[1]王子伟,疏佳铭,陆 斌,等.基于 GASF 及双输入 AlexNet-SVM 的变压器励磁涌流检测[J].机械与电子,2025,(08):30-39.
　WANG Ziwei,SHU Jiaming,LU Bin,et al.Transformer Inrush Current Detection Based on GASF and Dual-input AlexNet-SVM[J].Machinery & Electronics,2025,(08):30-39.
点击复制

基于 GASF 及双输入 AlexNet-SVM 的变压器励磁涌流检测()
分享到：

《机械与电子》[ISSN:1001-2257/CN:52-1052/TH]

卷:
期数:
2025年08期
页码:
30-39
栏目:
自动控制与检测
出版日期:
2025-08-25

文章信息/Info

Title:
Transformer Inrush Current Detection Based on GASF and Dual-input AlexNet-SVM
文章编号:
1001-2257 ( 2025 ) 08-0030-10
作者:
王子伟 1 ; 疏佳铭 2 ; 陆 斌 1 ; 张 敏 1 ; 倪继文 3
1. 华能澜沧江水电股份有限公司,云南 昆明 650214 ;
2. 南京南瑞继保电气有限公司,江苏 南京 211102 ;
3. 西安热工研究院有限公司,陕西 西安 710043
Author(s):
WANG Ziwei1 ;  SHU Jiaming2 ;  LU Bin1 ;  ZHANG Min1 ;  NI Jiwen3
( 1.Huaneng Lancang River Hydropower Co. , Ltd. , Kunming 650214 , China ;
2.Nanjing NR Electric Co. , Ltd. , Nanjing 211102 , China ;
3.Xi ’ an Thermal Power Research Institute Co. , Ltd. , Xi ’ an 710043 , China )
关键词:
电力变压器; 励磁涌流; 内部故障;  GASF 变换; 双输入 AlexNet-SVM 模型; 支持向量机
Keywords:
power transformer ;  inrush current ;  internal fault ;  GASF transformation ;  dual-input AlexNet-SVM model ;  support vector machine
分类号:
TP18 ;TM41
文献标志码:
A
摘要:
针对励磁涌流易于引发电力变压器差动保护误动作的问题,提出一种基于格拉姆角和场( GASF )及双输入 AlexNet-SVM 的变压器励磁涌流检测方法,旨在提升励磁涌流检测的准确性并缩短检测时间。首先,采用 GASF 变换及图像编码将三相差动电流转换为彩色图,提升励磁涌流与内部故障的工况特征区分度。其次,针对励磁涌流检测的快速性需求对 AlexNet 进行轻量化设计,并采用 SVM 作为分类器构建了 AlexNet-SVM 模型。最后,用不同的采样方式构建了 4 个励磁涌流检测模型,即单输入 AlexNet-SVM 模型、扩维单输入 AlexNet SVM 模型、并行双输入 AlexNet-SVM 模型和级联双输入 AlexNet-SVM 模型;并采用励磁涌流及内部故障仿真数据集对多模型进行了测试和评估。结果表明,级联双输入AlexNet-SVM 模型可在取得 100% 准确率的同时将模型平均检测时间降到 10.51 ms ,实现了励磁涌流检测准确性和实时性的同步提升。
Abstract:
To address the issue of differential protection maloperation caused by inrush current in power transformers , a method for identifying transformer inrush current based on GASF and dual-input AlexNet-SVM was proposed , aiming to enhance the accuracy of inrush current detection and reduce detection time.Firstly , the three-phase differential current was transformed into color images using Gramian angular summation field ( GASF ) transformation and image encoding , which enhances the distinction of working condition characteristics between inrush current and internal faults.Secondly , A lightweight AlexNet was designed to meet the real-time requirement for inrush current detection , and SVM was used as the classifier of AlexNet to construct the AlexNet-SVM model.Finally , four inrush current detection models were built according to different sampling methods as single input AlexNet-SVM model , dimension extended single input AlexNet-SVM model , parallel dual-input AlexNet-SVM model , and cascaded dual input AlexNet-SVM model ; and they were validated and assessed using simulation datasets containing inrush current and internal faults.The results indicate that the cascaded dual input model can achieve 100% detection accuracy while reducing the average detection time of the model to 10.51 ms , thereby the accuracy and real-time performance of inrush current detection are both improved.

参考文献/References:

[ 1 ] 张国栋,刘凯,蒲海涛,等 . 基于长短时记忆神经网络的励磁涌流与故障电流识别方法[ J ] . 上海交通大学学报,2024 , 58 ( 5 ): 730-738.

[ 2 ] 姚东晓,张凯,贺要锋,等 . 变压器多特征励磁涌流识别方案研究[ J ] . 电力系统保护与控制, 2017 , 45 ( 13 ): 149-154.
[ 3 ] 沈春城,严柏平,黄大卓,等 . 基于波形复杂特性的励磁涌流快速识别算法研究[ J ] . 电气工程学报,2024 , 19( 1 ): 243-253.
[ 4 ] 张运驰,高厚磊,杜士昌 . 基于综合形态算法的变压器励磁涌流识别方法 [ J ] . 电力系统自动化,2021 , 45( 24 ): 165-173.
[ 5 ] 黄大卓,严柏平,沈春城,等 . 基于二次谐波分量衰减特性的变压器励磁涌流识别方法[ J ] . 变压器,2023 , 60( 1 ): 28-34.
[ 6 ] LUAN Y , LI H , ZENG R , et al.Statistical similarity based identification of power transformer inrush current [ C ] ∥ 2022 7th Asia Conference on Power and Electrical Engineering ( ACPEE ), 2022 : 1568-1572.
[ 7 ] TAJDINIAN M , SAMET H.Application of probabilistic distance measures for inrush and internal fault currents discrimination in power transformer differential protection [ J ] .Electric power systems research , 2021 , 193 : 107012.
[ 8 ] SAMET H , SHADAEI M , TAJDINIAN M.Statistical discrimination index founded on rate of change of phase angle for immunization of transformer differential protection against inrush current [ J ] .International journal of electrical power and energy systems , 2022 , 134 : 107381.
[ 9 ] ETUMI A , ANAYI F.Current signal processing-based methods to discriminate internal faults from magnetizing inrush current [ J ] .Electrical engineering , 2021 , 103 ( 1 ):743-751.
[ 10 ] 李宗博,焦在滨,何安阳 . 基于卷积神经网络特征迁移策略的变压器智能保护方法[ J ] . 中国电机工程学报,2021 , 41 ( 15 ): 5201-5211.
[ 11 ] JIAO W , DONG K , ZHAO J.Intelligent transformer protection method based on convolutional neural network [ C ] ∥2021 4th International Conference on Energy , Electrical and Power Engineering ( CEEPE ), 2021 : 698-703.
[ 12 ] RUHAN Z , MANSOR N N B , ILLIAS H A.Identification of inrush current using a GSA-BP network[ J ] .Energies , 2023 , 16 ( 5 ): 2340.
[ 13 ] MENG J , DU J Y.Research on magnetizing inrush current and fault identification of transformer based on VMD-SVM [ C ] ∥2020 IEEE International Conference on Information Technology , Big Data and Artificial Intelligence ( ICIBA ), 2020 : 172-178.
[ 14 ] DUAN P , HE Y , ZHANG L F , et al.Research on intelligent identification of magnetizing inrush current based on empirical modal decomposition [ J ] .Journal of physics : conference series , 2023 , 2495 ( 1 ): 012001.
[ 15 ] AFRASIABI S , AFRASIABI M , PARANG B , et al.Fast GRNN-based method for distinguishing inrush currents in power transformers [ J ] .IEEE Transactions on industrial electronics , 2022 , 69 ( 8 ): 8501-8512.
[ 16 ] HE A , JIAO Z , LI Z , et al.Discrimination between internal faults and inrush currents in power transformers based on the discriminative-feature-focused CNN [ J ] .Electric power systems research , 2023 , 223 : 109701.
[ 17 ] AFSHARISEFAT R , JANNATI M , SHAMS M.A power transformer differential protection method based on variational mode decomposition and CNN BiLSTM techniques [ J ] .IET Generation , transmission and distribution , 2024 , 18 ( 4 ): 767-778.
[ 18 ] KEY S , SON G , NAM S.Deep learning-based algorithm for internal fault detection of power transformers during inrush current at distribution substations [ J ] .Energies ( Basel ), 2024 , 17 ( 4 ): 963.
[ 19 ] ZHANG Q , QI Z , CUI P Y , et al.Detection of single phase-to-ground faults in distribution networks based on Gramian angular field and improved convolutional neural networks [ J ] .Electric power systems research , 2023 , 221 : 109501.
[ 20 ] WANG X , LIANG Y , PAN Q , et al.A Gaussian approximation recursive filter for nonlinear systems with correlated noises [ J ] .Automatica , 2012 , 48 ( 9 ):2290-2297.
[ 21 ] 肖雨松,马宏忠 . 基于格拉姆角场和深度残差网络的变压器绕组松动故障诊断模型[ J ] . 电机与控制应用,2024 , 51 ( 1 ): 29-38.
[ 22 ] 苏斌,侯思祖,郭威 . 基于图像融合和双通道卷积神经网络的配电网故障选线方法研究[ J ] . 电子测量与仪器学报,2024 , 38 ( 9 ): 54-66.
[ 23 ] 赵扬,耿莉敏,胡循泉,等 . 采用格拉姆角场 卷积神经网络 时序卷积网络混合模型的锂离子电池健康状态估计[ J ] . 西安交通大学学报, 2024 , 58 ( 11 ): 27-38.
[ 24 ] KHAN A , SOHAIL A , ZAHOORA U , et al.A survey of the recent architectures of deep convolutional neural networks [ J ] .Artificial intelligence review , 2020 , 53 : 5455-5516.
[ 25 ] CUI Y , WANG R , SI Y , et al.T-type inverter fault diagnosis based on GASF and improved AlexNet [ J ] . Energy reports , 2023 , 9 : 2718-2731.
[ 26 ] 刘冬梅,霍龙龙,王浩然,等 . 基于 PSO-SVM的电流放大器故障诊断研究[ J ] . 传感器与微系统, 2021 , 40( 8 ): 50-52 , 56.
[ 27 ] GAO Z , XUE H , WAN S.Multiple discrimination and pairwise CNN for view-based 3D object retrieval [ J ] . Neural networks , 2020 , 125 : 290-302.
[ 28 ] 宁铎,尤磊,李英春,等 . 变压器差动保护动作特性的仿真研 究[ J ] . 电力系统保护与控制,2017 , 45 ( 4 ):99-106.

相似文献/References:

[1]王 峰.基于模糊神经网络的电力变压器绝缘状态预测方法[J].机械与电子,2022,(03):50.
　WANG Feng.Prediction Method of Insulation State of Power Transformer Based on Fuzzy Neural Network[J].Machinery & Electronics,2022,(08):50.

备注/Memo

备注/Memo:
收稿日期: 2025-02-10
基金项目:中国华能集团有限公司科技项目( HNKJ22-H88 )
作者简介:王子伟 ( 1967- ),男,云南永胜人,正高级工程师,研究方向为电力生产管理、电力二次系统控制;疏佳铭 ( 1990- ),男,安徽铜陵人,硕士,工程师,研究方向为电力主设备继电保护研究与开发。
更新日期/Last Update: 2025-09-05