中国电力 ›› 2016, Vol. 49 ›› Issue (6): 176-180.DOI: 10.11930/j.issn.1004-9649.2016.06.176.05

• 新能源 • 上一篇    下一篇

不同时间分辨率的风功率时间序列ARIMA模型预测

张立栋1,李继影2,吴颖3,余侃胜4,朱明亮5,迟俊宇6   

  1. 1. 东北电力大学 能源与动力工程学院,吉林市 132012;
    2. 中广核风电有限公司辽宁分公司,辽宁 沈阳 110000;
    3. 国网江西省电力公司检修分公司,江西 南昌 330096;
    4. 国网江西省电力科学研究院,江西 南昌 330096;
    5. 吉林省东能电力工程有限公司,吉林 长春 130033;
    6. 宾县大个岭风力发电有限公司,黑龙江 哈尔滨 150400
  • 收稿日期:2015-11-26 修回日期:2016-06-16 出版日期:2016-06-16 发布日期:2016-06-16
  • 作者简介:张立栋(1980-),男,吉林市人,副教授,从事风电场建模与仿真及颗粒混合装置研究。E-mail: nedu1015@aliyun.com

ARIMA Model Forecast for Wind Power Time Series with Different Temporal Resolutions

ZHANG Lidong1, LI Jiying2, WU Ying3, YU Kansheng4, ZHU Mingliang5, CHI Junyu6   

  1. 1. Institute of Energy and Power Engineering, Northeast Electric Power University, Jilin 132012, China;
    2. CGN Wind Energy Limited Liaoning Branch, Shenyang 110000, China;
    3. State Grid Jiangxi Maintenance Company, Nanchang 330096, China;
    4. State Grid Electric Power Research Institute of Jiangxi Province, Nanchang 330096, China;
    5. Jilin East Power Engineering Co., Ltd., Changchun 130033, China;
    6. Binxian Dageling Wind Power Generation Co., Ltd., 150400, China
  • Received:2015-11-26 Revised:2016-06-16 Online:2016-06-16 Published:2016-06-16

摘要: 以某风电场同一风力机为研究对象,采用自回归积分滑动平均模型(ARIMA)对5种时间分辨率实际输出功率的时间序列进行预测研究。结果表明:风功率时间序列某些明显的特征点,随着时间分辨率的减小而越来越少直至消失;对预测结果采用平均绝对误差分析,得出随着时间分辨率增大,ARIMA模型预测绝对误差呈现逐渐减小的趋势,1 min的时间分辨率误差最小。

关键词: 时间分辨率, 风电场, ARIMA, 功率预测

Abstract: By taking a wind turbine of a farm as a case, the ARIMA model is used to predict the wind power time series according to five time resolutions. The results show that the number of characteristic points of power time series decreases with the reduction of time resolution until the characteristic point disappears completely. The mean absolute error (MAE) is used to analyze the prediction results, and it is concluded that the absolute error of the ARIMA-predicted results decreases gradually with the increase of time resolution with the MAE of 1min resolution being the minimum.

Key words: time resolution, wind farm, ARIMA, power prediction

中图分类号: