摘要
随着分布式新能源的大量接入,变电站中低压线路中除了确定分类的负荷,还增加了许多不确定的可再生能源功率成分.从经济性和清洁性的角度出发,必要定量分析其中的新能源占比,提升用户的用电质量水平.针对低压馈线处的电能质量信息,提出一种基于改进经验小波能量熵和长短期记忆(Long Short Term Memory,LSTM)神经网络的定量分析算法.该方法以经验小波变换为基础,将长时间序列分为等长的小段时间序列,分别进行经验小波变换;得到各个经验小波函数分量后,计算出各个特征分量的能量熵序列;再引入峭度的概念,对各个能量熵序列加权差异化处理,突出特征信息;最后利用LSTM神经网络进行分类处理,判断不同可再生能源占比的电能质量信息属于哪个区间片段.通过搭建仿真模型模拟现场实际数据,高效准确地定量分析可再生能源占比,并与其他传统方法对比,验证了该方法的有效性和优越性.
With the massive access of distributed new energy sources,many uncertain renewable energy power components have been added to the low-voltage lines in substations,in addition to the loads of definite classificationFrom the perspective of economy and cleanliness,it is necessary to quantitatively analyze the percentage of new energy sources in them and improve the power quality level of usersA quantitative analysis algorithm based on improved empirical wavelet energy entropy and LSTM(Long Short Term Memory)neural network is proposed for the power quality information at the low-voltage feedersThe method is based on the empirical wavelet transform,which divides the long time series into equal length small time series and performs the empirical wavelet transform respectively;after obtaining each empirical wavelet function component,the energy entropy series of each feature component is calculated;then the concept of Kurtosis is introduced,and the energy entropy series are weighted differently to highlight the feature information;finally,the LSTM neural network is used for classification process to determine the energy quality of different renewable energy sourcesFinally,the LSTM neural network is used to classify and determine which interval fragment the power quality information belongs to with different renewable energy ratioBy building a simulation model to simulate the actual data in the field,we can quantitatively analyze the renewable energy share efficiently and accurately and compare with other traditional methods to verify the effectiveness and superiority of the method.
作者
鲍家伟
王青松
BAO Jiawei;WANG Qingsong(School of Electrical Engineering,Southeast University,Nanjing Jiangsu 210096)
出处
《东北电力大学学报》
2022年第5期33-43,共11页
Journal of Northeast Electric Power University
基金
国家自然科学基金面上项目(52177171)。
关键词
可再生能源
定量分析
经验小波变换
改进能量熵
LSTM神经网络
Renewable energy
Quantitative analysis
Empirical wavelet transform
Improved energy entropy
LSTM neural network
