Rotating machinery fault diagnosis using dimension expansion and AntisymNet lightweight convolutional neural network

Abstract

Deep learning-based methods have made remarkable progress in the field of fault diagnosis for rotating machinery. However, convolutional neural networks (CNNs) are not suitable for industrial applications due to their large model size and high computational complexity. To address this limitation, this paper proposes the Antisym module and constructs AntisymNet, which is combined with dimension expansion algorithms for fault diagnosis of rotating machinery. To begin with, the original vibration signal of the rolling machinery is subjected to time-frequency transformations using the Discrete Fourier Transform (DFT) and Discrete Wavelet Transform (DWT). Subsequently, each transformed time-frequency signal is expanded in dimensions, resulting in two-dimensional matrix single channel images. These single channel images are then fused into RGB images to enhance the sample features. Finally, the proposed AntisymNet is utilized for recognizing and classifying the expanded signals. To evaluate the performance of AntisymNet, the MiniImageNet image dataset is employed as a benchmark, and a comparison is made with other state-of-the-art lightweight convolutional neural networks (L-CNNs). Additionally, the effectiveness of the proposed fault diagnosis model is validated using the CWRU bearing dataset, Ottawa bearing dataset, and the hob dataset. The model achieves an impressive accuracy rate of 99.70% in the CWRU dataset, 99.26% in the Ottawa dataset, and an error rate of only 0.66% in the hob dataset. These results demonstrate the strong performance of the proposed fault diagnosis model.

Review of research on signal decomposition and fault diagnosis of rolling bearing based on vibration signal

“…Fault category Experiment samples Diagnosis accuracy (%) [227] FKP-SGECNN 10 -99.63 [228] Improved graph convolutional network (GCN) 28 1400 99.12 [229] BCMFDE-RF-mRMR-KNN 10 550 99.09 [230] NWMF-CNN 10 4000 99.80 [231] Reinforcement neural architecture search CNN 12 3600 99.65 [232] Dimension expansion and AntisymNet lightweight CNN 10 10 000 99.70 [233] Multi-scale weighted graph-MCGCN 10 3000 99.45 [234] Improved (ICEEMDAN)-ICA-FuEn 10 10 000 99.91 [235] MAM-DSDCNN 7 2800 99.63 [236] Sparse representation deep learning (SR-DEEP) 4 2000 100.00 [237] Online sequential extreme learning machine (OS-ELM) 4 9466 99.62 [238] IFE + CBAM-enhanced InceptionNet 10 1000 99.5 [239] SSCL method based on MSA mechanism and MCL 10 2000 99.97 [240] MRDNN-AG 10 120 000 98.85 [241] AMCEEMD-1DCNN 7 3500 99.50 [242] Modified AlexNet-SVM 4 -99.60 [243] FC-CLDCNN 10 10 000 99.95 [244] PCA-ICEEMDAN and BiLSTM-SCN-CCAM 10 1024 99.92 [245] 2ADA + MK-MMD 10 1960 99.76 [246] 1D feature matching domain adaptation 3 9000 100.00 [247] ICEEMDAN-Hilbert transform-CBAM 10 30 000 95.2 [248] Ensemble MSRCNN-BiLSTM 4 4800 98.43 [249] WKN-BiLSTM-AM 10 1750 99.7 [250] MVO-MOMEDA-SVM 4 400 92.50 [251] WPDPCC-DGCL 10 6000 98.65 [252] I-PixelHop framework based on Spark-GPU 10 -98.93…”