• CSCD核心库收录期刊
  • 中文核心期刊
  • 中国科技核心期刊

电力建设 ›› 2018, Vol. 39 ›› Issue (11): 42-50.doi: 10.3969/j.issn.1000-7229.2018.11.006

• 现代人工智能在电力系统中的应用 栏目主持 文福拴教授、赵俊华教授、颜拥博士 • 上一篇    下一篇

基于深度森林算法的电力系统短期负荷预测

陈吕鹏1,2, 殷林飞3, 余涛1,2, 王克英1,2   

  1. 1.华南理工大学电力学院,广州市 510640;2.广东省绿色能源技术重点实验室,广州市 510640;3.广西大学电气工程学院,南宁市 530004
  • 出版日期:2018-11-01
  • 作者简介:陈吕鹏(1995),男,硕士研究生,主要研究方向为电力系统负荷预测及负荷特性分析; 殷林飞(1990),男,博士,助理教授,通信作者,主要研究方向为电力系统智能发电控制技术和人工智能技术; 余涛(1974),男,博士,教授,主要研究方向为电力系统的非线性控制理论和仿真、智能控制算法; 王克英(1963),男,博士,教授,主要研究方向为电力系统保护控制与运行优化。
  • 基金资助:
    国家自然科学基金项目(51777078,51477055)

Short-term Power Load Forecasting Based on Deep Forest Algorithm

CHEN Lüpeng1,2, YIN Linfei3, YU Tao1,2, WANG Keying1,2   

  1. 1. College of Electric Power Engineering, South China University of Technology, Guangzhou  510640, China;2. Guangdong Key Laboratory of Clean Energy Technology, Guangzhou 510640, China;3. College of Electrical Engineering, Guangxi University,  Nanning 530004, China
  • Online:2018-11-01
  • Supported by:
    This work is supported by National Natural Science Foundation of China(No. 51777078,No. 51477055).

摘要: 为了提高电力系统短期负荷预测的精确度,解决目前基于机器学习算法的负荷预测需要人为凭经验对超参数进行大量设置和调整的问题,该文将深度森林算法引入了电力系统短期负荷预测领域。深度森林算法包含多粒度扫描阶段和级联森林阶段,具有表征学习的能力。与深度神经网络相比,深度森林算法能够进行高效并行训练,无须大量人为设置和调整超参数。该文选取了某地区实际电力负荷值以及气象因素数据,分别利用了前21天和前40天的数据对深度森林算法进行训练,并将其负荷预测结果与智能算法和传统分类算法的负荷预测结果进行了对比分析。试验结果表明深度森林算法具有高效的电力系统短期负荷预测的能力。

关键词: 深度森林, 短期负荷预测, 多粒度扫描, 级联森林, 超参数配置

Abstract: Conventional methods and the famous machine learning algorithms for short-term load forecasting have two shortcomings: (i) forecasting accuracy is low;(ii) experiences for the configuration of model hyper-parameters are needed. To mitigate the influence of these shortcomings, deep forest algorithm is applied to short-term load forecasting in power system. Deep forest algorithm, which can do representation learning, includes two procedures: multi-grained scanning procedure and cascading forest procedure. Compared with deep neural network, deep forest algorithm can be trained efficiently in parallel with the default settings for the hyper-parameters of deep forest. The data of systemic actual load and meteorological information are utilized to training the model of deep forest for short-term load forecasting. Two models of the forecasting are built in this paper, i.e., models with the data of previous 21-day and previous 40-day. The forecasting performances of deep forest algorithm are compared with that of numerous intelligent algorithms and conventional classification algorithms, and the results show that deep forest algorithm can forecast the short-term load effectively.

Key words: deep forest algorithm, short-term load forecasting, multi-grained scanning, cascading forest, configuration of hyper-parameters

中图分类号: