|本期目录/Table of Contents|

[1]曲晓峰,陈光伟.基于改进支持向量机的抽水蓄能发电机转子绕组接地故障检测方法[J].工业仪表与自动化装置,2023,(01):97-102.[doi:10.19950/j.cnki.cn61-1121/th.2023.01.019]
　QU Xiaofeng,CHEN Guangwei.Ground fault detection method of pumped storage generator rotor winding based on improved support vector machine[J].Industrial Instrumentation & Automation,2023,(01):97-102.[doi:10.19950/j.cnki.cn61-1121/th.2023.01.019]
点击复制

基于改进支持向量机的抽水蓄能发电机转子绕组接地故障检测方法

《工业仪表与自动化装置》[ISSN:1000-0682/CN:61-1121/TH]

卷:
期数:
2023年01期
页码:
97-102
栏目:
出版日期:
2023-02-15

文章信息/Info

Title:
Ground fault detection method of pumped storage generator rotor winding based on improved support vector machine
文章编号:
1000-0682（2023）01-0097-06
作者:
曲晓峰; 陈光伟
南方电网调峰调频发电有限公司检修试验分公司，广东 广州 510430
Author(s):
QU Xiaofeng; CHEN Guangwei
Generation Company of China Southern Power Grid Co..Ltd.,Guangzhou Guangdong 510430，China
关键词:
抽水蓄能; 发电机; 支持向量机; 遗传算法
Keywords:
pumped storage;  generator; support vector machine;  genetic algorithm
分类号:
TP391
DOI:
10.19950/j.cnki.cn61-1121/th.2023.01.019
文献标志码:
A
摘要:
抽水蓄能发电机转子发生一点接地故障时，如果不及时处理，将会造成励磁回路过热甚至损坏。为此需要准确检测接地故障，然而受瞬时频率的影响，造成发电机转子故障发生时的时频特征量采集效果较差，影响转子绕组接地故障检测精度，为此，提出了基于改进支持向量机(support vectormachine, SVM)的抽水蓄能发电机转子绕组接地故障检测方法。采用希尔伯特变换提取转子电流的瞬时频率，将瞬时频率处于两个周波期内的平均值与变化标准差视为时频特征量，采用伯格法提取转子电流的频谱特征量；搭建支持向量机分类器模型，利用遗传算法优化该模型的可调参数，并采用K折交叉验证误差及准确率，获取最佳故障检测分类器模型，向分类器中输入时频特征量及频谱特征量，获取抽水蓄能发电机转子绕组接地故障检测结果。实验结果表明：该方法能够高效提取抽水蓄能发电机转子电流信号的时频特征量及频谱特征量，并准确检测出抽水蓄能发电机转子绕组接地故障；具备极佳的收敛性能；寻优效果较好。
Abstract:
If the rotor of pumped storage generator has a little ground fault, if it is not handled in time, the excitation circuit will be overheated or even damaged. Therefore, it is necessary to accurately detect the ground fault. However, due to the influence of instantaneous frequency, the time-frequency feature collection effect of generator rotor fault is poor, which affects the accuracy of rotor winding ground fault detection. Therefore, a rotor winding ground fault detection method of pumped storage generator based on improved support vector machine (SVM) is proposed. Hilbert transform is used to extract the instantaneous frequency of rotor current, the average value and variation standard deviation of instantaneous frequency in two cycle periods are regarded as time-frequency characteristic quantity, and Berg method is used to extract the spectral characteristic quantity of rotor current; The support vector machine classifier model is built, the adjustable parameters of the model are optimized by genetic algorithm, and the k-fold cross validation error and accuracy are used to obtain the best fault detection classifier model. The time-frequency characteristic quantity and spectrum characteristic quantity are input into the classifier to obtain the detection results of rotor winding ground fault of pumped storage generator. The experimental results show that this method can efficiently extract the time-frequency characteristics and spectral characteristics of the rotor current signal of pumped storage generator, and accurately detect the grounding fault of the rotor winding of pumped storage generator; Excellent convergence performance; The optimization effect is good.

参考文献/References:

[1]刘忠,陈星宇,邹淑云,等.计及碳排放的风-光-抽水蓄能系统容量优化配置方法[J].电力系统自动化,2021,45(22):9-18.

[2]张飞,潘伟峰,江献玉,等.基于反时限方法的抽水蓄能机组振动保护模型[J].机械工程学报,2020,56(18):188-196.
[3]张高高,姜海军,徐青,等.基于机组运行特性曲线的变速抽水蓄能机组自适应协调控制方法[J].水力发电,2019,45(08):80-84.
[4]李玉平,桑建斌,朱宇聪,等.基于双端直流注入切换采样原理的发电机转子接地保护[J].电力系统自动化,2020,44(21):139-144.
[5]李永刚,王罗,李俊卿,等.基于多源信息融合的同步发电机转子绕组匝间短路故障识别[J].电力系统自动化,2019,43(16):162-167+191.
[6]武玉才,马倩倩,蔡波冲,等.水轮发电机转子绕组匝间短路故障的新型在线诊断方法[J].电机与控制学报,2018,22(11):19-25.
[7]王宾,崔鑫.基于伏安特性动态轨迹的谐振接地系统弧光高阻接地故障检测方法[J].中国电机工程学报,2021,41(20):6959-6968.
[8]谷艳昌,吴云星,黄海兵,等.基于遗传算法优化支持向量机的大坝安全性态预测模型[J].河海大学学报(自然科学版),2020,48(05):419-425.
[9]张举,王浩,罗舒婷,等.基于遗传算法的混合软件定义网络路由节能算法[J].计算机科学,2020,47(06):236-241.
[10]岳根霞.基于遗传算法的多模态病变图像关联挖掘仿真[J].计算机仿真,2021,38(02):225-229.

相似文献/References:

备注/Memo

备注/Memo:
收稿日期：2022-04-12
第一作者：曲晓峰（1983-），男，汉族，河南巩义人，硕士研究生，高级工程师，研究方向为电气工程及其自动化。
更新日期/Last Update: 1900-01-01