基于小波变换和神经网络的暂态电能质量扰动自动识别

针对短时电能质量变化和暂态扰动现象的不同特点,提出了一种暂态电能质量分类的新方法。先提取基波频段所在的小波系数将电压凹陷、电压凸起和电压中断分别检测出来;然后将小波包分解结果中的最佳子空间的熵值作为特征量,结合人工神经网络区分暂态脉冲和振荡。该方法利用小波和小波包各自的时频分解特点,实现了暂态电能质量扰动的自动检测和分类。经仿真分析,验证了此方法的准确性和高效性。

Auto recognition of transient power quality disturbances based on wavelet and neural network

Aimed at the different characteristics of power quality disturbance, this paper presents a new method for classification of transient power quality. First, the wavelet coefficients of basic frequency are extracted to identify the short duration disturbance like voltage sag, swell and interruption. Second, entropy features of the best wavelet packet of transient impulse and transient oscillation are calculated. Along with the artificial neural network, they can be effectively classified. Using the different time-frequency characteristics of wavelet and wavelet packet, the proposed approach can avoid the noise and complete the detection and classification of transient pow- er quality. The accuracy and efficiency are verified by simulation analysis.