东北大学学报（自然科学版）

东北大学学报(自然科学版) ›› 2006, Vol. 27 ›› Issue (10): 1083-1086.DOI: -

• 论著 • 上一篇    下一篇

基于动态神经网络的系统边际电价预测

林志玲;高立群;张大鹏;张强;   

  1. 东北大学信息科学与工程学院;东北大学信息科学与工程学院;东北大学信息科学与工程学院;东北大学信息科学与工程学院 辽宁沈阳110004;辽宁沈阳110004;辽宁沈阳110004;辽宁沈阳110004
  • 收稿日期:2013-06-23 修回日期:2013-06-23 出版日期:2006-10-15 发布日期:2013-06-23
  • 通讯作者: Lin, Z.-L.
  • 作者简介:-
  • 基金资助:
    国家自然科学基金资助项目(60374003);;

Forecasting system marginal price of electricity by dynamic neural network

Lin, Zhi-Ling (1); Gao, Li-Qun (1); Zhang, Da-Peng (1); Zhang, Qiang (1)   

  1. (1) School of Information Science and Engineering, Northeastern University, Shenyang 110004, China
  • Received:2013-06-23 Revised:2013-06-23 Online:2006-10-15 Published:2013-06-23
  • Contact: Lin, Z.-L.
  • About author:-
  • Supported by:
    -

RichHTML

1

PDF (PC)

732

摘要： 在分析系统边际电价(SMP)特点的基础上,确定了预测系统边际价格的主要依据为电力负荷、历史上对应时刻的SMP以及当天的SMP趋势.将电价看作是电力市场动态运行的结果,采用动态神经网络预测电价.由于动态神经网络结构及权值确定困难,采用二进制与实数编码相结合的联合编码,用遗传算法优化得到神经网络结构及对应权值.利用某电力市场的历史数据对该模型进行验证,结果表明该方法所建立的预测模型具有较高的预测精度.

关键词: 电力市场, 系统边际电价, 动态神经网络, 遗传算法, 预测, 仿真

Abstract: Analyzing the characteristics of SMP (system marginal price), the electrical load and the historically corresponding and current SMP trends are regarded as the three main influencing factors on forecasting the oncoming SMP value. A recurrent neural network is therefore introduced into forecasting SMP, because it is available to mapping dynamic system and SMP is regarded as a result of dynamic running on power market. To rise above the difficulty of determining NN's structure and weights, the GA optimization algorithm is used to get them by combining binary encoding with real encoding. The historically corresponding market data verified that this method is effective and the forecasting model is accurate.

中图分类号: