, Journal Year: 2025, Volume and Issue: 231, P. 112682 - 112682

Published: April 8, 2025

Language: Английский

---

MT-CDGAT: A multi-label diagnosis model for untrained planetary gearbox compound faults based on multi-task cross dynamic graph attention networks

Neurocomputing, Journal Year: 2025, Volume and Issue: unknown, P. 130307 - 130307

Published: April 1, 2025

Language: Английский

---

A Lightweight Cross-Axis Semantic Interaction Network with Receptive-Field-Based Attention for Industrial Surface Defect Detection

Expert Systems with Applications, Journal Year: 2025, Volume and Issue: unknown, P. 127949 - 127949

Published: May 1, 2025

Language: Английский

---

A channel-independent network based on wavelet enhancement for long-term time series forecasting

Engineering Applications of Artificial Intelligence, Journal Year: 2025, Volume and Issue: 154, P. 110964 - 110964

Published: May 10, 2025

Language: Английский

---