基于卷积神经网络的核电多相无刷励磁系统旋转整流器故障诊断

doi:10.19595/j.cnki.1000-6753.tces.221531

摘要
图/表
参考文献
相关文章 (15)

全文: PDF (1122 KB)
输出: BibTeX | EndNote (RIS)

摘要

多相无刷励磁系统广泛应用于大容量核电机组,对其旋转整流器故障的准确判断是保障机组安全稳定运行的重要前提,但是现有方法难以准确区分所有类型故障。为此,论文提出基于卷积神经网络的多相无刷励磁系统旋转整流器故障诊断方案。在给出故障诊断信号选取的理论基础上,将励磁电流作为输入特征,采用一维空洞卷积神经网络实现励磁电流的特征提取和故障模式分类,并结合置信度指标对诊断结果的可靠程度进行评价。进一步使用Score-CAM类激活映射算法分析模型的诊断机制,兼顾了准确性和可解释性。通过11相无刷励磁系统的动模实验进行验证,结果表明所提方法能够准确区分旋转整流器的不同二极管开路故障模式,在含噪声情况下仍然有较高的准确率,具有良好的应用前景。

	服务

	把本文推荐给朋友
	加入我的书架
	加入引用管理器
	E-mail Alert
	RSS
	作者相关文章
	梁郑秋
	郝亮亮
	周艳真
	段贤稳
	王光

关键词 ：无刷励磁系统, 旋转整流器, 深度学习, 卷积神经网络, 故障诊断

Abstract：

Multi-phase brushless excitation system is widely used in large-capacity nuclear power plants. The safe and stable operation of nuclear power plants depends on the accurate diagnosis of rotating rectifier fault. However, the existing methods cannot accurately distinguish all types of faults, which makes it difficult to meet the demands in practice. Recently, as a representative artificial intelligence method, deep learning provided new ideas for solving this problem. Therefore, this paper develops a fault diagnosis scheme for the rotating rectifier based on convolutional neural network.
First, taking an 11-phase brushless excitation system as an example, the excitation current characteristics after the rotating rectifier fault is analyzed. The results show that the excitation current waveforms contain amplitude and phase information of all fault harmonic, so that they can be various under different fault modes and used for fault diagnosis of the rotating rectifier.
Then, a one-dimensional dilated convolutional neural network (1D-DCNN) model that can realize the fault diagnosis is introduced. The 1D-DCNN establishes a mapping relationship from the input (excitation current waveforms) to the output (fault mode). Meanwhile, the difference between the largest probability and the second largest probability of the output is defined as the confidence based on the conditional probability of samples belonging to different fault modes given by the model, which is used to measure the reliability of the diagnosis results. The process of fault diagnosis based on 1D-DCNN model includes three stages. At stage 1, we acquire raw data at several voltage levels to ensure that the model has applicability and generalization ability to different voltage levels. Then, we need to preprocess the data in terms of sliding window segmentation, filtering and mean normalization. At stage 2, the 1D-DCNN model is trained and used for online diagnosis of rotating rectifier faults. At stage 3, the Score-CAM method is introduced to clarify the interpretability of the 1D-DCNN model.
Finally, the effectiveness of the proposed 1D-DCNN model is verified by the prototype experiment. We mainly study diode open-circuit faults of 11-phase brushless excitation system which include eight fault types. The processed samples are divided into training set, verification set and test set. The verification set is utilized to determine the hyperparameter of the model. We use the control variable method to optimize the hyperparameters of each layer in turn. Next, the Score-CAM method is used to provide interpretability for the 1D-DCNN model and explain how hyperparameters influence fault diagnosis performance. Besides, the test sets with noise are added as a comparison. We compare the diagnosis accuracy of different models in the original test set and the test sets with noise, and the 1D-DCNN model shows better anti-noise ability. In addition, we analyze the confidence of model diagnosis results in the original test set and the test sets with noise. The results show that selecting the appropriate confidence threshold contribute to improve the classification accuracy of the model.
In conclusion, the feasibility of fault diagnosis of rotating rectifiers based on convolutional neural network is verified, which provides a new data-driven idea for fault diagnosis of rotating rectifier.

Key words： Brushless excitation system rotating rectifier deep learning convolutional neural network fault diagnosis

PACS:

TM341

基金资助:

中央高校基本科研业务费专项资金资助（No.2020JBM070）; 中广核集团公司科技项目资助项目(3100077013)

通讯作者: 周艳真女,1990年生,博士,研究方向为电力系统分析与控制、人工智能在复杂电网调控及故障诊断中的应用。E-mail: zhouyzh@126.com

作者简介: 梁郑秋女,1999年生,硕士研究生,研究方向为多相无刷励磁系统故障监测。E-mail: 21121435@bjtu.edu.cn

引用本文:

梁郑秋, 郝亮亮, 周艳真, 段贤稳, 王光. 基于卷积神经网络的核电多相无刷励磁系统旋转整流器故障诊断[J]. 电工技术学报, 0, (): 20235406-20235406. Liang Zhengqiu, Hao Liangliang, Zhou Yanzhen, Duan Xianwen, Wang Guang. Fault Diagnosis of Rotating Rectifier in Nuclear Multi-phase Brushless Excitation System Based on Convolutional Neural Network. Transactions of China Electrotechnical Society, 0, (): 20235406-20235406.

链接本文:

https://dgjsxb.ces-transaction.com/CN/10.19595/j.cnki.1000-6753.tces.221531 https://dgjsxb.ces-transaction.com/CN/Y0/V/I/20235406

[1] 张蕴. 双碳目标下我国核电发展趋势分析[J]. 核科学与工程, 2021, 41(6): 1347-1351.
Zhang Yun.Analysis on the development trend of nuclear energy in China under the dual carbon target[J]. Nuclear Science and Engineering, 2021, 41(6): 1347-1351.
[2] 郝亮亮, 常江, 李洪学, 等. 核电多相角形无刷励磁机定子匝间短路故障特征分析[J]. 中国电机工程学报, 2019, 39(22): 6726-6736.
Hao Liangliang, Chang Jiang, Li Hongxue, et al.Characteristics analysis of stator inter-turn short circuits in nuclear power multi-phase angular brushless exciter[J]. Proceedings of the CSEE, 2019, 39(22): 6726-6736.
[3] 郝亮亮, 李佳慧, 段贤稳, 等.核电多相环形无刷励磁机转子绕组短路故障特征分析[J]. 电工技术学报, 2020, 35(06): 1251-1261.
Hao Liangliang, Li Jiahui, Duan Xianwen, et al.Characteristic analysis of short-circuit fault in rotor winding of nuclear power multi-phase annular brushless exciter[J]. Transactions of China Electrotechnical Society, 2020, 35(06): 1251-1261.
[4] 吴新振, 张伟玲, 董海涛, 等. 十二相整流发电机二极管开路的故障运行分析[J]. 电工技术学报, 2017, 32(03): 157-163.
Wu Xinzhen, Zhang Weiling, Dong Haitao, et al.Analysis of diode-open fault operation for 12-phase rectifier generators[J]. Transactions of China Electrotechnical Society, 2017, 32(03): 157-163.
[5] Bui H K, Bracikowski N, Hecquet M, et al.Simulation of a large power brushless synchronous generator (BLSG) with a rotating rectiﬁer by a reluctance network for fault analysis and diagnosis[J]. IEEE Transactions on Industry Applications, 2017, 53(5): 4327-4337.
[6] Wu Y, Cai B, Ma Q.An online diagnostic method for rotary diode open-circuit faults in brushless exciters[J]. IEEE Transactions on Energy Conversion, 2018, 33(4): 1677-1685.
[7] 张伟玲, 吴新振, 魏锟, 等. 十二相整流发电机缺桥故障运行分析[J]. 中国电机工程学报, 2016, 36(06): 1724-1730.
Zhang Weiling, Wu Xinzhen, Wei Kun, et al.Analysis of bridge-loss fault operation of 12-phase generators with rectifier-load[J]. Proceedings of the CSEE, 2016, 36(06): 1724-1730.
[8] 蔡波冲, 武玉才, 赵艳军. 基于定子电流谐波法的无刷励磁机旋转二极管开路故障检测[J]. 大电机技术, 2018(04): 61-65.
Cai Bochong, Wu Yucai, Zhao Yanjun.The detection of open-circuit fault of rotary diode in brushless exciter using stator current harmonic method[J]. Large Electric Machine and Hydraulic Turbine, 2018(04): 61-65.
[9] 郝亮亮, 李佳慧, 李洪学, 等. 核电多相角形无刷励磁系统旋转二极管开路故障特征分析[J]. 电力系统自动化, 2019, 43(11): 112-120.
Hao Liangliang, Li Jiahui, Li Hongxue, et al.Characteristic analysis of open-circuit fault of rotating diode in nuclear multi-phase angular brushless excitation system[J]. Automation of Electric Power Systems, 2019, 43(11): 112-120.
[10] 朱继忠, 骆腾燕, 吴皖莉, 等. 综合能源系统运行可靠性评估评述Ⅱ:数据驱动法与模型-数据混合驱动法[J]. 电工技术学报, 2022, 37(13): 3227-3240.
Zhu Jizhong, Luo Tengyan, Wu Wanli, et al.A Review of Operational Reliability Assessment of Integrated Energy Systems Ⅱ : Data-Driven Method and Model-Data Hybrid Driven Method[J]. Transactions of China Electrotechnical Society, 2022, 37(13): 3227-3240.
[11] 马伟明. 关于电工学科前沿技术发展的若干思考[J]. 电工技术学报, 2021, 36(22): 4627-4636.
Ma Weiming.Thoughts on the Development of Frontier Technology in Electrical Engineering[J]. Transactions of China Electrotechnical Society, 2021, 36(22): 4627-4636.
[12] 崔江, 冯赛, 张卓然, 等. 基于BLS的无刷发电机旋转整流器特征提取技术研究[J]. 中国电机工程学报, 2020, 40(12): 4004-4013.
Cui Jiang, Feng Sai, Zhang Zhuoran, et al.Research on feature extraction technology of brushless generator rotating rectifier based on BLS[J]. Proceedings of the CSEE, 2020, 40(12): 4004-4013.
[13] 崔江, 郭瑞东, 张卓然, 等. 基于改进DBN的发电机旋转整流器故障特征提取技术[J]. 中国电机工程学报, 2020, 40(07): 2369-2376+2415.
Cui Jiang, Guo Ruidong, Zhang Zhuoran, et al. Generator rotating rectifier fault feature extraction technique based on improved DBN[J]. Proceedings of the CSEE, 2020, 40(7): 2369-2376+2415.
[14] Rahnama M, Vahedi A, Alikhani A M, et al.Diode open-circuit fault detection in rectifier bridge of the brushless synchronous generator[C]//2018 XIII International Conference on Electrical Machines (ICEM), Alexandroupoli, Greece, 2018: 1821-1826.
[15] Alikhani A M, Rahnama M, Vahedi A, et al.Neighbors class colidarity feature selection for fault diagnosis of brushless generator using thermal imaging[J]. IEEE Transactions on Instrumentation and Measurement, 2020, 69(9): 6221-6227.
[16] 纪守领, 李进锋, 杜天宇, 等. 机器学习模型可解释性方法、应用与安全研究综述[J]. 计算机研究与发展, 2019, 56(10): 2071-2096.
Ji Shouling, Li Jinfeng, Du Tianyu, et al.Survey on techniques, applications and security of machine learning interpretability[J]. Journal of Computer Research and Development, 2019, 56(10): 2071-2096.
[17] 任守纲, 张景旭, 顾兴健, 等. 时间序列特征提取方法研究综述[J]. 小型微型计算机系统, 2021, 42(02): 271-278.
Ren Shougang, Zhang Jingxu, Gu Xingjian, et al.Overview of feature extraction algorithms for time series[J]. Journal of Chinese Computer Systems, 2021, 42(02): 271-278.
[18] Zhang Q, Wu Y N, Zhu S C.Interpretable convolutional neural networks[C]//2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Salt Lake City, USA, 2018: 8827-8836.
[19] Selvaraju R R, Cogswell M, Das A, et al.Grad-CAM: visual explanations from deep networks via gradient-based localization[J]. International Journal of Computer Vision, 2020, 128(2):336-359.
[20] Chattopadhyay A, Sarkar A, Howlader P, et al.Grad-CAM++: improved visual explanations for deep convolutional networks[C]//2018 IEEE Winter Conference on Applications of Computer Vision. Nevada, USA, 2018.
[21] Wang H, Wang Z, Du M, et al.Score-CAM: score-weighted visual explanations for convolutional neural networks[C]//2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Seattle, USA, 2020.
[22] Zeiler M D, Fergus R.Visualizing and Understanding Convolutional Networks[C]//Proceedings of the 13th European Conference on Computer Vision. Zurich, Switzerland, 2014:818-83.
[23] FongG R, Vedaldi A. Interpretable explanations of black boxes by meaningful perturbation[C]//2017 IEEE International Conference on Computer Vision. Venice, Italy, 2017:3449-3457.
[24] Chen J, Hao L, Li H, et al.Time-frequency characteristics analysis and diagnosis of rotating rectifier faults in multi-phase annular brushless system[J]. IEEE Transactions on Industrial Electronics, 2023, 70(4):3233-3244.
[25] 张大海,张晓炜,孙浩等.基于卷积神经网络的交直流输电系统故障诊断[J].电力系统自动化,2022,46(05):132-145.
Zhang Dahai, Zhang Xiaowei, Sun Hao, et al.Fault Diagnosis for AC/DC Transmission System Based on Convolutional Neural Network[J]. Automation of Electric Power Systems, 2022,46(05):132-145.
[26] 王艳新,闫静,王建华等.基于域对抗迁移卷积神经网络的小样本GIS绝缘缺陷智能诊断方法[J].电工技术学报,2022,37(09):2150-2160.
Wang Yanxin, Yan Jing, Wang Jianhua, et al.Intelligent Diagnosis for GIS with Small Samples Using a Novel Adversarial Transfer Learning in Convolutional Neural Network[J]. Transactions of China Electrotechnical Society, 2022,37(09):2150-2160.
[27] 蔡智超,孙翼虎,赵振勇等.基于时频分析和深度学习的表面粗糙度超声模式识别方法[J].电工技术学报,2022,37(15):3743-3752.
Cai Zhichao, Sun Yihu, Zhao Zhenyong, et al.A Deep Learning-Based Electromagnetic Ultrasonic Recognition Method for Surface Roughness of Workpeice[J]. Transactions of China Electrotechnical Society, 2022,37(15):3743-3752.
[28] Yu F, Koltun V .Multi-scale context aggregation by dilated convolutions[C]//4th International Conference on Learning Representations,San Juan,Puerto Rico, 2016.
[29] Bridle J S.Probabilistic interpretation of feedforward classification network outputs, with relationships to statistical pattern recognition[J]. Neurocomputing: Algorithms, Architectures and Applications, 1989, 68:227-236.
[30] Rubinstein R.The cross-entropy method for combinatorial and continuous optimization[J]. Methodology & Computing in Applied Probability, 1999, 1(2):127-190.
[31] Maaten L, Hinton G.Visualizing data using t-SNE[J]. Journal of Machine Learning Research, 2008, 9(86): 2579-2605.