摘要
针对目前基于数字信号处理的防孤岛检测手段,提出一种小波包变换(Wavelet Package Transform, WPT)与长短时记忆(Long Short-Term Memory, LSTM)神经网络相结合的孤岛检测方法。该方法首先采集公共耦合点处不同工况下的电压波形,再将其分别进行4层小波分解,重构波形将1-7次低频和40次以上谐波成分滤除。最后将重构信号作为LSTM神经网络的输入对孤岛情况进行分类识别,完成孤岛检测。实验结果表明,所提方法网络学习效率高,检测准确率高。
Anti-island detection devices must work rapidly and reliably due to the reason that unintentional island operation could disturb the stable state of power system.A new island detection method is proposed based on Wavelet Package Transform(WPT) and Long Short-Term Memory Neural Network(LSTM) according to the current island detection method which using digital signal processing.First, the voltage signals from common coupling point are sampled.Then they were decomposed by 4-level wavelet package transform, 1-7 order and over 40 order harmonic is filtered by WPT reconstruction.Finally, the LSTM Neural network is used to classify and detect the island operation condition. the simulation shows that this detection method has the advantages of fast learning speed, short detection time and high detection accuracy.
作者
黄文聪
王增雯
常雨芳
王明辉
HUANG Wencong;WANG Zengwen;CHANG Yufang;WANG Minghui(School of Electrical and Electronic Engineering,Hubei Univ.of Tech.,Wuhan 430068,China;Jiangsu Goldwind Software Technology Co.,Ltd,Wuxi 214000,China)
出处
《湖北工业大学学报》
2022年第5期1-5,22,共6页
Journal of Hubei University of Technology
基金
国家自然科学基金(61903129)。
关键词
防孤岛检测
被动检测法
小波包变换
LSTM神经网络
分布式光伏电站
unintentional island detection
passive detection
wavelet transform
LSTM neural network
distributed photovoltaic generator