Aug 2026· 2026 IEEE International Conference on Mechatronics and Automation (ICMA)· pp. 1228-1233· 0 citations· 10 references
Abstract
Multi-sensor data fusion is widely recognized as a key enabler for reliable aero-engine condition monitoring under complex operating conditions. However, this paradigm inherently increases system complexity. This paper conducts a systematic empirical study to evaluate the practical efficacy and boundaries of deep feature mining using only a single channel. We propose a Local-Global State Space Model Network (LG-SSMNet), which uses multi-scale 1D CNNs to fit local transient pulse responses and employs a Selective State Space Model (Mamba) to construct a long-term periodic integrator, achieving feature decoupling within a single channel. To address the discretization divergence issue of continuous vibration signals, an empirical stabilization configuration involving dual normalization and stratified learning rates is introduced. Deep quantitative analysis on the HIT aero-engine bearing dataset shows that the four acceleration channels achieve stable accuracies over 99.6% (approaching 100%), while displacement channels show significant divergence. The underperforming channel (Sensor 1) reached only 91.31% due to a critical missed detection rate of outer race faults (4.3% misclassified as healthy). Through joint empirical evidence using ROC-AUC, F1-scores, and qualitative feature visualization, we demonstrate that with reasonably selected acceleration sensors, single-channel deep mining can serve as a practical alternative to multi-channel fusion.
A hybrid deep learning architecture is proposed for robust vibration-based fault diagnosis in industrial machinery by jointly modeling time-domain, frequency-domain, and temporal dynamics, enabling coherent cross-domain interaction and robust fault characterization.
Canan Taştimur· Information Technology and C...· 0 citations
In complex industrial environments, data-driven fault diagnosis of rolling bearings is often constrained by time-varying operating conditions, severe noise interference, and the ‘black-box’ nature of deep learning models. To overcome these challenges, this study systematically integrates an improved Harris Hawks optimization (IHHO) algorithm, a convolutional neural network, and a bidirectional long short-term memory network to establish a physics-consistent fault diagnosis framework. First, a multi-channel image fusion architecture based on continuous wavelet transform is constructed to effectively encode the deep spatiotemporal correlations embedded in triaxial vibration signals. To enhance model adaptability, a Cauchy mutation–based IHHO strategy is introduced for autonomous hyperparameter optimization, effectively alleviating premature convergence to local optima. The core innovation of this work lies in the development of a quantitative interpretability framework centered on a novel metric termed the energy capture ratio (ECR). By jointly employing SHAP and gradient-weighted class activation mapping visualization techniques, the proposed model is demonstrated to exhibit strong physical consistency, enabling adaptive suppression of noise-dominated channels while accurately focusing on theoretically relevant fault resonance frequency bands. Extensive experiments on the HUST, case western reserve university, and southeast university benchmark datasets demonstrate that the proposed framework achieves average diagnostic accuracies of 99.0%, 100.0%, and 99.0%, respectively; even under severe noise interference at −5 dB, the accuracy remains as high as 89.6%. Quantitative evaluations show that the principal-axis relative ECR (RECR) consistently remains above 1.0 across all fault conditions, providing strong evidence for the mechanistic reliability of the model. These findings indicate that the proposed method provides an accurate and reliable solution for predictive maintenance of bearings in practical engineering applications.
Tingliang Chen, Wenhai Zhou, Rui Liang et al.· Engineering Research Express· 0 citations
Slewing bearings in low-speed, heavy-load equipment generate weak and heterogeneous fault signatures that are difficult to characterize using a single sensor. This study develops a compact dual-branch feature-fusion framework that jointly exploits six-channel vibration and one-channel acoustic-emission (AE) signals. To prevent source-record leakage, complete raw recording groups are assigned to training, validation, and test subsets before segmentation; non-overlapping 1024-point windows are then generated, and channel normalization is fitted using training groups only. Five matched random-seed runs are performed, with the best checkpoint selected exclusively by validation Macro-F1. The fusion model achieves mean test accuracy of 99.10% and Macro-F1 of 0.9910, compared with 97.38%/0.9738 for vibration-only and 98.03%/0.9802 for AE-only. The improvement over vibration-only is statistically significant (p = 0.022 for both Accuracy and Macro-F1), whereas the improvement over AE-only is numerical but does not reach the 0.05 significance level (p = 0.063). The fusion model also obtains the lowest mean Davies-Bouldin index (0.963). These results support compact vibration-AE fusion as an effective diagnostic baseline while also defining its statistical and deployment limitations.
A multi-scale linear attention multi-source subdomain adaptation network (MLAMSAN) that integrates the multi-scale linear attention (MLA) that can achieve fault diagnosis under cross operating conditions through subdomain feature alignments that exhibits the superior diagnostic performance and the strong generalization ability.
Zheng Han, Yuqi Fan, Yaping Wang et al.· Engineering Research Express· 0 citations
Predicting the Remaining Useful Life (RUL) of rolling bearings is a central problem in predictive maintenance, but most existing data-driven solutions still rely on a single vibration channel and a CNN–LSTM architecture. This combination ignores cross-axis degradation information, depends on noise-sensitive Fourier features, carries a heavy parameter count from standard convolutions, and is limited by the sequential nature of LSTM. In this paper, we propose a Dual-Channel Wavelet Scattering Network(DC-WSNet) that addresses these four limitations in a single architecture. The model uses two parallel depthwise separable CNN branches to fuse the horizontal and vertical vibration channels. It then simultaneously extracts features from the wavelet, Fourier, and time-statistical domains, recalibrates the fused channels using an Efficient Channel Attention (ECA) module, and uses a Temporal Convolutional Network (TCN) for temporal modeling. Experimental and interpretability analysis demonstrate that the proposed model effectively captures the cross-axis degradation patterns and leverages multi-axis information for improved RUL prediction.
S. T, Poornesh M, Siddhartha Reddy Salla et al.· 2026 4th International Confe...· 0 citations
An unsupervised hybrid deep learning framework for unknown bearing fault diagnosis and severity assessment using vibration signals that combines Continuous Wavelet Transform, Convolutional Neural Networks, and Long Short-Term Memory autoencoders is presented.
Edris Shamsulhaq, Fikri Arif Wicaksana· Jambura Journal of Electrica...· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
