电器与能效管理技术 ›› 2022, Vol. 0 ›› Issue (8): 23-32.doi: 10.16628/j.cnki.2095-8188.2022.08.004

• 研究与分析 • 上一篇    下一篇

基于因果关系分析的短期负荷预测方法研究*

张光儒1, 马振祺1, 杨军亭1, 张家午1, 苏娟2, 高天2,3, 丁泽琦2, 方舒2   

  1. 1. 国网甘肃省电力公司 电力科学研究院, 甘肃 兰州 730070
    2. 中国农业大学 信息与电气工程学院, 北京 100083
    3. 国网山东省电力公司 泰安供电公司, 山东 泰安 271000
  • 收稿日期:2022-02-27 出版日期:2022-08-30 发布日期:2022-10-11
  • 作者简介:张光儒(1989—),男,高级工程师,主要从事主动配电网及智能配电网技术研究。|马振祺(1980—),男,高级工程师,主要从事配电网运维管理研究工作。|杨军亭(1986—),男,高级工程师,主要从事配网运检和高电压技术研究工作。
  • 基金资助:
    *大学城绿色低碳多能互补综合能源系统关键技术研究与示范(52090020002X)

Research on Forecasting Methods of Short-Term Load Based on Causal Relationship Analysis

ZHANG Guangru1, MA Zhenqi1, YANG Junting1, ZHANG Jiawu1, SU Juan2, GAO Tian2,3, DING Zeqi2, FANG Shu2   

  1. 1. State Grid Gansu Electric Power Company Electric Power Research Institute,Lanzhou 730070,China
    2. College of Information and Electrical Engineering,China Agricultural University,Beijing 100083,China
    3. Tai’an Power Supply Company,State Grid Shandong Electric Power Company,Tai’an 271000,China
  • Received:2022-02-27 Online:2022-08-30 Published:2022-10-11

摘要:

随着高比例可再生能源和电力市场的快速发展,电力系统不确定性增大。为提高市场环境下负荷预测精度,提出一种基于因果关系分析的短期负荷预测方法。首先,采用灰色关联度分析法量化气象因素和市场因素与负荷的相关性;然后,采用最优模态分解法对负荷模态分解,利用Granger因果分析法将影响因素与模态子序列进行匹配;最后,对子序列分别采用差分自回归移动平均(ARIMA)模型和双向长短时记忆(Bi-LSTM)神经网络模型进行预测,将预测结果叠加得到短期负荷预测结果。仿真结果表明,所提方法的预测精度可达到92%,验证了方法的准确性和有效性。

关键词: 短期负荷预测, 最优模态分解, Granger因果分析, ARIMA, Bi-LSTM

Abstract:

With the high proportion of renewable energy and the rapid development of the electricity market,the uncertainty of the power system increases.In order to improve the accuracy of load forecasting in the market environment,a short-term load forecasting method based on causal relationship analysis is proposed.First,the grey correlation analysis method is used to quantify the correlation between meteorological factors and market factors and the load.Then,the optimal modal decomposition method is used to decompose the load mode,and the Granger causal analysis method is used to match the influencing factors with the modal subsequences.Finally,the subsequence is predicted using the differential autoregressive moving average (ARIMA) model and the bidirectional long-short-term memory (Bi-LSTM) neural network model.The prediction results are superimposed to obtain the short-term load prediction results.The simulation results show that the prediction accuracy of the proposed method can reach 92%,which can verifie the accuracy and effectiveness of the proposed method.

Key words: short-term load forecasting, optimal modal decomposition, Granger causal analysis, ARIMA, Bi-LSTM

中图分类号: