中国电力 ›› 2016, Vol. 49 ›› Issue (5): 157-162.DOI: 10.11930/j.issn.1004-9649.2016.05.157.06

• 新能源 • 上一篇    下一篇

基于典型天气类型计及随机预测误差的光伏发电短期预测研究

陈瑶琪   

  1. 南京师范大学电气与自动化工程学院,江苏南京210042
  • 收稿日期:2015-10-14 出版日期:2016-05-16 发布日期:2016-05-16
  • 作者简介:陈瑶琪(1995—),女,江苏泰州人,从事电气工程研究。E-mail: 408315064@qq.com

Short-term Photovoltaic Power Prediction Based on Typical Climate Types and Stochastic Prediction Error

CHEN Yaoqi   

  1. College of Electrical and Automatic Engineering, Nanjing Normal University, Nanjing 210042, China
  • Received:2015-10-14 Online:2016-05-16 Published:2016-05-16
  • Supported by:
    Keywords: photovoltaic power prediction; typical climate type; relative affective factor; probabilistic modeling; t Location-Scale distribution

摘要: 光伏发电出力与天气类型有直接关系,越是多云或是阴雨天气,光伏出力预测的误差越大。基于典型天气类型的划分,提出了一种计及预测相对误差的考虑“相关影响因子”的光伏出力预测模型。定义了典型天气类型并据此对历史数据进行了划分,提出了光伏出力预测的“相关影响因子”;利用t Location-Scale分布建立了光伏出力预测的相对误差概率模型,采用拉丁超立方技术进行了预测相对误差的样本抽取;将光伏预测的相对误差其与预测值进行叠加得到了最终的预测结果。利用算例证明了所建模型的可行性和有效性。

关键词: 光伏预测, 典型天气类型, 相关影响因子, 概率建模, t Location-Scale分布

Abstract: Photovoltaic(PV) power output is directly related to the climate types and the prediction accuracy of PV power decreases for cloudy and rainy weathers. Based on the categories of typical climate types, a PV power forecast model is proposed with consideration of the predictive relative tolerance and the relative affective factor(RAF). At first, the historical climate data are categorized according to the definition of typical climate types and the RAF is put forward; Then the probabilistic model of predictive relative tolerance is established by using the t Location-Scale distribution and the Latin hypercube technique is used for sampling of the predictive relative tolerances; Finally, the predictive relative tolerance and the predicted value are superimposed to obtain the final prediction results. A case study has proved the feasibility and effectiveness of the model.

中图分类号: