摘要
为解决配电网发生单相接地故障时仅依靠比较动作阈值出口跳闸,不能对单相接地具体故障类型进行有效识别的问题,提出了一种基于卷积神经网络(CNN)自适应识别电弧接地和电阻接地的方法。研究了4种单相接地的故障类型,在PSCAD/EMTDC中搭建了10 kV配电网模型进行仿真。利用希尔伯特-黄变换(HHT)构造出故障信号的时频谱图,以此作为CNN的输入,在故障特征量被CNN自主提取后能够分类识别单相接地故障类型。Matlab仿真结果表明:该方法与传统机器学习算法相比具有更高的准确率。试验结果表明:投入消弧线圈、调整网络结构和加入噪声污染后,对单相接地故障类型的识别也具有良好的适应性。
In order to solve the problem that the outlet tripping only depends on the comparison of action threshold when the single-phase grounding fault occurs in the distribution network, and the specific fault types of single-phase grounding cannot be effectively identified, a method based on convolutional neural network(CNN) for adaptive identification of arc grounding and resistance grounding is proposed. Four single-phase grounding fault types are studied, and a 10 kV distribution network model is built in PSCAD/EMTDC for simulation. Hilbert-Huang transform(HHT) is used to construct the time-frequency spectrum of fault signal, which is used as the input of CNN. After the fault feature is independently extracted by CNN, the single-phase grounding fault type can be classified and identified. The verification in matlab environment shows that this method has higher accuracy than the traditional machine learning algorithm. At the same time, the test results show that the identification of single-phase grounding fault type has good adaptability by adding arc suppression coil, adjusting network structure and adding noise pollution.
作者
杨佳
陈勇
冯波
王佳豪
潘鑫
钟加勇
YANG Jia;CHEN Yong;FENG Bo;WANG Jiahao;PAN Xin;ZHONG Jiayong(School of Electrical and Electronic Engineering,Chongqing University of Technology,Chongqing 400054,China;Chongqing Engineering Research Center of Energy Internet,Chongqing 400054,China;State Grid Chongqing Electric Power Research Institute,Chongqing 401123,China)
出处
《重庆理工大学学报（自然科学）》
CAS
北大核心
2022年第8期236-245,共10页
Journal of Chongqing University of Technology：Natural Science
基金
重庆市教委科学技术研究重点项目(KJZD-K201901102)
重庆市教育委员会科学技术研究计划青年项目(KJQN201801113,KJQN202101147)
重庆市技术创新与应用发展专项面上项目(cstc2020jscx-msxmX0210)
重庆理工大学研究生创新项目(clgycx20203037)。
