Multi-branch global Transformer‐assisted network for fault diagnosis

Fault Diagnosis (FD) is critical in smart manufacturing, enabling predictive maintenance, reducing operational costs, and enhancing system reliability. To deal with this task more accurately, this paper proposes a generative, effective, and novel framework, a multi-branch global Transformer-assisted...

Celý popis

Uloženo v:

Podrobná bibliografie
Vydáno v:	Applied soft computing Ročník 182; s. 113572
Hlavní autoři:	Shao, Xiaorui, Kim, Chang-Soo
Médium:	Journal Article
Jazyk:	angličtina
Vydáno:	Elsevier B.V 01.10.2025
Témata:	1DCNN Fault diagnosis multi-supervised algorithm multi-view global features Transformer LN PV TAS MHA DSMT-1DCNN multi-view global features DL-MSCNN RNN MBGTNet FFT DWT WARS WDCNNProposed IR GMP RF FD MSCNN WDCNN ANN Transformer CORAL GFC LFCS SVM BN SA AWF Fault diagnosis 2DCNN NLP CBAM SNR ACEL ProposedWo-MSS multi-supervised algorithm ViT CWT 1DCNN CE SU LSTM OR ProposedWo-LFCS MBLF MSS MGF CNN-LSTMProposed WSCNN-GMPProposed MBG-WideConv1D MFPT MSCNN-LSTMPropsoed CNNCombine DB
ISSN:	1568-4946
On-line přístup:	Získat plný text
Tagy:	Přidat tag Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!

Popis
Shrnutí:	Fault Diagnosis (FD) is critical in smart manufacturing, enabling predictive maintenance, reducing operational costs, and enhancing system reliability. To deal with this task more accurately, this paper proposes a generative, effective, and novel framework, a multi-branch global Transformer-assisted network (MBGTNet), for accurate FD. First, a multi-branch global-wide one-dimension convolution operation (MBG-WideConv1D) is proposed to obtain global features in different views. Meanwhile, a Transformer assist scheme (TAS) is designed to leverage the Transformer's global feature extraction capacity. The features extracted by the Transformer are fused with those extracted with MBG-WideConv1D by minimizing their pairwise correlation alignment (CORAL) distances. Benefiting from the well-designed MBG-WideConv1D and TAS, the global features hidden in the raw signals are fully extracted from multiple viewpoints. Each branch of global features is then fed into a one-dimension convolutional neural network (1DCNN) to extract local features in a multi-supervised scheme (MSS) that helps each branch learn thoroughly. Furthermore, the proposed method employs a local feature correlation enhancement scheme (LFCS) to reduce distribution differences and increase robustness among the local features of each branch. As a result, the final features used for FD are a fusion of multi-view global and local features with strong robustness, enabling accurate FD in noisy environments. Comparative experiments on four datasets, including CWRU, MFPT, SU Bearing, and SU Gear, validate the proposed method's effectiveness, achieving over 99.6 % accuracy across four datasets. Moreover, the TAS and LFCS's generalities have been demonstrated on two 1DCNNs and hybrid CNN-LSTMs with four subsets. Also, the effectiveness of each component in the proposed framework has been thoroughly analyzed. •The TAS ensures global feature extraction without meta-information loss.•The MB-WideConv1D extracts multi-global features.•The MSS guides each 1DCNN to learn local features thoroughly.•The LFCS increases the robustness of local features.•MBGTNet achieves an average accuracy of over 99.6 % for fault diagnosis.
ISSN:	1568-4946
DOI:	10.1016/j.asoc.2025.113572