Addressing the scarcity of data reflecting realistic variable-speed operating conditions in current research on aero-engine inter-shaft bearing fault diagnosis, and overcoming limitations in existing methods concerning time-frequency resolution, global cross-channel interaction capability, and robustness under variable speeds, we proposed a High-Channel Triple Attention and LSK Module Fusion Network (HC-TSKNet) characterized by high precision and strong robustness. The method first employs Synchrosqueezing Wavelet Wavelet Packet Transform (SSWPT) to transform the raw vibration signals into high-resolution time-frequency spectrograms. This effectively mitigates the spectral smearing issue prevalent in conventional time-frequency analysis such as Continuous Wavelet Transform (CWT), significantly enhancing the clarity of discriminative fault signatures. The core innovation of HC-TSKNet lies in its synergistic integration of three key techniques: through an increased channel count, the model size is compressed and the per-sample inference time is reduced, while preserving a satisfactory level of diagnostic accuracy; utilizing a Triplet Attention mechanism to capture global cross-channel and spatial interaction dependencies, overcoming the limitations in cross-dimensional interaction inherent in conventional attention mechanisms like CBAM and SENet; and introducing a Large Selective Kernel (LSK) module to adaptively weight local features extracted by deep convolutional kernels of different scales spatially, enabling dynamic focus on localized fault patterns under variable-speed conditions. Validation experiments conducted on an inter-shaft bearing fault dataset from aero-engine dual-rotor simulator test rig demonstrate that HC-TSKNet achieves exceptional diagnostic accuracy of 99.75% under variable-speed conditions, significantly outperforming mainstream methods including Swin-Transformer and recent state-of-the-art models in related domains. It is particularly noteworthy that the model successfully overcomes significant variations in signal characteristics induced by differences in sensor mounting positions, demonstrating robust adaptability to diverse sensor configurations. Ablation studies, hyperparameter experiment, and robustness validation further substantiate the effectiveness of its key modules, the superiority of its structural design, and its substantial potential for engineering applications.
10 August 2026 in Measurement Science and Technology
To address the challenge of accurately identifying and evaluating fault development stages in the full-lifecycle fault diagnosis of rolling bearings caused by dynamically evolving fault morphologies, we propose a high-precision diagnosis method based on synchrosqueezed wavelet packet transform (SSWPT) and the DilateEPLKBNet deep learning model. First, SSWPT constructs high-quality time–frequency image samples that clearly represent the evolutionary laws of vibration signal characteristics throughout the bearing’s lifecycle. Second, the fault evolution state is comprehensively determined using root mean square, peak, kurtosis, and crest factors, dividing the lifecycle into initial incubation, progressive degradation, and critical failure phases. Subsequently, the novel DilateEPLKBNet diagnostic model is introduced. By integrating the DilateFormer structure with multi-scale dilated attention and the enhanced partial large kernel block (EPLKB) module, this model achieves synergistic optimization of multi-scale spatial features and adaptive channel attention, accurately capturing key discriminative features related to fault evolution in time–frequency images. Comprehensive experiments on the bearing accelerated lifecycle dataset from Xi’an Jiaotong University and Changxing Sumyoung Technology demonstrate leading performance, with an overall accuracy of 99.09% and a macro-average F 1-score of 98.49%. Hyperparameter and ablation experiments validate the optimal parameter selection and the architecture’s rationality. Furthermore, comparative experiments confirm the superiority of both the DilateEPLKBNet model and the SSWPT preprocessing method. Crucially, lifecycle stage transition experiments reveal the model’s ability to internalize the dynamic laws of continuous state evolution, proving its deep capacity to perceive the fault ‘evolution’ process rather than merely performing static classification. Ultimately, SSWPT-DilateEPLKBNet provides a highly effective solution for intelligent evolutionary condition monitoring and early fault warning throughout a bearing’s full lifecycle.
16 March 2026 in International Journal of Precision Engineering and Manufacturing