摘要
为提高柔性直流换流阀故障分类正确率,提出基于改进蜣螂优化算法(IDBO)的卷积神经网络融合-核极限学习机(CNN-KELM)柔性直流换流阀故障分类方法。对柔性直流换流阀故障特征库归一化处理,利用CNN进行故障特征提取,采用IDBO优化KELM的核参数和惩罚因子。将IDBO-CNN-KELM作为分类器对提取到的柔性直流换流阀故障库进行分类。通过实验证明,IDBO-CNN-KELM模型在故障测试集中的分类正确率达到97.727%,相较传统KELM、PSO(粒子群算法)-CNN-KELM提升了1.136%、0.577%,证明了IDBO-CNN-KELM模型的精确性。该方法有效提高了柔性直流换流阀故障分类的准确性和效率,增强了电网直流输电可靠性。
In order to improve the correct rate of flexible DC converter valve fault classification,a flexible DC converter valve fault classification method based on the convolutional neural network fused kernel-extreme learning machine(CNN-KELM)optimized by the Improved dung beetle optimization algorithm(IDBO)is proposed.The flexible DC converter valve fault feature library is normalized,the CNN network is used for fault feature extraction,and IDBO is used to optimize the kernel parameters and penalty factors of KELM.IDBO-CNN-KELM is used as a classifier to classify the extracted flexible DC converter valve fault library.Through experiments,the IDBO-CNN-KELM model achieves a classification correctness of 97.727%in the fault test set,which improves 1.136%and 0.577%compared with the traditional KELM and PSO-CNN-KELM,proving the accuracy of the IDBO-CNN-KELM model.The method effectively improves the accuracy and efficiency of fault classification of flexible DC converter valves,and enhances the reliability of DC transmission in power grids.
作者
马群
戈一航
刘黎
田蘅
MA Qun;GE Yihang;LIU Li;TIAN Heng(State Grid Electric Power Research Institute,Nanjing 211106,China;Beijing Kedong Electric Power Control System Co.,Ltd.,Beijing 100092,China;Zhoushan Power Supply Company of State Grid Zhejiang Electric Power Co.,Ltd.,Zhoushan 316000,Zhejiang,China)
出处
《科技和产业》
2024年第21期311-318,共8页
Science Technology and Industry
基金
国家电网公司科技项目(5500-202140505A-0-5-ZN)。
关键词
柔性直流换流阀
蜣螂优化算法
卷积神经网络
核极限学习机
故障辅助决策
flexible DC converter valve
dung beetle optimisation algorithm
convolutional neural network
nuclear limit learning machine
fault-assisted decision making
作者简介
马群(1987-),男,山东济宁人,硕士,工程师,研究方向为电力系统仿真、电力设备三维数字孪生技术;戈一航(1998-),男,河北石家庄人,硕士,助理工程师,研究方向为电力系统自动化;刘黎(1987-),女,浙江舟山人,硕士,高级工程师,研究方向为柔性直流输电运维检修技术;田蘅(1970-),男,江苏南京人,硕士,教授级高级工程师,研究方向为电力系统运行监控、电力系统建模与仿真、特高压直流技术及设备装置研发与应用。
