| 7 | 0 | 26 |
| 下载次数 | 被引频次 | 阅读次数 |
真空断路器分闸过程中产生的电弧会直接影响触头的使用寿命，为探究电弧形态与触头寿命之间的关系，需利用高速相机采集断路器断口处电弧图像，精准捕捉分闸过程中的电弧形态。传统的图像特征提取方法受不同阶段图像亮度差异干扰，存在泛化能力弱、抗噪性不足及精度较低等问题。基于深度学习方法设计电弧形态提取网络，结合空洞卷积和残差结构搭建特征提取网络，引入Dense-ASPP结构融合特征，采用辅助分类器提升分割精度。结果表明，该方法在保证精度的同时提升了电弧形态特征提取的泛化能力，算法对电弧图像的交并比可达94%，为后续研究电弧形态对触头的影响机理提供数据支撑。
Abstract：The arc generated during the opening process of a vacuum circuit breaker directly affects the service life of the contacts. To investigate the relationship between the arc morphology and the contact life, it is necessary to use a high-speed camera to collect the arc images at the breaking point of the circuit breaker and accurately capture the arc morphology during the opening process. Traditional image feature extraction methods are disturbed by the brightness differences of images at different stages, and have problems such as weak generalization ability, insufficient noise resistance and low accuracy. A deep learning-based arc morphology extraction network is designed, and a feature extraction network is built by combining dilated convolution and residual structure. The Dense-ASPP structure is introduced to fuse features, and an auxiliary classifier is adopted to improve the segmentation accuracy. The results show that this method improves the generalization ability of arc morphology feature extraction while ensuring accuracy. The intersection over union of the algorithm for arc images can reach 94%, providing data support for the subsequent study of the mechanism of arc morphology on the contacts.
[1]董华军，温超阳，刘林林，等．触头系统结构对真空灭弧室温度场影响仿真[J]．电机与控制学报，2023,27(8):164-172.DONG H J, WEN C Y, LIU L L, et al. Simulation of the influence of the structure of the contact system on the temperature field of the vacuum interrupter[J]. Electric Machines and Control, 2023, 27(8):164-172.
[2]LIU S, YUAN Z, LIU L M, et al. Experimental study on dynamic characteristics of charged particles in vacuum arc[J]. Measurement Science and Technology, 2024, 35(2):025402.
[3]马丽婷，赵一鉴，邢喜超，等．真空灭弧室内电弧形态演变研究[J]．广东电力，2024, 37(6):104-110.MA L T, ZHAO Y J, XING X C, et al. Research and analysis on evolution of arc morphology in vacuum arc extinguishing chambers[J]. Guangdong Electric Power,2024, 37(6):104-110.
[4]岳宇鑫．真空断路器失效机理分析及健康状况检测方法[J]．仪器仪表与分析监测，2023(1):27-31.YUE Y X. Failure mechanism analysis of vacuum circuit breaker and health status detection method[J]. Instrumentation Analysis Monitoring, 2023(1):27-31.
[5]郑一鸣，李星，李晨，等．真空开关选相操作特性在线检测消除电感性电流开断的操作过电压[J]．高压电器，2021, 57(7):18-23.ZHENG Y M, LI X, LI C, et al. Online detection of phase selection operation performance of vacuum switch for switching overvoltage with elimination of inductive current interruption[J]. High Voltage Apparatus, 2021,57(7):18-23.
[6]孙华，王宾．基于分合闸电气量的真空断路器灭弧能力在线检测研究[J]．电力系统保护与控制，2022, 50(2):102-109.SUN H, WANG B. Arc-quenching ability detection of a high voltage vacuum circuit breaker based on switching and closing current[J]. Power System Protection and Control, 2022, 50(2):102-109.
[7]李奕，王振兴，刘志远，等．横向磁场下真空电弧运动形态三维重构研究[J]．真空电子技术，2019(5):14-20.LI Y, WANG Z X, LIU Z Y, et al. Three-dimensional reconstruction of vacuum arc motion under transverse magnetic field[J]. Vacuum Electronics, 2019(5):14-20.
[8]董华军，张作凡，李博，等．基于PIV技术的真空电弧扩散阶段流场分析[J]．大连交通大学学报，2023,44(4):109-113.DONG H J, ZHANG Z F, LI B, et al. Research on quantitative analysis of flow field in diffusion stage of vacuum switch based on PIV technology[J]. Journal of Dalian Jiaotong University, 2023, 44(4):109-113.
[9]徐鸿飞，李争博，付思，等．磁控触头旋转开断过程中直流真空电弧形态与电压特性[J]．真空科学与技术学报，2024, 44(7):636-645.XU H F, LI Z B, FU S, et al. The morphology and voltage characteristics of DC vacuum arc during the rotation and disconnection process of magnetic control contacts[J]. Chinese Journal of Vacuum Science and Technology, 2024, 44(7):636-645.
[10]李雨鑫．基于深度学习的红外与可见光图像融合算法研究及硬件实现[D]．西安：西安理工大学，2023.LI Y X. Research and hardware implementation of infrared and visible image fusion algorithm based on deep learning[D]. Xi’an:Xi’an University of Technology, 2023.
[11]WANG J, ZHANG X L, YAN T H, et al. DPNet:Dualpyramid semantic segmentation network based on improved Deeplabv3 plus[J]. Electronics, 2023, 12(14):3161.
[12]何配林．基于深度学习的工业仪表识别读数算法研究及应用[D]．成都：电子科技大学，2020.HE P L. Deep learning-based recognition algorithm for industry meters and its applications[D]. Chengdu:University of Electronic Science and Technology of China, 2020.
[13]向川，王惠，史鹏飞，等．基于改进脉冲耦合神经网络模型的真空电弧燃烧过程研究[J]．电工技术学报，2019, 34(19):4028-4037.XIANG C, WANG H, SHI P F, et al. Research on the combustion process of vacuum arc based on an improved pulse coupled neural network model[J]. Transactions of China Electrotechnical Society, 2019, 34(19):4028-4037.
[14]林宁宁，高心丹．DeepLabV3+改进的树木图像分割[J]．计算机工程与设计，2023, 44(1):232-239.LIN N N, GAO X D. Improved trees image segmentation based on DeepLab V3+[J]. Computer Engineering and Design, 2023, 44(1):232-239.
[15]DUAN L X, TSANG I W, XU D, et al. Domain adaptation from multiple sourcesviaauxiliary classifiers[C]∥Proceedings of the 26th Annual International Conference on Machine Learning. New York:ACM, 2009:289-296.
基本信息:
DOI：10.13291/j.cnki.djdxac.2026.04.018
中图分类号:TP18;TM561.2
引用信息:
[1]王海燕,李小钊,柴娜,等.基于深度学习的电弧形态提取算法[J].大连交通大学学报().DOI:10.13291/j.cnki.djdxac.2026.04.018.
2026-08-17
2026-08-17
2026-08-17