摘要
针对温度快速上升与温度缓慢上升两类电力设备载流故障形态,提出了基于相关性分析与主元提取的载流故障检测方法。首先分析了触头在空间上的纵横关系,并基于电气规范提出了纵横对比的数据处理方法来消除负载影响;再通过主成分分析(PCA)法提取特征值;最后通过对特征值超阈值情况的分析来给出预警信号。测试结果表明,该方法能够有效地实现对多种载流故障的预警,并在对温度缓慢上升的故障实施预警时,预警时间显著提前。
A current-carrying fault early warning approach based on relevance analysis and principal component analysis（PCA） al- gorithm is proposed, which focus on both quiek and slow rise of the electric equipment contact temperature. Firstly, the spatial relationship among contacts is analyzed and a method based on electrical specifications to eliminate the load influence is proposed. Secondly, the characteristic value is extracted by using PCA. Finally, the warning signal is given by analyzing the characteristic value. The above method was verified with the actual temperature data of a power station, and the result shows that the method can realize current-carrying fault early warning accurately, especially for the slow mode.
出处
《中国科技论文》
CAS
北大核心
2015年第11期1240-1244,共5页
China Sciencepaper
基金
高等学校博士学科点专项科研基金资助项目(20130101110111)
关键词
载流故障
时间序列
空间相关性
主成分分析(PCA)
故障检测
current-carrying fault
time series
spatial correlation analysts
principal component analysis （PCA）
failure detection
