基于图正则化堆叠自编码器的风机轴承故障诊断方法

doi:10.12096/j.2096-4528.pgt.23175

发电技术 ›› 2024, Vol. 45 ›› Issue (6): 1146-1152.DOI: 10.12096/j.2096-4528.pgt.23175

• 新能源 • 上一篇

基于图正则化堆叠自编码器的风机轴承故障诊断方法

刘展¹, 刘健洵², 包琰洋², 李大字²

^1.北京能高普康测控技术有限公司，北京市丰台区 100070
^2.北京化工大学信息科学;与技术学院，北京市朝阳区 100029

收稿日期:2023-12-15 修回日期:2024-03-02 出版日期:2024-12-31 发布日期:2024-12-30
通讯作者: 李大字
作者简介:刘展(1982)，男，博士，高级工程师，主要从事风力发电行业基于工业物联网的核心关键设备服役能力保持与主动安全保障技术研究、产品开发及工业物联网大数据分析服务等相关工作，LiuZhan@bj-pukang.com；
李大字(1970)，女，博士，教授，主要从事机器学习与人工智能、故障诊断、先进控制、复杂系统建模与优化等研究，本文通信作者，lidz@mail.buct.edu.cn。
基金资助:
国家自然科学基金项目(62273026);工信部高技术船舶科研项目(MC-202025-S02)

Bearing Faults Diagnosis Method Based on Stacked Auto-Encoder With Graph Regularization for Wind Turbines

Zhan LIU¹, Jianxun LIU², Yanyang BAO², Dazi LI²

^1.Beijing Pukang Measurement and Monitoring Tech Co. , Ltd. , Fengtai District, Beijing 100070, China
^2.College of Information Science and Technology, Beijing University of Chemical Technology, Chaoyang District, Beijing 100029, China

Received:2023-12-15 Revised:2024-03-02 Published:2024-12-31 Online:2024-12-30
Contact: Dazi LI
Supported by:
National Natural Science Foundation of China(62273026);High-Tech Ship Research Project of Ministry of Industry and Information Technology(MC-202025-S02)

摘要/Abstract

摘要：

目的为解决风电机组轴承故障诊断中的故障特征提取效率低、特征表示不够精准以及现有方法难以适应复杂信号需求的问题，提出一种基于图正则化的故障诊断方法，提升了对振动信号的分析能力，从而实现对不同故障类型的准确分类和可靠诊断。方法采用基于图正则化自编码器的技术，结合图嵌入思想，指导堆叠监督自编码器进行特征提取。在故障特征提取阶段，首先对诊断信号进行图表示，然后在堆叠自编码器中添加图正则化项，以确保嵌入后的低维特征保持流形结构，从而提取数据深层的复杂几何特征。结果所提取的特征能够实现对不同故障类型的准确分类，展现出在故障特征捕捉方面的显著优势。实验结果表明，该方法在实际风场数据的应用中，具有更高的诊断精度和可靠性，有效提高了故障特征的提取效率和分类准确性。结论所提方法在风电轴承故障诊断领域表现出显著的有效性和优越性，为实际应用提供了可靠的技术支持。

关键词: 风电机组, 轴承, 故障诊断, 图正则化, 自编码器, 图嵌入, 特征提取, 振动信号

Abstract:

Objectives In order to solve the problems of low efficiency of fault feature extraction, inaccurate feature representation, and difficulty in adapting existing methods to complex signal requirements in wind turbine bearing fault diagnosis, a fault diagnosis method based on graph regularization was proposed. The method helps to improve the analysis ability of vibration signals, thereby achieving accurate classification and reliable diagnosis of different fault types. Methods The technology based on graph regularization auto-encoder was adopted, combined with the idea of graph embedding, to guide the stack-supervised auto-encoder to carry out feature extraction. In the fault feature extraction stage, the diagnostic signal was first graphically represented, and then graph regularization terms were added to the stacked auto-encoder to ensure that the embedded low-dimensional features maintain the manifold structure, thereby extracting complex geometric features deep in the data. Results The extracted features can accurately classify different fault types, showing significant advantages in fault feature capture. Experimental results show that this method has higher diagnosis accuracy and reliability in the application of actual wind farm data, and effectively improves the extraction efficiency and classification accuracy of fault features. Conclusions The proposed method has shown significant effectiveness and advantages in the field of wind power bearing fault diagnosis, and provides reliable technical support for practical applications.

Key words: wind turbine, bearing, fault diagnosis, graph regularization, auto-encoder, graph embedding, feature extraction, vibration signal

中图分类号:

TK 83

刘展, 刘健洵, 包琰洋, 李大字. 基于图正则化堆叠自编码器的风机轴承故障诊断方法[J]. 发电技术, 2024, 45(6): 1146-1152.

Zhan LIU, Jianxun LIU, Yanyang BAO, Dazi LI. Bearing Faults Diagnosis Method Based on Stacked Auto-Encoder With Graph Regularization for Wind Turbines[J]. Power Generation Technology, 2024, 45(6): 1146-1152.

图/表 8

图1 自编码器网络结构

Fig. 1 Autoencoder network architecture

图2 故障特征图表征

Fig. 2 Graphical representations of fault characteristics

表1 不同小波函数的性能比较

Tab. 1 Performance comparison of different wavelet functions

小波函数	MSE	SNR	PSNR
Coiflet	15.463 2	11.944 8	12.513 6
Meyer	16.485 9	11.666 7	12.235 4
Daubechies	17.040 3	11.523 0	12.091 7
Symlet	17.018 2	11.528 7	12.097 4

表2 特征参数

Tab. 2 Characteristic parameters

时域特征	频域特征
$p 1 = ∑ n = 1 N x (n) N$	$p 12 = ∑ k = 1 K s (k) K$
$P 2 = ∑ n = 1 N [x (n) - p 1] 2 N - 1$	$p 13 = ∑ k = 1 K [s (k) - p 12] 2 K - 1$
$p 3 = (∑ n = 1 N x (n) N) 2$	$p 14 = ∑ k = 1 K [s (k) - p 12] 3 K (p 13) 3$
$p 4 = ∑ n = 1 N [x (n)] 2 N$	$p 15 = ∑ k = 1 K [s (k) - p 12] 4 K p 132$
$p 5 = m a x x (n)$	$p 16 = ∑ k = 1 K f k s (k) ∑ k = 1 K s (k)$
$p 6 = ∑ n = 1 N [x (n) - p 1] 3 (N - 1) p 23$	$p 17 = ∑ k = 1 K (f k - P 16) 2 s (k) K$
$p 7 = ∑ n = 1 N [x (n) - p 1] 4 (N - 1) p 24$	$p 18 = ∑ k = 1 K f k 2 s (k) ∑ k = 1 K s s (k)$
$p 8 = p 5 p 4$	$p 19 = ∑ k = 1 K f k 4 s (k) ∑ k = 1 K f k 2 s (k)$
$p 9 = p 5 p 3$	$p 20 = ∑ k = 1 K f k 2 s (k) ∑ k = 1 K s (k) ∑ k = 1 K f k 4 (k)$
$p 10 = p 4 1 N ∑ n = 1 N x (n)$	$p 21 = p 17 p 16$
$p 11 = p 5 1 N ∑ n = 1 N x (n)$	$p 22 = ∑ k = 1 K (f k - p 16) 3 s (k) K p 173$
	$p 23 = ∑ k = 1 K (f k - p 16) 4 s (k) K p 174$
	$p 24 = ∑ k = 1 K (f k - p 16) 1 / 2 s (k) K p 17$

表2 特征参数

Tab. 2 Characteristic parameters

时域特征	频域特征
$p 1 = ∑ n = 1 N x (n) N$	$p 12 = ∑ k = 1 K s (k) K$
$P 2 = ∑ n = 1 N [x (n) - p 1] 2 N - 1$	$p 13 = ∑ k = 1 K [s (k) - p 12] 2 K - 1$
$p 3 = (∑ n = 1 N x (n) N) 2$	$p 14 = ∑ k = 1 K [s (k) - p 12] 3 K (p 13) 3$
$p 4 = ∑ n = 1 N [x (n)] 2 N$	$p 15 = ∑ k = 1 K [s (k) - p 12] 4 K p 132$
$p 5 = m a x x (n)$	$p 16 = ∑ k = 1 K f k s (k) ∑ k = 1 K s (k)$
$p 6 = ∑ n = 1 N [x (n) - p 1] 3 (N - 1) p 23$	$p 17 = ∑ k = 1 K (f k - P 16) 2 s (k) K$
$p 7 = ∑ n = 1 N [x (n) - p 1] 4 (N - 1) p 24$	$p 18 = ∑ k = 1 K f k 2 s (k) ∑ k = 1 K s s (k)$
$p 8 = p 5 p 4$	$p 19 = ∑ k = 1 K f k 4 s (k) ∑ k = 1 K f k 2 s (k)$
$p 9 = p 5 p 3$	$p 20 = ∑ k = 1 K f k 2 s (k) ∑ k = 1 K s (k) ∑ k = 1 K f k 4 (k)$
$p 10 = p 4 1 N ∑ n = 1 N x (n)$	$p 21 = p 17 p 16$
$p 11 = p 5 1 N ∑ n = 1 N x (n)$	$p 22 = ∑ k = 1 K (f k - p 16) 3 s (k) K p 173$
	$p 23 = ∑ k = 1 K (f k - p 16) 4 s (k) K p 174$
	$p 24 = ∑ k = 1 K (f k - p 16) 1 / 2 s (k) K p 17$

图3 齿轮箱轴承数据预处理前后的时域信号对比

Fig. 3 Comparison of time-domain signals before and after preprocessing of gearbox bearing data

图4 故障特征可视化结果

Fig. 4 Visualization results of fault characteristics

表3 采用不同特征提取方法后的轴承磨损故障类型分类准确率

Tab. 3 Cassification accuracy of bearing wear fault types using different feature extraction methods

测点位置	图正则化自编码器	传统时频域特征提取
主轴轴承	0.976	0.908
齿轮箱轴承	0.933	0.879
发电机轴承	0.949	0.913

表4 采用不同特征提取方法后的内外环破损故障类型分类准确率

Tab. 4 Classification accuracy of inner and outer ring damage fault types using different feature extraction methods

测点位置	图正则化自编码器	传统时频域特征提取
主轴轴承	0.953	0.900
齿轮箱轴承	0.978	0.853
发电机轴承	0.992	0.950

参考文献 22

1	BADAL F R，DAS P， SARKER S K，et al ．A survey on control issues in renewable energy integration and microgrid[J]．Protection and Control of Modern Power Systems，2019，4(1)：8-14． doi:10.1186/s41601-019-0122-8
2	姜红丽，刘羽茜，冯一铭，等．碳达峰、碳中和背景下“十四五”时期发电技术趋势分析[J]．发电技术，2022，43(1)：54-64． doi:10.12096/j.2096-4528.pgt.21030
	JIANG H L， LIU Y X， FENG Y M，et al ．Analysis of power generation technology trend in 14th Five-Year Plan under the background of carbon peak and carbon neutrality[J]．Power Generation Technology，2022，43(1)：54-64． doi:10.12096/j.2096-4528.pgt.21030
3	SPINATO F， TAVNER P J， VAN BUSSEL G J W，et al ．Reliability of wind turbine subassemblies[J]．IET Renewable Power Generation，2009，3(4)：387-395． doi:10.1049/iet-rpg.2008.0060
4	李俊卿，马亚鹏，胡晓东，等．基于CBAM-InceptionV2-双流CNN的风电机组轴承故障诊断[J]．智慧电力，2023，51(6)：28-33．
	LI J Q， MA Y P， HU X D，et al ．Wind turbine bearing fault diagnosis based on CBAM-InceptionV2-two-stream CNN[J]．Smart Power，2023，51(6)：28-33．
5	田德，陶立壮．计及齿轮分度圆误差的风电齿轮箱振动特性分析[J]．可再生能源，2023，41(4)：480-486．
	TIAN D， TAO L Z ．Vibration characteristics research of wind turbine gearbox considering flank pitch error[J]．Renewable Energy Resources，2023，41(4)：480-486．
6	刘展，包琰洋，李大字．基于重采样降噪与主成分分析的宽卷积深度神经网络风机故障诊断方法[J]．发电技术，2023，44(6)：824-832．
	LIU Z， BAO Y Y， LI D Z ．Fault diagnosis method of wind turbines based on wide deep convolutional neural network with resampling and principal component analysis[J]．Power Generation Technology，2023，44(6)：824-832．
7	陈阳，刘永前，韩爽，等．基于多模态时频图融合的风电机组齿轮箱故障诊断方法[J]．分布式能源，2023，8(3)：17-23．
	CHEN Y， LIU Y Q， HAN S，et al ．Fault diagnosis method of wind turbine gearbox based on multi-mode time-frequency image fusion[J]．Distributed Energy，2023，8(3)：17-23．
8	LIU Z， ZHANG L ．A review of failure modes，condition monitoring and fault diagnosis methods for large-scale wind turbine bearings[J]．Measurement，2020，149：107002． doi:10.1016/j.measurement.2019.107002
9	LIU B， LIU B X， DAI Q B ．A review of bearing fault diagnosis for wind turbines[J]．IOP Conference Series：Earth and Environmental Science，2021，675(1)：012094． doi:10.1088/1755-1315/675/1/012094
10	金晓航，孙毅，单继宏，等．风力发电机组故障诊断与预测技术研究综述[J]．仪器仪表学报，2017，38(5)：1041-1053．
	JIN X H， SUN Y， SHAN J H，et al ．Fault diagnosis and prognosis for wind turbines：an overview[J]. Chinese Journal of Scientific Instrument，2017，38(5)：1041-1053．
11	JIN T， YAN C， CHEN C，et al ．Light neural network with fewer parameters based on CNN for fault diagnosis of rotating machinery[J]．Measurement，2021，181：109639． doi:10.1016/j.measurement.2021.109639
12	WANG X， QIN Y， WANG Y，et al ．ReLTanh：an activation function with vanishing gradient resistance for SAE-based DNNs and its application to rotating machinery fault diagnosis[J]．Neurocomputing，2019，363：88-98． doi:10.1016/j.neucom.2019.07.017
13	PANG Y， JIA L， ZHANG X，et al ．Design and implementation of automatic fault diagnosis system for wind turbine[J]．Computers & Electrical Engineering，2020，87：106754． doi:10.1016/j.compeleceng.2020.106754
14	LIU J Q， PAN C L， LEI F，et al ．Fault prediction of bearings based on LSTM and statistical process analysis[J]．Reliability Engineering & System Safety，2021，214：107646． doi:10.1016/j.ress.2021.107646
15	SHAH A K， YADAV A， MALIK H ．EMD and ANN based intelligent model for bearing fault diagnosis[J]．Journal of Intelligent & Fuzzy Systems，2018，35(5)：5391-5402． doi:10.3233/jifs-169821
16	ZHAO Y， ZHOU M， WANG L，et al ．Bearing fault diagnosis with cascaded space projection and a CNN[J]．Control Theory and Technology，2022，20(1)：103-113． doi:10.1007/s11768-022-00084-0
17	WANG M， WANG W， ZHANG X，et al ．A new fault diagnosis of rolling bearing based on Markov transition field and CNN[J]．Entropy，2022，24(6)：751． doi:10.3390/e24060751
18	丁雪，邓艾东，李晶，等．基于多尺度和注意力机制的滚动轴承故障诊断[J]．东南大学学报(自然科学版)，2022，52(1)：172-178．
	DING X， DENG A D， LI J，et al ．Fault diagnosis of rolling bearing based on multi-scale and attention mechanism[J]．Journal of Southeast University (Natural Science Edition)，2022，52(1)：172-178．
19	KE Y， YAO C， SONG E，et al ．Intelligent fault diagnosis method of common rail injector based on composite hierarchical dispersion entropy and improved least squares support vector machine[J]．Digital Signal Processing，2021，114：103054． doi:10.1016/j.dsp.2021.103054
20	POTHISARN C， KLOMJIT J， NGAOPITAKKUL A ．Comparison of various mother wavelets for fault classification in electrical systems[J]．Applied Sciences，2020，10(4)：1203-1210． doi:10.3390/app10041203
21	LEI Y， HE Z， ZI Y，et al ．Fault diagnosis of rotating machinery based on multiple ANFIS combination with GAs[J]．Mechanical Systems and Signal Processing，2007，21(5)：2280-2294． doi:10.1016/j.ymssp.2006.11.003
22	李树钰．改进的小波阈值去噪方法及其在MATLAB中的仿真[J]．噪声与振动控制，2010，30(2)：121-124．
	LI S Y ．Improved wavelet threshold denoising method and its simulation using MATALB[J]．Noise and Vibration Control，2010，30(2)：121-124．

[1]	王耀函, 张扬帆, 赵庆旭, 王晓东, 梁恺, 王玙. 低电压穿越过程中风电机组载荷特性联合仿真研究[J]. 发电技术, 2024, 45(4): 705-715.
[2]	樊昂, 李录平, 刘瑞, 欧阳敏南, 陈尚年. 不同风速对单桩式海上风电机组塔筒动态特性的影响[J]. 发电技术, 2024, 45(2): 312-322.
[3]	陈代俊, 陈里里, 李阳涛. 联合循环发电站电力输出预测[J]. 发电技术, 2024, 45(1): 99-105.
[4]	孙财新, 张波, 唐巍, 周昳鸣, 付明志, 秦猛, 郭小江. 海上风电机组国产化研究与实践[J]. 发电技术, 2023, 44(5): 696-702.
[5]	董光德, 李道明, 陈咏涛, 马兴, 付昂, 穆钢, 肖白. 基于粒子群优化与卷积神经网络的电能质量扰动分类方法[J]. 发电技术, 2023, 44(1): 136-142.
[6]	张航, 周传杰, 张林, 陈节涛, 徐春梅, 彭道刚. 基于概率神经网络-小波神经网络-DS信息融合的电厂引风机故障诊断[J]. 发电技术, 2022, 43(6): 951-958.
[7]	白明亮, 张冬雪, 刘金福, 刘娇, 于达仁. 基于深度自编码器和支持向量数据描述的燃气轮机高温部件异常检测[J]. 发电技术, 2021, 42(4): 422-430.
[8]	崔颖, 张晗, 黄宇熙. 考虑接触刚度的周向拉杆转子-轴承系统动力学特性分析[J]. 发电技术, 2021, 42(4): 447-453.
[9]	陈尚年, 李录平, 张世海, 欧阳敏南, 樊昂, 文贤馗. 汽轮发电机组振动故障诊断技术研究进展[J]. 发电技术, 2021, 42(4): 489-499.
[10]	李亚光, 李蒙. 基于深度小世界神经网络的风电机组异常检测[J]. 发电技术, 2021, 42(3): 313-321.
[11]	杨昌杏, 李录平. 风力机叶片损伤故障检测技术研究进展[J]. 发电技术, 2020, 41(6): 599-607.
[12]	赵海川, 芦彦东, 郑浩康, 邢作霞. 低压微网中小型直驱永磁风电机组低电压穿越技术研究[J]. 发电技术, 2020, 41(6): 659-666.
[13]	于永军,王利超,张明远,肖仕武,张馨元. 基于阻抗特性多项式拟合的直驱风电机组次同步振荡稳定判据[J]. 发电技术, 2020, 41(4): 429-436.
[14]	危威. 大型火电厂风机质量不平衡故障分析及处理[J]. 发电技术, 2020, 41(2): 198-201.
[15]	胡伟,戚宇辰,张鸿轩,董凌,李延和. 风光水多能源电力系统互补智能优化运行策略[J]. 发电技术, 2020, 41(1): 9-18.

基于图正则化堆叠自编码器的风机轴承故障诊断方法

Bearing Faults Diagnosis Method Based on Stacked Auto-Encoder With Graph Regularization for Wind Turbines

RichHTML

PDF (PC)

可视化

摘要/Abstract

引用本文

图/表 8

参考文献 22

相关文章 15

编辑推荐

Metrics