Lightweight OLTC mechanical fault recognition method based on secondary&nbsp;decomposition time-frequency maps and SE-DSMC-BSA

PDF(4997 KB)

Journal of Vibration and Shock ›› 2025, Vol. 44 ›› Issue (11) : 268-279.

Lightweight OLTC mechanical fault recognition method based on secondary decomposition time-frequency maps and SE-DSMC-BSA

LI Siqi1, XIA Mao1, LU Sizhao*1, BI Guihong1, HUANG Yichao1, RUAN Yanjun2, LI Liangchuang2

Author information +

History +

Abstract

On-load Tap-Changer (OLTC) is the only movable component in an on-load voltage regulating transformer, and its frequent switching can lead to mechanical failures. To achieve online monitoring of the OLTC's mechanical condition, this paper proposes a lightweight fault identification method that combines secondary decomposition time-frequency maps, Depthwise Separable Multiscale Convolutions (DSMC), SE attention mechanism, and Broadcast Self-Attention Mechanism (BSA). First, an OLTC fault simulation experimental platform is established to collect vibration signals. Based on this, secondary decomposition and Hilbert transform are introduced to convert all the decomposed components into time-frequency maps. Then, SE-DSMC is used for multi-scale feature extraction of the time-frequency maps and for enhancing channel features. Finally, BSA is employed to extract global features to improve the accuracy of fault identification. Compared to existing methods, this approach offers advantages, particularly in small sample scenarios, such as faster recognition speed, higher accuracy, and lightweight characteristics.

Key words

On-load Tap-Changer / Fault Identification / Secondary Decomposition / Depthwise Separable Multiscale Convolutions / Lightweight Design.

Cite this article

EndNote

Ris (Procite)

Bibtex

Download Citations

LI Siqi1, XIA Mao1, LU Sizhao1, BI Guihong1, HUANG Yichao1, RUAN Yanjun2, LI Liangchuang2. Lightweight OLTC mechanical fault recognition method based on secondary decomposition time-frequency maps and SE-DSMC-BSA[J]. Journal of Vibration and Shock, 2025, 44(11): 268-279

References

[1] 马宏忠, 严岩. 基于混沌理论和GOA-K-means算法的有载分接开关状态特征分析计算方法[J].电工技术学报,2021,36(07):1399-1406.
MA Hongzhong, YAN Yan. A Method for Analyzing and Calculating the State Characteristics of Load Tap Changers Based on Chaos Theory and GOA-K-means Algorithm[J]. Transactions of China Electrotechnical Society, 2021, 36(07): 1399-1406.
[2] 赵彤, 李庆民, 陈平. OLTC振动信号特征提取的动力学分析方法[J]. 电工技术学报,2007,(01):41-46.
ZHAO Tong, LI Qingmin, Chen Ping. Dynamic Analysis Method for Feature Extraction of OLTC Vibration Signals[J]. Transactions of China Electrotechnical Society, 2007, (01): 41-46.
[3] 梁玄鸿, 王有元, 古洪瑞. 基于换档全过程振动强度的变压器CM型有载分接开关机械故障诊断方法[J]. 中国电机工程学报,2024,44(10):4145-4157.
LIANG Xuanhong, WANG Youyuan, GU Hongrui. Mechanical Fault Diagnosis Method for Transformer CM-Type Load Tap Changers Based on Vibration Intensity During the Whole Gear Shifting Process[J]. Proceedings of the CSEE, 2024, 44(10): 4145-4157.
[4] 张燕, 方瑞明. 基于油中溶解气体动态网络标志物模型的变压器缺陷预警与辨识[J]. 电工技术学报,2020,35(09):2032-2041.
ZHANG Yan, FANG Ruiming. Transformer Defect Early Warning and Identification Based on Dynamic Network Marker Model of Dissolved Gas in Oil[J]. Transactions of China Electrotechnical Society, 2020, 35(09): 2032-2041.
[5] 韩帅, 高飞, 王博闻,等. 基于Mel频谱滤波和CNN的有载分接开关可听声辨识方法[J]. 电网技术,2021,45(09):3609-3617.
HAN Shuai, GAO Fei, WANG Bowen, et al. Audible Sound Identifi-cation Method for Load Tap Changers Based on Mel Spectrum Filtering and CNN[J]. Power System Technology, 2021, 45(09): 3609-3617.
[6] 王丰华, 曾全昊, 郑一鸣,等. 基于Bayes估计相空间融合和CM-SVDD的有载分接开关机械故障诊断[J]. 中国电机工程学报,2020,40(01):358-368+402.
WANG Fenghua, ZENG Quanhao, ZHENG Yiming, et al. Mechanical Fault Diagnosis of Load Tap Changers Based on Bayesian Estimation Phase Space Fusion and CM-SVDD[J]. Proceedings of the CSEE, 2020, 40(01): 358-368+402.
[7] 樊家昊, 汪卫国, 王丰华,等. 真空灭弧有载分接开关振动特性分析[J]. 电网技术,2019,43(12):4425-4432.
FAN Jiahao, WANG Weiguo, WANG Fenghua, et al. Vibration Characteristics Analysis of Vacuum Interrupter Load Tap Changers[J]. Power System Technology, 2019, 43(12): 4425-4432.
[8] 蔡宇琦, 方瑞明, 彭长青,等. 基于ANHGA-VMD和耦合隐马尔可夫模型的有载分接开关机械故障诊断[J]. 高电压技术,2021,47(11):3949-3959.
CAI Yuqi, FANG Ruiming, PENG Changqing, et al. Mechanical Fault Diagnosis of Load Tap Changers Based on ANHGA-VMD and Coupled Hidden Markov Model[J]. High Voltage Engineering, 2021, 47(11): 3949-3959.
[9] 赵彤. 有载分接开关机械状态的在线监测与故障诊断技术研究[D]. 山东大学,2008.
ZHAO Tong. Research on Online Monitoring and Fault Diagnosis Technology for the Mechanical State of Load Tap Changers[D]. Shandong University, 2008.
[10] RIVAS E, BURGOS J C, GARCIA-PRADA J C. Condition assessment of power OLTC by vibration analysis using wavelet transform[J]. IEEE Transactions on Power Delivery, 2009, 24(2): 687-694.
[11] E. Rivas, J. C. Burgos, J. C. G. Prada. Vibration analysis using envelope wavelet for detecting faults in the OLTC tap selector. IEEE Transactions on Power Delivery, 2010, 25(3): 1629-1636.
[12] 段若晨, 王丰华, 周荔丹,等. 利用窄带噪声辅助多元经验模态分解算法检测换流变压器用有载分接开关机械状态[J]. 电工技术学报,2017,32(10):182-189.
DUAN Ruocheng, WANG Fenghua, ZHOU Lidan, et al. Detection of Mechanical State of Load Tap Changers for Converter Transformers Using Narrowband Noise-Assisted Multivariate Empirical Mode Decomposition Algorithm[J]. Transactions of China Electrotechnical So-ciety, 2017, 32(10): 182-189.
[13] J. Liu, G. Wang, T. Zhao, L. Zhang. Fault diagnosis of on-load tap-changer based on variational mode decomposition and relevance vector machine. Energies, 2017, 10(7): 946-959
[14] 周翔, 王丰华, 傅坚,等. 基于混沌理论和K-means 聚类的有载分接开关机械状态监测[J]. 中国电机工程学报, 2015, 35(6): 1541-1548.
ZHOU Xiang, WANG Fenghua, FU Jian, et al. Mechanical Condition Monitoring of On-load Tap-Changers Based on Chaos Theory and K-means Clustering[J]. Proceedings of the CSEE, 2015, 35(6): 1541-1548.
[15] 张惠峰, 马宏忠, 陈凯,等. 基于振动信号EMD-HT时频分析的变压器有载分接开关故障诊断. 高压电器, 2012, 48(1): 76-81.
ZHANG Huifeng, MA Hongzhong, CHEN Kai, et al. Fault diagnosis of transformer on-load tap changers based on vibration signal EMD-HT time-frequency analysis. High Voltage Apparatus, 2012, 48(1): 76-81.
[16] 段若晨, 王丰华, 周荔丹. 基于优化HHT算法与洛仑兹信息量度的换流变用有载分接开关机械特征提取[J]. 中国电机工程学报,2016,36(11):3101-3109.
DUAN Ruochen, WANG Fenghua, ZHOU Lidan. Mechanical feature extraction of on-load tap-changer for commutation based on optimized HHT algorithm and Lorentz information measurement[J]. Proceedings of the CSEE,2016,36(11):3101-
3109.
[17] 石卓见, 冉启武, 徐福聪. 基于聚合二次模态分解及Informer的短期负荷预测[J]. 电网技术,2024,48(06):2574-2583.
SHI Zhuojian, RAN Qiwu, XU Fucong. Short-Term Load Forecasting Based on Aggregated Secondary Mode Decomposition and Informer[J]. Power System Technology, 2024, 48(06): 2574-2583.
[18] 李文武, 张鹏宇, 石强, 等. 基于聚合混合模态分解和时序卷积神经网络的综合能源系统负荷修正预测[J]. 电网技术,2022,46(09):3345-3357.
LI Wenwu, ZHANG Pengyu, SHI Qiang, et al. Load Correction Forecasting for Integrated Energy Systems Based on Aggregated Hybrid Mode Decomposition and Temporal Convolutional Neural Network[J]. Power System Technology, 2022, 46(09): 3345-3357.
[19] 彭君哲. 基于时频分析和多通道卷积神经网络的有载分接开关故障诊断研究[D]. 华中科技大学,2022.
PENG Junzhe. Research on Fault Diagnosis of Load Tap Changers Based on Time-Frequency Analysis and Multi-Channel Convolutional Neural Network[D]. Huazhong University of Science and Technology, 2022.
[20] 余长厅,黎大健,陈梁远,等.基于优化TQWT及孪生SVM的有载分接开关机械故障诊断[J].高压电器,2024,60(10):110-118.
YU Changting, LI Dajian, CHEN Liangyuan, et al. Mechanical Fault Diagnosis of On-load Tap-Changer Based on Optimized TQWT and Twin SVM [J]. High Voltage Apparatus, 2024, 60(10): 110-118.
[21] 宋长铭,李岩,王飞,等.基于Transformer的有载分接开关故障诊断研究[J].自动化与仪器仪表,2024,(03):26-29+34.
SONG Changming, LI Yan, WANG Fei, et al. Research on Fault Diagnosis of On-load Tap-Changer Based on Transformer [J]. Automation and Instrumentation, 2024, (03): 26-29+34.
[22] 杨雨昕, 姚伟, 邓贤哲,等. 基于Koopman时序延拓和CNN-Transformer模型的频率稳定指标预测[J/OL]. 电网技术:1-15[2024-06-25]. https://doi.org/10.13335/j.1000-3673.pst.2024.0686
YANG Yuxin, YAO Wei, DENG Xianzhe, et al. Frequency Stability Index Prediction Based on Koopman Time Series Extension and CNN-Transformer Model[J/OL]. Power System Technology: 1-15 [2024-06-25]. https://doi.org/10.13335/j.1000-3673.pst.2024.0686
[23] 周志飞, 李华, 冯毅雄,等. 轻量化深度卷积神经网络设计研究进展[J/OL]. 计算机工程与应用,1-24[2024-08-17].
ZHOU Zhifei, LI Hua, FENG Yixiong, et al. Research progress on the design of lightweight deep convolutional neural network[J/OL]. Computer Engineering and Applications,1-24[2024-08-17].
[24] Yan S, Shao H, Wang J, et al. LiConvFormer: A light-weight fault diagnosis framework using separable multiscale convolution and broadcast self-attention[J]. Expert Systems with Applications, 2024, 237: 121338.