期刊文献+

基于双储备池回声状态网的电力负荷预测 被引量:5

Research on Power Load Forecasting Based on Echo State Network with Double-reservoir
原文传递
导出
摘要 精确的电力负荷预测不仅在发电侧优化能源产能,而且在供给侧实现经济调度、绿色用电。调度工作能够根据预测负荷峰值科学有效的实现电网安全可靠运行。首先,提出了一种包含两个储备池的新型回声状态网预测方法,来预测每日电力负荷;其次,在传统的回声状态网的基础上,选取两个储备池进行串行连接,得到一种新型的回声状态网(Echo State Network with Double Reservoir,called DR-ESN)。DR-ESN能够更加有效的提取预测对象的特征信息,从而可以得到精度更高的预测结果;并利用批量梯度和岭回归算法来优化训练过程中的DR-ESN的6个参数。对广州市的实际用电量进行仿真,所得结果表明了预测方法的有效性。 Accurate power load forecasting not only optimizes energy capacity in power generation side,but also realizes economic operation and efficient use of electricity in the supply side.The dispatching work can realize the safe and reliable operation of the power grid scientifically and effectively according to the predicted load peak.Firstly,a new type of echo state network containing two reservoirs is proposed to predict the daily power load.Based on traditional echo state network,a new type of echo state network(echo state network with double reservoir,called DR-ESN)is gained by connecting two reservoirs in series.DR-ESN can extract characteristics of the object more effectively,which can get higher precision of prediction results.Batch gradient and ridge regression algorithm are used to optimize the six parameters in the training process.A simulation,using the actual power load data of Guangzhou,is given to demonstrate the validity of the forecasting method.
作者 庄仲 伍铭妍 刘冲 ZHUANG Zhong;WU Ming-yan;LIU Chong(China Southern Power Grid Guangzhou Power Supply Bureau,Guangzhou 510000,China;College of Information Science and Engineering,Northeastern University,Shenyang 110819,China)
出处 《控制工程》 CSCD 北大核心 2020年第6期1032-1036,共5页 Control Engineering of China
关键词 电力负荷预测 回声状态网 双储备池 梯度下降法 Power load forecasting echo state network double reservoir gradient descent
  • 相关文献

参考文献2

二级参考文献28

共引文献18

同被引文献64

引证文献5

二级引证文献12

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部