Mlamsan-based cross-case fault diagnosis of rotor-bearing systems

Deep Learning Approaches for Bearing Fault Diagnosis Using Motor Current Signals

Motor Current Signature Analysis (MCSA) is a non-invasive technique that enables the detection of bearing faults in rotating electrical machines without the need for additional sensors. In this study, the Paderborn University bearing dataset was utilized to perform a two-stage analysis. In the first stage, motor current data were processed directly using 1-Dimensional Convolutional Neural Networks (1D-CNN). In the second stage, scalogram images obtained via Continuous Wavelet Transform (CWT) were used to train five different deep learning models, with the ResNet18-based 2D-CNN model providing the best performance. To prevent data leakage, the training and testing sets were partitioned based on individual bearings to ensure complete isolation. The experimental results demonstrated that both 1D-CNN and ResNet18-based 2D-CNN methods achieved 100% accuracy in detecting outer race faults. However, it was observed that the impact of inner race faults on the stator current remains weak due to the complex physical transmission path of the fault signal, resulting in significantly lower detection rates.

Robust Bearing Fault Diagnosis Using Transfer Learning and SISA-Based Machine Unlearning

A fault diagnosis method for aero-engine inter-shaft bearings based on 1DCNN-Transformer-BiGRU.

To address the challenges of difficult fault feature extraction and feature aliasing in aero-engine inter-shaft bearings under strong noise conditions, this paper proposes a fusion diagnostic method that integrates a dual-scale one-dimensional convolutional neural network, a multi-head self-attention Transformer, and a bidirectional gated recurrent unit. The method employs a three-stage progressive network architecture for end-to-end fault diagnosis. The dual-scale 1DCNN extracts local temporal features from vibration signals, and batch normalization and dropout are applied to stabilize training and reduce potential overfitting. The Transformer encoder models dependencies among the extracted feature representations, supporting the representation of fault-sensitive features. The BiGRU captures bidirectional temporal dependencies in the fault evolution process. Experimental validation on the Harbin Institute of Technology aero-engine inter-shaft bearing dataset shows that the proposed model achieves 97% diagnostic accuracy under extreme noise conditions (SNR = -5 dB). Compared with existing methods, these results indicate that the proposed network effectively maintains diagnostic performance under controlled noise conditions.

Fault Diagnosis of Propulsion Shaft Bearings Based on Deep Multi-Scale Multi-Source Subdomain Adaptation

To improve the fault diagnosis accuracy of marine propulsion shaft bearings under variable working conditions and with limited samples, a Multi-scale Deep Multi-source Subdomain Adaptation Network (MDMSAN) is proposed. First, DCGAN is employed to augment source domain samples. A shared feature extraction network integrating multi-scale convolutions, LSTM, and ECA mechanism is constructed to extract domain-invariant features. Private feature extractors are designed for each source-target domain pair, utilizing LMMD for fine-grained subdomain alignment. A transferability-aware weighted fusion strategy integrates multi-source classification results. Experiments on CWRU and PT500 datasets demonstrate that MDMSAN achieves average accuracies of 99.92% and 99.73% under variable working conditions, and maintains 98.85% accuracy with only 1/16 of the target samples, significantly outperforming existing methods.

Bearing Fault Diagnosis Based on a Spectral-Guided Adaptive Multi-Scale Convolutional Network

To address the degradation of cross-condition diagnostic performance caused by feature-scale drift in rolling bearing vibration signals under variable operating conditions, this paper proposes a spectral-guided adaptive multi-scale convolutional neural network (SAMACNN). First, PSD sequences and time-frequency features are introduced as dual-stream inputs. While the time-frequency main branch extracts local information, the Spectral Transformer spectral bypass branch captures long-range dependencies in harmonic structures. Second, dynamic gating weights are generated for the multi-scale convolutional branches, enabling sample-conditioned multi-scale feature selection and fusion and alleviating scale mismatch caused by fixed receptive fields and static fusion. Finally, data collected from two bearing fault simulation test rigs are used to verify the effectiveness and superiority of the proposed algorithm. The experimental results show that the proposed SAMACNN method achieves average accuracies of 97.16% and 97.56% on the two datasets, respectively, outperforming the ablation variants and demonstrating strong robustness and generalization capability in complex variable-condition measurement environments.

Motor-Current-Based Bearing Fault Detection Under Unseen Operating Conditions

Reliable motor-current-based bearing diagnosis requires evaluation on unseen physical bearings and operating conditions. This study uses the Paderborn University benchmark, acquired from a 425 W permanent-magnet synchronous motor (PMSM) test rig, to evaluate time–frequency deep transfer learning under strict bearing-level grouping. Four representations—continuous wavelet transform (CWT), short-time Fourier transform (STFT), wavelet synchrosqueezed transform (WSST), and Fourier synchrosqueezed transform (FSST)—are combined with pretrained CNN backbones across four binary targets: aged-only A/B and artificial-plus-aged C/D, with mixed-fault bearings excluded/included within each pair. The workflow includes pooled-condition candidate discovery, exploratory Main-split leave-one-operating-condition-out (LOCO) screening, and a retrospective multi-split LOCO audit. The audit contains 288 crossed condition–split–seed evaluations. Because pooled test summaries and Main-split LOCO results informed later stages, these evaluations provide descriptive robustness evidence rather than an independent post-selection test. Target A achieved the highest all-split mean balanced accuracy (0.736 for CWT–EfficientNetB0). The pairs for Targets B and C were near-ties, and the Target D ordering reversed when Main was excluded. Across the eight audited candidates, mean sensitivity ranged from 0.618 to 0.948, whereas specificity ranged from 0.092 to 0.564. Target D combined approximately 0.89 sensitivity with an approximately 0.90 false-alarm rate. Thus, operating condition, fault-class composition, bearing split, and error-cost priorities all affect model interpretation. A matched current-domain baseline audit added 336 evaluations using handcrafted-feature RBF–SVM and Random Forest models and a compact raw-current 1D-CNN. The results show that instability is broader than the TF–CNN pipeline but is not uniform across model families: TF candidates were clearly stronger for Targets A and C, feature-based models were stronger for Target B, and Target D remained mixed and protocol-sensitive. A complementary bearing-level source-group analysis quantified six healthy/fault-source categories across the 37 current features; among the eight features with the largest mean between-group variance fraction, only spectral entropy and dominant power fraction preserved the same mixed-versus-non-mixed contrast direction across all four operating conditions. An additional matched Target D reference held the binary target, physical-bearing split, held-out condition, seed, fourth-order FSST representation, ResNet-50 backbone, and training settings fixed while changing the sensing channel. Across 24 matched runs, vibration showed descriptively higher mean balanced accuracy (0.642 vs. 0.503) and specificity (0.476 vs. 0.146), while sensitivity was slightly lower (0.807 vs. 0.859); the comparison does not establish universal modality superiority. Pooled-condition performance is useful for candidate discovery, but credible condition-generalization claims require explicit separation of exploratory selection and confirmatory testing. The numerical findings are specific to the evaluated PMSM benchmark and do not establish universal performance across electric-machine types.