MEMS photoacoustic sensor transformer fault diagnosis method based on GAM-MTN

• Autorzy: Peng Jiaqi , Fang Baomin , Li Shou , Luo Jin , Zhu Xiaobin , Ma Shunqing

Identyfikatory

DOI

10.29354/diag/202868

• Języki publikacji: EN

Abstrakty

EN

As the service life of electrical equipment increases, it may suffer from various faults, such as overheating, partial discharge, etc., resulting in the generation of dissolved gases in the oil. MEMS photoacoustic sensors analyze acoustic signals through photoacoustic spectroscopy and signal processing technology to detect the concentration of dissolved gases in oil. Regarding the data traits of photoacoustic sensors, this document suggests a graph mutual mapping transmission network (GAM-MTN) method. First, an improved aggregation weight graph convolutional neural network is introduced, and the node aggregation weight function is designed using the Manhattan distance metric, so that the model can dynamically adjust the aggregation weight according to the similarity between nodes during the message passing aggregation process. Secondly, the graph mutual mapping transmission network is proposed to achieve uniform spread of origin field and destination field samples through sample mapping technology, which effectively improves the overall migration effect of the model. Finally, unsupervised adaptation of the classifier and domain discriminator is utilized to enhance the generalization capability of the system. Test outcomes demonstrate that the suggested GAM-MTN network can effectively improve the learning efficiency and diagnosis accuracy of transformer fault characteristics. Compared with other advanced neural network models, the recognition accuracy is as high as 96.37%.

Słowa kluczowe

EN

MEMS photoacoustic sensor; dissolved gas in oil; graph convolutional neural networks; chart reciprocal mapping transfer web; defect detection

• Wydawca: Polskie Towarzystwo Diagnostyki Technicznej
• Czasopismo: Diagnostyka
• Rocznik: 2025
• Tom: Vol. 26, No. 2
• Strony: art. no. 2025201
• Opis fizyczny: Bibliogr. 21 poz., rys.

Twórcy

autor

Peng Jiaqi

• Haidong Power Supply Company, State Grid Qinghai Provincial Electric Power Company, Qinghai 810699, China

• https://orcid.org/0009-0003-9825-9685

autor

Fang Baomin

• Haidong Power Supply Company, State Grid Qinghai Provincial Electric Power Company, Qinghai 810699, China

• https://orcid.org/0009-0008-1694-4619

autor

Li Shou

• Haidong Power Supply Company, State Grid Qinghai Provincial Electric Power Company, Qinghai 810699, China

• https://orcid.org/0009-0007-4546-6989

autor

Luo Jin

• Haidong Power Supply Company, State Grid Qinghai Provincial Electric Power Company, Qinghai 810699, China

• https://orcid.org/0009-0000-4717-7688

autor

Zhu Xiaobin

• Haidong Power Supply Company, State Grid Qinghai Provincial Electric Power Company, Qinghai 810699, China

• https://orcid.org/0009-0000-4932-7037

autor

Ma Shunqing

• Haidong Power Supply Company, State Grid Qinghai Provincial Electric Power Company, Qinghai 810699, China

• https://orcid.org/0009-0009-0499-2637

Bibliografia

• 1. Shao YY, Wang X, Peng P, et al. Intelligent diagnosis method for dry transformer condition based on machine vision and auditory response. Noise and Vibration Control. 2024;44(04):199-204.
• 2. Sun T, Chen X, Du MS, et al. Transformer fault diagnosis method based on DGA using Light GBMICOA-CN. Proceedings of the CSEE, 1-10.
• 3. Yang X, Zhou W, Cheng L, et al. Research on fault diagnosis of transformers based on multi-source information fusion using KPCA-SVM. 2024 4th International Conference on New Energy and Power Engineering (ICNEPE). 2025. https://doi.org/10.1109/ICNEPE64067.2024.108605 20.
• 4. Wang FR, Li Z. Transformer fault diagnosis model based on SCSSA-BiLSTM. Southern Power System Technology. 2024:1-9.
• 5. Wen J, Yuan J, Zheng Q, et al. Hierarchical domain adaptation with local feature patterns. Pattern Recognit. 2022;124:108445. https://doi.org/10.1016/j.patcog.2021.108445.
• 6. Hu T, Guo Y, Gu L, et al. Remaining useful life prediction of bearings under different working conditions using a deep feature disentanglement based transfer learning method. Reliab. Eng. Syst. Saf. 2022; 219:108265. https://doi.org/10.1016/j.ress.2021.108265.
• 7. Li T. Wavelet Kernel Net: An interpretable deep neural network for industrial intelligent diagnosis. IEEE Trans. Syst., Man, Cybern., Syst. 2022;52(4): 2302-2312. https://doi.org/10.1109/TSMC.2020.3048950.
• 8. Wang D, Chen Y, Shen C, Zhong J, Peng Z, and Li C. Fully interpretable neural network for locating resonance frequency bands for machine condition monitoring. Mech. Syst. Signal. Process. 2022;168: 108673. https://doi.org/10.1016/j.ymssp.2021.108673.
• 9. Ganin Y. Domain-adversarial training of neural networks. J. Mach. Learn. Res. 2016;17(1):2096-2030.
• 10. Zhao D, Zhang H, Liu S, Wei Y, and Xiao S. Deep rational attention network with threshold strategy embedded for mechanical fault diagnosis. IEEE Trans. Instrum. Meas. 2021;70:3519715. https://doi.org/10.1109/TIM.2021.3085951.
• 11. Scarselli F, Gori M, Tsoi AC, et al. The graph neural net-work model J. IEEE Transactions on Neural Networks. 2008;20(1):61-80. https://doi.org/10.1109/TNN.2008.2005605.
• 12. Li T, Zhao Z, Sun C, et al. Domain adversarial graph convolutional network for fault diagnosis under variable working conditions. IEEE Trans. Instrum. Meas. 2021;70:1-10. https://doi.org/10.1109/TIM.2021.3075016.
• 13. Bruna J, Mallat S. Invariant scattering convolutionnetworks. IEEE Transactions on Pattern Analysis and Machine Intelligence. 2013;35(8):1872-1886. https://doi.org/10.1109/TPAMI.2012.230.
• 14. Defferrard M, Bresson X, Vandergheynst P. Convolutional neural networks on graphs with fast localized spectral filtering. Advances in Neural Information Processing Systems. 2016;29:3844-3852.
• 15. Pang S, Yang X, Zhang X, et al. Fault diagnosis of rotating machinery components with deep ELM ensemble induced by real-valued output-based diversity metric. Mech. Syst. Signal Process. 2021;159:107821. https://doi.org/10.1016/j.ymssp.2021.107821.
• 16. Goay CH, Ahmad NS, Goh P. Transient simulations of high-speed channels using CNN-LSTM with an adaptive successive halving algorithm for automated hyperparameter optimizations. IEEE Access. 2021;9: 127644-127663. https://doi.org/10.1109/ACCESS.2021.3112134.
• 17. Zhu J, Ju Y, Xia M. Vehicle recognition model based on improved CNN-SVM. in Proc. 2nd Int. Seminar Artif. Intell., Netw. Inf. Technol. 2021:294-297. https://doi.org/10.1109/AINIT54228.2021.00065.
• 18. Jin T, Yan C, Chen C, et al. Light neural network with fewer parameters based on CNN for fault diagnosis of rotating machinery. Measurement. 2021;181:109639. https://doi.org/10.1016/j.measurement.2021.109639.
• 19. Lavasani HM, Wanling P, Harrington B, et al. A 76 dB 1.7 GHz 0.18 μm CMOS tunable TIA using broadband current pre-amplifier for high frequency lateral MEMS oscillators. IEEE J. Solid-State Circuits. 2011;46:224-235. https://doi.org/10.1109/JSSC.2010.2085890.
• 20. Kamada Y, Isobe A, Oshima T, et al. Capacitive MEMS accelerometer with perforated and electrically separated mass structure for low noise and low power. J. Microelectromech. Syst. 2019;28(3):401-408. https://doi.org/10.1109/JMEMS.2019.2903349.
• 21. Akita I, Okazawa T, Kurui Y, et al. A feedforward noise reduction technique in capacitive MEMS accelerometer analog front-end for ultra-low-power IoT applications. IEEE J. Solid-State Circuits. 2020;55(6):1599-1609. https://doi.org/10.1109/JSSC.2019.2952837.