现代电力 ›› 2020, Vol. 37 ›› Issue (4): 351-357.doi: 10.19725/j.cnki.1007-2322.2019.0929

• 新能源电力系统 • 上一篇    下一篇

基于DPK-means和ELM的日前光伏发电功率预测

李雯, 魏斌, 韩肖清, 郭玲娟   

  1. 电力系统运行与控制山西省重点实验室(太原理工大学),山西省 太原市 030024
  • 收稿日期:2019-12-20 出版日期:2020-08-10 发布日期:2020-08-07
  • 作者简介:李雯(1993),女,硕士研究生,研究方向为新能源发电功率预测,E-mail:cocojaliwen@163.com;魏斌(1989),男,博士研究生,研究方向为分布式发电及微电网优化控制与能量管理,E-mail:weibin279@sina.cn;韩肖清(1964),女,教授,博士生导师,通信作者,研究方向为电力系统运行与控制、微电网及新能源技术,E-mail:hanxiaoqing@tyut.edu.cn;郭玲娟(1995),女,硕士研究生,研究方向为微电网储能配置,E-mail:1043718129@qq.com
  • 基金资助:
    国家重点研发计划项目(2018YFB0904700);山西省科技重大专项(20181102028)

Day-ahead Photovoltaic Power Generation Forecasting Based on DPK-means and ELM

LI Wen, WEI Bin, HAN Xiaoqing, GUO Lingjuan   

  1. Shanxi Key Laboratory of Power System Operation and Control (Taiyuan Universityof Technology), Taiyuan 030024, Shanxi Province, China
  • Received:2019-12-20 Online:2020-08-10 Published:2020-08-07
  • Supported by:
    Project supported by the National Key Research and Development Program of China (2018YFB0904700), Science and Technology Major Project of Shanxi Province (20181102028)

摘要: 日前光伏发电功率预测是电网经济调度的重要依据。针对K均值(K-means)聚类算法初始聚类中心和聚类数目不易确定的问题和传统神经网络训练参数较多、易陷入局部最优等缺陷,构建了DPK-means和极限学习机(extreme learning machine,ELM)的组合预测算法实现日前光伏发电功率的预测模型。首先,采用密度峰值法(density peaks clustering,DPC)对K-means聚类进行优化,解决了K-means算法初始聚类中心和聚类数目不易确定的问题。然后,在利用DPK-means算法对历史气象数据样本聚类分析的基础上,建立ELM预测模型实现日前光伏发电功率的预测。经实测数据验证可知,所提出的组合预测算法可得到较好的预测结果,具有较强的实用性。

关键词: 光伏发电功率, 日前预测, K-means聚类, 密度峰值法, 极限学习机

Abstract: The day-ahead photovoltaic (PV) power generation forecasting was an important evidence in power grid economic scheduling, and the K-means clustering algorithm and neural network were widely used in the power generation forecasting. In allusion to the defects in K-means clustering algorithm, i.e., not easy to determine the initial clustering center and the number of clusters, and the imperfection of neural network, i.e., too many training parameters and easy to falling into local optimum, an algorithm, in which the DPK-means was combined with extreme learning machine (ELM), was constructed to implement the day-ahead forecasting of PV power generation. Firstly, utilizing density peaks clustering (DPC) abovementioned defects in K-means clustering was revised. Secondly, the DPK-means algorithm was used to carry out the clustering analysis on historical meteorological data samples, on this basis an ELM forecasting model was established to implement the forecasting of day-ahead PV power generation. Case study results show that using the proposed combined forecasting algorithm a better forecasting result can be obtained, so the proposed algorithm is feasible.

Key words: photovoltaic power, day-ahead forecast, K-means clustering, density peaks clustering, extreme learning machine

中图分类号: 

  • TM7