Reactor Fault Diagnosis Based on CNN-LSTM Hybrid Neural Network

DOI: https://doi.org/10.62517/jes.202402414

Author(s)

Xiangdong Li*, Hui Dan, Chunsheng Wang, Liuxin Nie

Affiliation(s)

Jiaozuo Guangyuan Electric Power Group Co., Ltd, Jiaozuo, Henan, China *Corresponding Author.

Abstract

The acoustic signals generated by electrical insulation defects and mechanical failures during the operation of dry reactor contain a large amount of equipment state information, which can be used as an important feature parameter for the diagnosis of defects and failures. In view of the insufficiency of a single neural network to extract the temporal sequence features, this paper proposes a reactor fault diagnosis algorithm based on CNN-LSTM network, using CNN to mine the local spatial feature information of the spectrogram, and LSTM to mine the temporal sequence information of the spectrogram. Physical experimental platforms for electrical insulation defects and mechanical faults in reactors were established. Three fault samples were created for each fault type, and acoustic wave data and acoustic signals were collected during the evolution of insulation defects and mechanical faults in dry-type reactors. Experiments show that the method in this paper significantly improves the feature mining ability of the spectrogram, and effectively improves the fault diagnosis accuracy.

Keywords

Reactor; Acoustic Signal; Spectrogram Feature Extraction; Neural Network; Fault Diagnosis

References

[1] XIAN Richang, LU Yao, CHEN Lei, et al. Fault characteristics of an inter-turn short circuit of a dry-type air core series reactor. Power System Protection and Control, 2021, 49(18): 10-16. [2] WEI Xu, WU Shuyu, JIANG Ning, et al. Theoretical and Experimental Analyses on High-voltage Reactor Vibration Characteristics. High Voltage Apparatus, 2019, 55(11): 66-73. [3] YU Changting, LI Dajian, CHEN Liangyuan, et al. Transformer Fault Diagnosis Technique Based on Voiceprint and Vibration. High Voltage Apparatus, 2019, 55(11): 248-254. [4] Lecun Y, Bottou L, Bengio Y, et al. Gradient-based learning applied to document recognition. Proceed-ings of the IEEE, 1998,86(11): 2278-2324. [5] Krizhevsky A, Sutskever I, Hinton G E. Imagenet clas-sification with deep convolutional neural networks. Communications of the ACM, 2017, 60(6): 84-90. [6] WU Chenfang, YANG Shixi, HUANG Haizhou, et al. An improved fault diagnosis method of rolling bearings based on Le Net-5. Journal of Vibration and Shock, 2021, 40(12): 55-61. [7] ZHANG Long, HU Yanqing, ZHAO Lijuan, et al. Fault Diagnosis of Rolling Bearings Using Recurrence Plot Coding Technique and Residual Network. Journal of Xi'an Jiaotong University, 2023, 57(02): 110-120. [8] XUE Jiantong, MA Hongzhong, YANG Hongsu, et al. A fault diagnosis method for transformer winding looseness based on Gramian angular field and transfer learning-AlexNet. Power System Protection and Control, 2023, 51(24): 154-163. [9] GU Yingkui, WU Kuan, LI Cheng, et al. Rolling bearing fault diagnosis based on Gram angle field and transfer deep residual neural network. Journal of Vibration and Shock, 2022, 41(21): 228-237. [10] GAO Shu-guo, MENG Ling-ming, ZHANG Yu-kun, et al. Fault diagnosis method for core loose of high voltage shunt reactor using vibration sensing array. Journal of Vibration Engineering, 2023, 36(03): 875-884. [11] SHAO Haidong, XIAO Yiming, YAN Shen, et al. Simulation Data-driven Enhanced Unsupervised Domain Adaptation for Bearing Fault Diagnosis. Journal of Mechanical Engineering, 2023, 59(03): 76-85. [12] CUI Guiyan, ZHONG Qianwen, ZHENG Shubin, et al. Multi-sensor fusion bearing fault diagnosis based on VMD gray level image coding and CNN. Journal of Vibration and Shock, 2023, 42(21): 316-326.