摘要
电能质量扰动识别过程中噪声的存在会增加误判,为了提高分类的正确率,消噪对于电能质量扰动识别是一项非常重要的工作。论文应用Daubechies小波时频分解的噪声能量保持特性来估计扰动信号中不同分解尺度上的噪声能量,从而由含噪声信号能量分布和所估计的噪声能量确定实际扰动信号的能量特征,完成了消噪,对消噪处理后电能质量扰动信号应用数据挖掘中的决策树算法进行识别。仿真计算表明,该消噪方法能提高识别精度,是一种非常有效的电能质量扰动信号消噪方法。
In the process of power quality disturbances classification, the presence of noise may result in increased false classification rate, denoising is an extremely important work in order to enhance classified accuracy. The noise energy preserving property of Danbechies wavelet across the time-frequency scales was used to estimate the noise energy at different resolution levels in this paper, so the energy features of signals can be determined by that of signals with noise and estimated noise. The efficiency and validity of the denoising method in classification of power quality disturbances was verified by classification algorithm of decision tree algorithm of data mining.
出处
《华北电力大学学报（自然科学版）》
CAS
北大核心
2008年第4期87-92,共6页
Journal of North China Electric Power University：Natural Science Edition
关键词
电能质量扰动
小波变换
噪声能量估计
特征提取
决策树
power quality disturbance
wavelet transform
energy features estimation
features extraction
decision tree
