基于深度网络的可控混合式磁力耦合器退磁诊断

doi:10.3785/j.issn.1008-973X.2026.02.006

浙江大学学报(工学版)

2026, Vol. 60

Issue (2): 279-286 DOI: 10.3785/j.issn.1008-973X.2026.02.006

能源工程、机械工程

基于深度网络的可控混合式磁力耦合器退磁诊断

王爽1,2,3(

),章熙泰2,3,郭永存1,2,3,孙守锁2,3

1. 安徽理工大学煤炭无人化开采数智技术全国重点实验室，安徽淮南 232001
2. 安徽理工大学矿山智能技术与装备省部共建协同创新中心，安徽淮南 232001
3. 安徽理工大学机电工程学院，安徽淮南 232001

Demagnetization fault diagnosis of controllable hybrid magnetic couplers based on deep neural networks

Shuang WANG1,2,3(

),Xitai ZHANG2,3,Yongcun GUO1,2,3,Shousuo SUN2,3

1. State Key Laboratory of Digital Intelligent Technology for Unmanned Coal Mining, Anhui University of Science and Technology, Huainan 232001, China
2. Collaborative Innovation Center for Mining Intelligent Technology and Equipment, Anhui University of Science and Technology, Huainan 232001, China
3. School of Mechatronics Engineering, Anhui University of Science and Technology, Huainan 232001, China

全文: PDF(2447 KB) HTML

摘要：

可控混合式磁力耦合器的退磁程度直接影响耦合器的传动性能，准确诊断是保障稳定运行的关键. 为此，将格拉姆角和场（GASF）与改进残差网络（ResNet）结合提出退磁故障诊断方法. 通过有限元仿真提取耦合器气隙内不同位置的电流密度数据，利用GASF方法将一维时序数据转换为二维图像，以增强特征表达能力. 在此基础上，将注意力机制嵌入ResNet结构，提升网络对故障细节特征的学习与分类能力. 实验表明，所提方法在测试集上的平均诊断准确率为97.67%，较EfficientNet、RepVGG、ViT、AlexNet分别提升15.04、14.73、10.66和 9.37个百分点. 通过实验平台验证所提方法在可控混合式磁力耦合器退磁故障诊断中的有效性和优越性，该方法对真实退磁故障的识别准确率为96.5%.

关键词： 磁力耦合器; 格拉姆角和场（GASF）; 残差网络（ResNet）; 注意力机制; 故障诊断

Abstract:

The demagnetization degree of a controllable hybrid magnetic coupler has a significant influence on its transmission performance, and accurate diagnosis is essential for stable operation. A demagnetization fault diagnosis method combining the Gramian angular summation field (GASF) and an improved residual network (ResNet) was proposed. Current density data at different positions in the air gap were obtained through finite element simulation. One-dimensional time-series data were converted into two-dimensional images using the GASF method to enhance feature representation. An attention mechanism was embedded into the ResNet structure to improve the network’s ability to learn fault features and perform classification. Experimental results show that the proposed method achieves an average diagnostic accuracy of 97.67%, which is higher than those of EfficientNet, RepVGG, ViT, and AlexNet by 15.04, 14.73, 10.66, and 9.37 percentage points, respectively. Furthermore, an experimental platform was established to verify the effectiveness of the proposed method in demagnetization fault diagnosis of controllable hybrid magnetic couplers, and an identification accuracy of 96.5% was achieved for actual demagnetization faults.

Key words: magnetic coupler Gramian angular summation field (GASF) residual network (ResNet) attention mechanism fault diagnosis

收稿日期: 2025-02-08 出版日期: 2026-02-03

CLC:

TH 133

基金资助: 安徽省高校杰出青年科研项目（2022AH020056）；国家自然科学基金资助项目（52274152）；安徽省自然科学优秀青年科研基金资助项目（2308085Y37）.

作者简介: 王爽（1991—），女，教授，博导，博士. 从事高效磁力传动研究. orcid.org/0000-0002-6452-778X. E-mail：shuangw094@126.com

	服务
	把本文推荐给朋友
	加入引用管理器
	E-mail Alert
	作者相关文章
	王爽
	章熙泰
	郭永存
	孙守锁

引用本文:

王爽,章熙泰,郭永存,孙守锁. 基于深度网络的可控混合式磁力耦合器退磁诊断[J]. 浙江大学学报(工学版), 2026, 60(2): 279-286.

Shuang WANG,Xitai ZHANG,Yongcun GUO,Shousuo SUN. Demagnetization fault diagnosis of controllable hybrid magnetic couplers based on deep neural networks. Journal of ZheJiang University (Engineering Science), 2026, 60(2): 279-286.

链接本文:

https://www.zjujournals.com/eng/CN/10.3785/j.issn.1008-973X.2026.02.006 或 https://www.zjujournals.com/eng/CN/Y2026/V60/I2/279

图 1 可控混合式磁力耦合器结构模型

图 2 可控混合式磁力耦合器主磁通等效磁路图

表 1 可控混合式磁力耦合器模型参数

图 3 烧结钕铁硼N35磁体在不同退磁程度下的磁性曲线

图 4 不同退磁程度下的永磁体磁密分布、磁场强度示意图

图 5 格拉姆角和场编码流程

表 2 电流密度数据集内容

图 6 不同残差模块的结构示意图

图 7 基于注意力机制的改进残差网络结构示意图

表 3 退磁故障诊断模型参数

表 4 空间残差模块参数

图 8 退磁故障诊断模型10次实验的准确率

表 5 退磁故障诊断模型故障诊断结果

表 6 退磁故障诊断模型的模块消融实验结果

图 9 退磁故障诊断模型的模块性能参数对比

图 10 可控混合式磁力耦合器实验台

表 7 永磁体剩磁强度实测数据

表 8 磁力耦合器电流信号特征分析

图 11 退磁故障诊断模型在真实数据上的分类结果

9	郑近德, 潘海洋, 程军圣, 等基于自适应经验傅里叶分解的机械故障诊断方法[J]. 机械工程学报, 2020, 56 (9): 125- 136 ZHENG Jinde, PAN Haiyang, CHENG Junsheng, et al Adaptive empirical Fourier decomposition based mechanical fault diagnosis method[J]. Journal of Mechanical Engineering, 2020, 56 (9): 125- 136
10	李兵, 韩睿, 何怡刚, 等改进随机森林算法在电机轴承故障诊断中的应用[J]. 中国电机工程学报, 2020, 40 (4): 1310- 1319 LI Bing, HAN Rui, HE Yigang, et al Applications of the improved random forest algorithm in fault diagnosis of motor bearings[J]. Proceedings of the CSEE, 2020, 40 (4): 1310- 1319
11	赵国栋. 基于多层感知机的永磁同步电机退磁故障诊断 [D]. 哈尔滨: 哈尔滨工业大学, 2023: 1–91. ZHAO Guodong. Demagnetization fault diagnosis of permanent magnet synchronous motor based on multi-layer perceptron [D]. Harbin: Harbin Institute of Technology, 2023: 1–91.
12	宋雪玮, 赵吉文, 董菲, 等基于PSO-LSSVM的永磁同步直线电机局部退磁故障识别[J]. 中国电机工程学报, 2019, 39 (8): 2426- 2435 SONG Xuewei, ZHAO Jiwen, DONG Fei, et al Local demagnetization fault recognition of permanent magnet synchronous linear motor based on PSO-LSSVM[J]. Proceedings of the CSEE, 2019, 39 (8): 2426- 2435
13	蔡锷, 李春明, 刘东民, 等基于局部均值分解和K近邻算法的滚动轴承故障诊断方法[J]. 现代电子技术, 2015, 38 (13): 50- 52 CAI E, LI Chunming, LIU Dongmin, et al Fault diagnosis method based on LMD and KNN algorithms for rolling bearing[J]. Modern Electronics Technique, 2015, 38 (13): 50- 52
14	张丹, 赵吉文, 董菲, 等基于概率神经网络算法的永磁同步直线电机局部退磁故障诊断研究[J]. 中国电机工程学报, 2019, 39 (1): 296- 306 ZHANG Dan, ZHAO Jiwen, DONG Fei, et al Partial demagnetization fault diagnosis research of permanent magnet synchronous motors based on the PNN algorithm[J]. Proceedings of the CSEE, 2019, 39 (1): 296- 306
1	寇宝泉, 张鲁, 邢丰, 等高性能永磁同步平面电机及其关键技术发展综述[J]. 中国电机工程学报, 2013, 33 (9): 79- 87 KOU Baoquan, ZHANG Lu, XING Feng, et al Development of the high-performance synchronous permanent magnet planar motor and its key technologies[J]. Proceedings of the CSEE, 2013, 33 (9): 79- 87
15	谢乐, 衡熙丹, 刘洋, 等基于线性判别分析和分步机器学习的变压器故障诊断[J]. 浙江大学学报: 工学版, 2020, 54 (11): 2266- 2272 XIE Le, HENG Xidan, LIU Yang, et al Transformer fault diagnosis based on linear discriminant analysis and step-by-step machine learning[J]. Journal of Zhejiang University: Engineering Science, 2020, 54 (11): 2266- 2272
16	SU Y, MENG L, KONG X, et al Small sample fault diagnosis method for wind turbine gearbox based on optimized generative adversarial networks[J]. Engineering Failure Analysis, 2022, 140: 106573
17	ALSUMAIDAEE Y A M, PAW J K S, YAW C T, et al Fault detection for medium voltage switchgear using a deep learning hybrid 1D-CNN-LSTM model[J]. IEEE Access, 2023, 11: 97574- 97589
18	韩康, 战洪飞, 余军合, 等基于空洞卷积和增强型多尺度特征自适应融合的滚动轴承故障诊断[J]. 浙江大学学报: 工学版, 2024, 58 (6): 1285- 1295 HAN Kang, ZHAN Hongfei, YU Junhe, et al Rolling bearing fault diagnosis based on dilated convolution and enhanced multi-scale feature adaptive fusion[J]. Journal of Zhejiang University: Engineering Science, 2024, 58 (6): 1285- 1295
19	邱建琪, 沈佳晨, 史涔溦, 等基于残差卷积网络的多传感器融合永磁同步电机故障诊断[J]. 电机与控制学报, 2024, 28 (7): 24- 33 QIU Jianqi, SHEN Jiachen, SHI Cenwei, et al Fault diagnosis of multi-sensor fusion permanent magnet synchronous motor based on residual convolutional neural network[J]. Electric Machines and Control, 2024, 28 (7): 24- 33
2	陈本永, 潘科荣, 杨涛, 等磁场同步跟随式电磁悬浮微驱动器的理论分析与建模[J]. 中国机械工程, 2008, 19 (21): 2637- 2642 CHEN Benyong, PAN Kerong, YANG Tao, et al Analysis and modeling of an electromagnetic levitation micro-actuator[J]. China Mechanical Engineering, 2008, 19 (21): 2637- 2642
3	MENG Z, ZHU Z, SUN Y 3-D analysis for the torque of permanent magnet coupler[J]. IEEE Transactions on Magnetics, 2015, 51 (4): 8002008
4	FLORIO F, SINHA G, SUNDARARAMAN R Designing high-accuracy permanent magnets for low-power magnetic resonance imaging[J]. IEEE Transactions on Magnetics, 2018, 54 (5): 5300209 doi: 10.1109/tmag.2018.2795592
5	何富君, 仲于海, 张瑞杰, 等永磁涡流耦合传动特性研究[J]. 机械工程学报, 2016, 52 (8): 23- 28 HE Fujun, ZHONG Yuhai, ZHANG Ruijie, et al Research on characteristics of permanent magnet eddy-current coupling drive[J]. Journal of Mechanical Engineering, 2016, 52 (8): 23- 28
6	HÖGBERG S, BENDIXEN F B, MIJATOVIC N, et al. Influence of demagnetization-temperature on magnetic performance of recycled Nd-Fe-B magnets [C]// Proceedings of the IEEE International Electric Machines and Drives Conference. Coeur d’Alene: IEEE, 2016: 1242–1246.
7	赵学智, 叶邦彦, 陈统坚基于小波—奇异值分解差分谱的弱故障特征提取方法[J]. 机械工程学报, 2012, 48 (7): 37- 48 ZHAO Xuezhi, YE Bangyan, CHEN Tongjian Extraction method of faint fault feature based on wavelet-SVD difference spectrum[J]. Journal of Mechanical Engineering, 2012, 48 (7): 37- 48
8	耿志强, 陈威, 马波, 等基于连续小波卷积神经网络的轴承智能故障诊断方法[J]. 浙江大学学报: 工学版, 2024, 58 (10): 2069- 2075 GENG Zhiqiang, CHEN Wei, MA Bo, et al Bearing intelligent fault diagnosis method based on continuous wavelet convolutional neural network[J]. Journal of Zhejiang University: Engineering Science, 2024, 58 (10): 2069- 2075
20	CALIN M D, HELEREA E. Temperature influence on magnetic characteristics of NdFeB permanent magnets [C]// Proceedings of the 7th International Symposium on Advanced Topics in Electrical Engineering. Bucharest: IEEE, 2011: 1–6.

[1]	周思瑶,夏楠,江佳鸿. 姿态引导的双分支换装行人重识别网络[J]. 浙江大学学报(工学版), 2026, 60(1): 71-80.
[2]	张学军,梁书滨,白万荣,张奉鹤,黄海燕,郭梅凤,陈卓. 基于异构图表征的源代码漏洞检测方法[J]. 浙江大学学报(工学版), 2025, 59(8): 1644-1652.
[3]	林宜山,左景,卢树华. 基于多头自注意力机制与MLP-Interactor的多模态情感分析[J]. 浙江大学学报(工学版), 2025, 59(8): 1653-1661.
[4]	翟亚红,陈雅玲,徐龙艳,龚玉. 改进YOLOv8s的轻量级无人机航拍小目标检测算法[J]. 浙江大学学报(工学版), 2025, 59(8): 1708-1717.
[5]	付家瑞,李兆飞,周豪,黄惟. 基于Convnextv2与纹理边缘引导的伪装目标检测[J]. 浙江大学学报(工学版), 2025, 59(8): 1718-1726.
[6]	杨荣泰,邵玉斌,杜庆治. 基于结构感知的少样本知识补全[J]. 浙江大学学报(工学版), 2025, 59(7): 1394-1402.
[7]	杨宇豪,郭永存,李德永,王爽. 基于视觉信息的煤矸识别分割定位方法[J]. 浙江大学学报(工学版), 2025, 59(7): 1421-1433.
[8]	王圣举,张赞. 基于加速扩散模型的缺失值插补算法[J]. 浙江大学学报(工学版), 2025, 59(7): 1471-1480.
[9]	黄爱颖,李晓辉,孙淑娴,朱逸群. 基于因果解耦的域自适应滚动轴承故障诊断[J]. 浙江大学学报(工学版), 2025, 59(7): 1523-1531.
[10]	蔡永青,韩成,权巍,陈兀迪. 基于注意力机制的视觉诱导晕动症评估模型[J]. 浙江大学学报(工学版), 2025, 59(6): 1110-1118.
[11]	鞠文博,董华军. 基于上下文信息融合与动态采样的主板缺陷检测方法[J]. 浙江大学学报(工学版), 2025, 59(6): 1159-1168.
[12]	周翔宇,刘毅志,赵肄江,廖祝华,张德城. 面向目的地预测的层次化空间嵌入BiGRU模型[J]. 浙江大学学报(工学版), 2025, 59(6): 1211-1218.
[13]	王爽,孙守锁,郭永存,胡泽永. 电磁混合式耦合器调隙装置多目标参数优化[J]. 浙江大学学报(工学版), 2025, 59(5): 1007-1017.
[14]	李宗民,徐畅,白云,鲜世洋,戎光彩. 面向点云理解的双邻域图卷积方法[J]. 浙江大学学报(工学版), 2025, 59(5): 879-889.
[15]	周昶清,侯耀春,武鹏,杨帅,吴大转. 自适应齿轮箱稀疏表示原子构建方法[J]. 浙江大学学报(工学版), 2025, 59(5): 1018-1030.

Viewed

Full text

Abstract

Cited

Shared

Discussed