摘要
高压套管是变电站的关键设备,频域介电谱可有效地反映其绝缘状态变化,但无法得到具体的老化与受潮状态。为了有效评估高压套管的老化与受潮状态,提出了一种高压套管绝缘状态评估方法。首先,利用最小输出编码原理设计了基于最小二乘支持向量机(LS-SVM)的多分类模型,然后提出了基于布谷鸟算法优化多分类LS-SVM的评估流程,最后利用试验套管获取训练与测试样本,对评估模型进行了训练与验证。结果表明:提出的多分类LS-SVM模型参数寻优中布谷鸟算法的寻优性能明显优于PSO与GA等传统寻优算法;提出的高压套管绝缘状态评估方法能够有效地区分套管绝缘缺陷类型,并有效地评估状态。
High-voltage bushing is the key equipment in a substation.The frequency domain dielectric spectrum can effectively reflect changes of its insulation state,but details about its aging and wet condition cannot be obtained.For the purpose of effective assessment of its aging and wet conditions,an assessment method for the insulation state of high-voltage bushing was proposed in this paper.Firstly,a multi-classification model based on least square support vector machines(LS-SVM)was designed by using the principle of minimum output coding.Then,an optimized multi-classification LS-SVM assessment process based on the cuckoo algorithm was proposed.Finally,training and testing samples were obtained by use of experimental bushing,and the assessment model was trained and verified.The results showed that:The searching performance of proposed multi-classification LS-SVM model based on the cuckoo algorithm was obviously better than traditional searching algorithms like PSO and GA;The proposed assessment method for the insulation state of the high-voltage bushing could effectively distinguish types of insulation defects and assess the state.
作者
马旭斌
杨勇
马鹏
Ma Xubin;Yang Yong;Ma Peng(Alxa Electric Power Bureau of Inner Mongolia Electric Power(Group)Co.,Ltd.,Alxa Nei Mongol 750306,China)
出处
《电气自动化》
2021年第4期60-62,共3页
Electrical Automation
关键词
高压套管
绝缘状态
频域介电谱测试
布谷鸟算法
最小二乘支持向量机
high-voltage bushing
insulation state
frequency domain dielectric spectrum test
cuckoo algorithm
least square support vector machines(LS-SVM)
