多区域注意力机制变压器高压套管爆炸起火部位判定方法

朴文泉; 杨洋; 李晓光; 赵普志; 刘磊

doi:10.13296/j.1001-1609.hva.2026.07.024

您当前的位置：

首页 >

文章列表页 >

多区域注意力机制变压器高压套管爆炸起火部位判定方法

技术讨论 | 更新时间：2026-07-06

- 多区域注意力机制变压器高压套管爆炸起火部位判定方法
- Method for Determining the Explosion and Fire Location of High Voltage Bushings with Multi-region Attention Mechanism
- 高压电器 2026年62卷第7期页码：230-236
- 作者机构：
  
  国网新疆电力有限公司电力科学研究院，乌鲁木齐 830011
  新疆输变电设备极端环境运行与检测技术重点实验室，乌鲁木齐 830011
  国网新疆电力有限公司，乌鲁木齐 830011
- 作者简介：
  
  朴文泉（1983—），男，硕士，高级工程师，主要研究方向为输变电工程安全技术、电网设备技术监督、变电设备状态分析等（通信作者）（E-mail:llbxdamcs@163.com）。
- 基金信息：
  
  国网新疆电力有限公司电力科学研究院2022年第二批成本专项计划;the 2nd Batch of Special Cost Plan for 2022 of State Grid Xinjiang Electric Power Research Institute(新电科研科【2022】8号);国网新疆电科院2022年生产指挥中心生产任务管控技术服务项目;the 2022 Production Task Management and Control Technical Service Project of the Production Command Center, State Grid Xinjiang Electric Power Research Institute(SGXJDKOOGYJS2200128)
- DOI：10.13296/j.1001-1609.hva.2026.07.024
  中图分类号：
- 收稿：2026-01-11，
  
  修回：2026-03-01，
  
  纸质出版：2026-07-16
- 稿件说明：
移动端阅览
朴文泉, 杨洋, 李晓光, 等. 多区域注意力机制变压器高压套管爆炸起火部位判定方法[J]. 高压电器, 2026,62(7):230-236.

PIAO Wenquan, YANG Yang, LI Xiaoguang, et al. Method for Determining the Explosion and Fire Location of High Voltage Bushings with Multi-region Attention Mechanism[J]. High Voltage Apparatus, 2026, 62(7): 230-236.
朴文泉, 杨洋, 李晓光, 等. 多区域注意力机制变压器高压套管爆炸起火部位判定方法[J]. 高压电器, 2026,62(7):230-236. DOI： 10.13296/j.1001-1609.hva.2026.07.024.

PIAO Wenquan, YANG Yang, LI Xiaoguang, et al. Method for Determining the Explosion and Fire Location of High Voltage Bushings with Multi-region Attention Mechanism[J]. High Voltage Apparatus, 2026, 62(7): 230-236. DOI： 10.13296/j.1001-1609.hva.2026.07.024.

摘要

为准确定位击穿位置，明确变压器高压套管爆炸起火部位，提出基于多区域注意力机制的变压器高压套管爆炸起火部位判定方法。构建变压器高压套管区域检测网络模型，将变压器高压套管原始图像作为特征提取网络的输入，获取其全局特征后，利用多区域判定模块确定可能发生爆炸起火的感兴趣区域，通过非局部关注模块捕捉各感兴趣区域位置间的联系，以感兴趣区域为注意力信息，通过基于CBAM注意力模块的预测网络实现原始变压器高压套管图像和感兴趣区域图像通道、空间信息的融合，利用通道注意力权重提升模型对关键特征通道信息的捕捉，实现变压器高压套管爆炸起火部位的精准判定。实验结果表明：该方法具有变压器高压套管爆炸起火部位判定能力，当提取的感兴趣区域为3、threshold阈值为0.3时，判定精度最优。

Abstract

To ccurately locate the breakdown position and identify the explosion and fire location of the transformer high-voltage bushing

the location determination method for explosion and fire of the transformer high-voltage bushing based on the multi-region attention mechanism is proposed. A detection network model for the transformer highvoltage bushing area is constructed.The original image of the transformer high-voltage bushing is used as the input of the feature extraction network to obtain its global features. Then

a multi-region determination module is used to identify the regions of interest where explosions and fires may occur. A non-local attention module is adopted to capture the relationship among the locations of each region of interest. Through a prediction network based on CBAM attention module

the fusion of channel and spatial information between the original transformer high-voltage bushing image and the region-of-interest images is achieved. The channel attention weights are used to enhance the model's ability to capture key feature channel information

thereby realizing precise determination of the explosion and fire location of the transformer high-voltage bushing. Experimental results show that this method is capable of determining the explosion and fire location of transformer high-voltage bushing. When the number of extracted region of interest is 3 and the threshold value is 0.3

the determination accuracy is optimal.

关键词

Keywords

references

张少华,蒋慧灵,刘国平.某特高压变压器爆炸事故数值模拟重构[J].消防科学与技术,2021,40(3):356-359.

Zhang Shaohua, Jiang Huiling, Liu Guoping. Numerical simulation and reconstruction of explosion accident of an UHV transformer[J]. Fire Science and Technology,2021,40(3):356-359.

邱晟璇,王松,刘子瑞,等.基于FRA法和CNN的变压器绕组故障诊断研究[J].大电机技术,2025(3):108-116.

Qiu Shengxuan, Wang Song, Liu Zirui, et al.Research on fault diagnosis of transformer windings based on FRA method and CNN[J]. Large Electric Machine and Hydraulic Turbine ,2025(3):108-116.

余阳,阮娅琳,杨黎波,等.换流变压器油蒸汽爆炸的三维仿真研究[J].电瓷避雷器,2021(2):155-161.

Yu Yang, Ruan Yalin, Yang Libo, et al. 3D simulation of oil vapor explosion in converter transformer[J]. Insulators and Surge Arresters,2021(2):155-161.

马鑫,尚毅梓,胡昊,等.基于数据特征增强和残差收缩网络的变压器故障识别方法[J].电力系统自动化,2022,46(3):175-183.

Ma Xin, Shang Yizi, Hu Hao, et al. Identification method of transformer fault based on data feature enhancement and residual shrinkage network[J]. Automation of Electric Power Systems,2022, 46(3):175-183.

裴丽君,徐慧,陈晗,等.基于PSO优化的SA-LSTM变压器绕组温度预测方法[J].供用电,2025,42(8):112-119.

Pei Lijun, Xu Hui, Chen Han, et al. Prediction method for winding temperature of SA-LSTM transformer based on PSO optimization[J]. Distribution & Utilization,2025,42(8):112-119.

刘云鹏,董王英,许自强,等.基于卷积神经网络的变压器套管故障红外图像识别方法[J].高压电器,2021,57(10):134-140.

Liu Yunpeng, Dong Wangying, Xu Ziqiang, et al. Infrared image recognition method on fault of transformer bushing based on convolutional neutral networks[J]. High Voltage Apparatus,2021, 57(10):134-140.

严敏超,陈明浩,王稼轩,等.两起变压器套管将军帽发热处理与分析[J].变压器,2025,62 (4):65-70.

Yan Minchao, Chen Minghao, Wang Jiaxuan, et al. Treatment and analysis of two cases of general cap of transformer bushing[J]. Transformer,2025,62 (4):65-70.

陈晋鹏,杨昊,张璐,等.环境温度变化对含局部过热缺陷的油纸套管温升特性影响研究[J].电瓷避雷器,2024(3):131-139.

Chen Jinpeng, Yang Hao, Zhang Lu, et al.Impact of ambient temperature changes on temperature rise characteristics of oip bushings with local overheating defects[J]. Insulators and Surge Arresters,2024(3):131-139.

廖文龙,刘凡,陈家慧,等.一起500 kV高压电抗器套管内部环流过热缺陷分析[J].变压器,2024,61(3):67-71.

Liao Wenlong, Liu Fan, Chen Jiahui, et al. Analysis of a circulating current overheating defect in 500 kV high-voltage reactor bushing[J]. Transformer,2024,61(3):67-71.

李旭,王文森,郭丽,等.基于多传感器融合的电力变压器内部放电定位与辨识技术[J].电网与清洁能源,2024,40(3):22-31.

Li Xu, Wang Wensen, Guo Li, et al.Internal discharge location and identification for power transformers based on multi-sensor fusion[J]. Power System and Clean Energy,2024,40(3):22-31.

何楠,刘弘景,闫春江,等.基于改进深度自注意力网络的变压器及其部件高分辨率图像识别技术[J].变压器,2022,59(8):23-27.

He Nan, Liu Hongjing, Yan Chunjiang, et al. Recognition of transformer and its components based on modified deep selfattention network[J]. Transformer,2022,59(8):23-27.

万吉林,吴国强,管敏渊,等.基于RetinaNet的变压器图像小部件智能识别方法[J].电力系统保护与控制,2021,49(12):166-173.

Wan Jilin, Wu Guoqiang, Guan Minyuan, et al. Intelligent recognition method for transformer small components based on RetinaNet[J]. Power System Protection and Control,2021,49(12):166-173.

刘洋,金忠.一种结合非局部和多区域注意力机制的细粒度图像识别方法[J].计算机科学,2021,48(1):197-203.

Liu Yang, Jin Zhong. Fine-grained image recognition method combining with non-local and multi-region attention mechanism[J]. Computer Science,2021,48(1):197-203.

彭晏飞,梅金业,王恺欣,等.基于区域注意力机制的遥感图像检索[J].激光与光电子学进展,2020,57(10):172-180.

Peng Yanfei, Mei Jinye, Wang Kaixin, et al. Remote sensing image retrieval based on regional attention mechanism[J]. Laser &Optoelectronics Progress,2020,57(10):172-180.

赵洪山,孙京杰,彭轶灏,等.基于多尺度窗口和区域注意力残差网络的无线电力终端身份识别方法[J].电工技术学报,2023, 38(1):107-116.

Zhao Hongshan, Sun Jingjie, Peng Yihao, et al. Wireless power terminal identification method based on multiscale windowed deep residual network[J]. Transactions of China Electrotechnical Society,2023,38(1):107-116.

孙明浩,王洪元,吴琳钰,等.基于特征金字塔分支和非局部关注的行人重识别[J].数据采集与处理,2023,38(1):121-131.

Sun Minghao, Wang Hongyuan, Wu Linyu, et al. Person reidentification based on feature pyramid branch and non-local attention[J]. Journal of Data Acquisition & Processing,2023,38(1):121-131.

刘紫燕,朱明成,袁磊,等.基于非局部关注和多重特征融合的视频行人重识别[J].计算机应用,2021,41(2):530-536.

Liu Ziyan, Zhu Mingcheng, Yuan Lei, et al. Video person reidentification based on non-local attention and multi-feature fusion[J]. Journal of Computer Applications,2021,41(2):530-536.

林润超,黄荣,董爱华.基于注意力机制和元特征二次重加权的小样本目标检测[J].计算机应用,2022,42(10):3025-3032.

Lin Runchao, Huang Rong, Dong Aihua. Few-shot object detection based on attention mechanism and secondary reweighting of metafeatures[J]. Journal of Computer Applications,2022,42(10):3025-3032.

朱威,王立凯,靳作宝,等.引入注意力机制的轻量级小目标检测网络[J].光学精密工程,2022,30(8):998-1010.

Zhu Wei, Wang Likai, Jin Zuobao, et al. Lightweight small object detection network with attention mechanism[J]. Optics and Precision Engineering,2022,30(8):998-1010.

Sudha G, Thyagesh G K, Prabhakaran V A, et al. High pressure turbine casing expansion measurements in thermal power stations using smart sensor[J]. Test Engineering and Management,2021 (82):7629-7634.

Baojun G, Jiong Z, Tao D.Temperature prediction and cooling structure optimization of explosion-proof high pressure water-cooled double speed motor[J]. Energy Reports,2022(8):3891-3901.

Smith P. Determining design pressure and temperature limits for a new high - speed, very high pressure steam turbine casing[J]. Chemical Engineering World,2021(7):56.

Mao X, Liankang H E, Liu J, et al. Mechanism of the strong earthquake triggered by high pressure fluid in reservoir:a case study of the 5.12 Wenchuan earthquake[J]. Journal of Geomechanics,2021, 27(4):628-642.

Guo X Q, Liu J, Huang L, et al.Wear experiment and influencing factors analysis of 13Cr-L80 tubing string in high-pressure, hightemperature and high-yield gas wells[J]. Materials Science Forum, 2021(1026):169-175.

Huang Lijuan, Wang Yu, Zhang Liang, et al. Influence of pressure on the flammability limits and explosion pressure of ethane/propane-air mixtures in a cylinder vessel[J]. Journal of Loss Prevention in the Process Industries,2022(74):104638-104639.

Huang Lijuan, Wang Yu, Li Zongfa, et al. Experimental study on piloted ignition temperature and auto ignition temperature of heavy oils at high pressure[J]. Energy,2021,229(4):120644.

Huang Lijuan, Li Zongfa, Wang Yu, et al. Experimental assessment on the explosion pressure of CH 4 -air mixtures at flammability limits under high pressure and temperature conditions[J ] . Fuel,2021 (1):299.

Nguyen T, Nguyen L, Tran P, et al.Improving transformer-based neural machine translation with prior alignments[J]. Complexity, 2021(16):2021.

Gao S, Zhao J, Liu Y, et al. Research into power transformer health assessment technology based on uncertainty of information and deep architecture design[J]. Mathematical Problems in Engineering,2021(6):1-12.

浏览量

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

交、直流电压下126 kV GIS支柱绝缘子内部气泡缺陷局部放电特性

泄漏电流与高频脉冲信号联合检测的集成式传感器研究

基于振动信号的断路器刚分刚合点计算方法

基于注意力—门控循环—时间卷积网络的变压器油温预测

基于场景生成的构网型储能优化配置技术研究