Explainable Fault Classification and Severity Diagnosis in Rotating Machinery Using Kolmogorov–Arnold Networks

Entropy, Год журнала: 2025, Номер 27(4), С. 403 - 403

Опубликована: Апрель 9, 2025

Rolling element bearings are critical components of rotating machinery, with their performance directly influencing the efficiency and reliability industrial systems. At same time, bearing faults a leading cause machinery failures, often resulting in costly downtime, reduced productivity, and, extreme cases, catastrophic damage. This study presents methodology that utilizes Kolmogorov–Arnold Networks—a recent deep learning alternative to Multilayer Perceptrons. The proposed method automatically selects most relevant features from sensor data searches for optimal hyper-parameters within single unified approach. By using shallow network architectures fewer features, models lightweight, easily interpretable, practical real-time applications. Validated on two widely recognized datasets fault diagnosis, framework achieved perfect F1-Scores detection high severity classification tasks, including 100% cases. Notably, it demonstrated adaptability by handling diverse types, such as imbalance misalignment, dataset. availability symbolic representations provided model interpretability, while feature attribution offered insights into types or signals each studied task. These results highlight framework’s potential applications, monitoring, scientific research requiring efficient explainable models.

Язык: Английский

---

Hybrid CNN-BiLSTM-MHSA Model for Accurate Fault Diagnosis of Rotor Motor Bearings

Mathematics, Год журнала: 2025, Номер 13(3), С. 334 - 334

Опубликована: Янв. 21, 2025

Rotor motor fault diagnosis in Unmanned Aerial Vehicles (UAVs) presents significant challenges under variable speeds. Recent advances deep learning offer promising solutions. To address extracting spatial, temporal, and hierarchical features from raw vibration signals, a hybrid CNN-BiLSTM-MHSA model is developed. This leverages Convolutional Neural Networks (CNNs) to identify spatial patterns, Bidirectional Long Short-Term Memory (BiLSTM) network capture long- short-term temporal dependencies, Multi-Head Self-Attention (MHSA) mechanism highlight essential diagnostic features. Experiments on rotor data preprocessed with Butterworth band-stop filters were conducted laboratory real-world conditions. The proposed achieves 99.33% accuracy identifying faulty bearings, outperforming traditional models like CNN (93.33%) LSTM (62.00%) recent including CNN-LSTM (98.87%), the Attention Recurrent Autoencoder Model (ARAE) (66.00%), Lightweight Time-focused Network (LTFM-Net) (96.67%), Wavelet Denoising (WDCNN-LSTM) (96.00%). model’s high stability varying conditions underscore its robustness, making it reliable solution for rolling bearing motors, particularly dynamic UAV applications.

Язык: Английский

---

Lightweight mechanical equipment fault diagnosis framework based on GCGAN-MDSCNN-ICA model

Scientific Reports, Год журнала: 2025, Номер 15(1)

Опубликована: Фев. 10, 2025

In response to the challenges posed by imbalanced failure diagnosis samples, limited labeled data, and significant computational costs in actual industrial production settings, this paper introduces a high-precision, low-resource, end-to-end fault framework. On one hand, we propose data augmentation method based on GCGAN, which combines CNN GRU construct core network structures for generator discriminator. We integrate novel Smoothed Hinge-Cross-Entropy loss function facilitate training process, effectively mitigating mode collapse vanishing gradient issues. other design lightweight model MDSCNN-ICA-BiGRU. By substituting standard convolutions with depthwise separable deeper channels, complexity is significantly reduced, facilitating effective extraction of multiscale spatial features. The improved Coordinate Attention (CA) mechanism filters out noise enhances high-frequency characteristics. Combined BiGRU, captures global temporal associations, achieving fusion spatiotemporal Experimental results demonstrate that proposed approach performs well both publicly available simulation datasets private laboratory datasets. Compared benchmark methods, GCGAN module augmentation, improving classification accuracy CNNs 10%. When compared classic convolutional networks such as DRSN WDCNN, our MDSCNN-ICA-BiGRU shows faster more stable convergence rates, near-100% test sets an average computation cost reduction approximately 70%. Even noisy environments, maintains high slow rate precision decay, indicating robustness generalization capabilities.

Язык: Английский

---

An Intelligent Onboard Condition Monitoring Method of Wheelset Bearings under Time-varying Speed

IEEE Transactions on Instrumentation and Measurement, Год журнала: 2025, Номер 74, С. 1 - 15

Опубликована: Янв. 1, 2025

Язык: Английский

---

Scalable bearing fault diagnosis using metaheuristic feature selection and machine learning for diverse operating conditions

Systems Science & Control Engineering, Год журнала: 2025, Номер 13(1)

Опубликована: Март 2, 2025

Язык: Английский

---

An efficient transfer fault diagnosis method integrating feature redundancy selection and multi-strategy parameter optimization

Expert Systems with Applications, Год журнала: 2025, Номер unknown, С. 127267 - 127267

Опубликована: Март 1, 2025

Язык: Английский

---

Early Soft Fault Diagnosis of Circuits Based on CPO-VMD

Lecture notes in electrical engineering, Год журнала: 2025, Номер unknown, С. 346 - 350

Опубликована: Янв. 1, 2025

Язык: Английский

---

A multi-scale deep neural networks for early fault diagnosis in rolling ball bearings

Soft Computing, Год журнала: 2025, Номер unknown

Опубликована: Май 13, 2025

Язык: Английский

---

An efficient bearing fault detection strategy based on a hybrid machine learning technique

Scientific Reports, Год журнала: 2025, Номер 15(1)

Опубликована: Май 28, 2025

Язык: Английский

---

A Spectral-Based Blade Fault Detection in Shot Blast Machines with XGBoost and Feature Importance

Journal of Sensor and Actuator Networks, Год журнала: 2024, Номер 13(5), С. 64 - 64

Опубликована: Окт. 9, 2024

The optimal functionality and dependability of mechanical systems are important for the sustained productivity operational reliability industrial machinery, have a direct impact on its longevity profitability. Therefore, failure system or any components would be detrimental to production continuity availability. Consequently, this study proposes robust diagnostic framework analyzing blade conditions shot blast machinery. explores spectral characteristics vibration signals generated by discriminative feature excitement. Furthermore, peak detection algorithm is introduced identify extract unique features present in magnitudes each signal spectrum. A importance then deployed as selection tool, these selected fed into ten machine learning classifiers (MLCs), with extreme gradient boosting (XGBoost (version 2.1.1)) core classifier. results show that XGBoost classifier achieved best accuracy 98.05%, cost-efficient computational cost 0.83 s. Other global assessment metrics were also implemented further validate model.

Язык: Английский

---