摘要
提出一种新颖的基于决策树方法的电能质量扰动自动分类方法.该方法首先对采集到的扰动信号进行小波多分辨率分解,扰动信号在每个小波分解尺度的能量分布构成一个特征向量;然后利用CART决策树算法从这些特征向量构成的训练样本中自动提取相应的分类规则,得到决策树分类模型,并将该模型应用到电能质量扰动测试数据中.仿真结果表明所提电能质量扰动数据分类挖掘方法的有效性和鲁棒性.
A new approach for power quality disturbances (PQDs) classification was proposed, which was based on wavelet transform and decision tree classification algorithm. By wavelet multi-resolution analysis, the energy distribution of PQDs at different frequency bands was constructed as feature vectors. Then the decision tree was obtained based on the features through inductive inference. And the classification rules extracted from the decision tree were applied to classify various PQDs. Numerical experiments have shown that the proposed approach can provide high accurate classification results,and its performance is fast enough to fit real time PQDs classification.
出处
《湖南大学学报(自然科学版)》
EI
CAS
CSCD
北大核心
2007年第10期57-61,共5页
Journal of Hunan University:Natural Sciences
基金
湖南省自然科学基金资助项目(05JJ40001)
长沙市科研基金项目(K051150-72)
关键词
电能质量
扰动分类
小波变换
决策树
power quality
disturbances classification
wavelet transform
decision tree