2种风电功率预测模型的比较.pdfVIP

第 卷 第 期 23 6 能 源 技 术 经 济 Vol．23 No． 6 第 期 6 年 月 智 能 电 网 2011 6 Energy Technology and Economics Jun. 2011 文章编号：1674-8441(2011)06-0031-05 2 种风电功率预测模型的比较 ， 1 2 2 3 时庆华 ， 高 山 ， 陈 昊 (1. 山东省电力公司日照莒县供电公司，山东日照 276500 ；2. 东南大学，江苏南京 210096 ；3. 南京供电公司，江苏南京 210008） 摘 要：采用AR MA 模型对风电功率进行了预测，并由AR MA 方程推导出卡尔曼滤波状态方程和测 量方程，从而将预测问题转化到状态空间，并利用卡尔曼滤波法预测了风电功率，比较了2 种方法的 预测效果。实例表明，卡尔曼滤波法能够提高风电功率的预测精度，并在一定程度上解决了时间序列 分析法的预测时延问题，对电力系统的安全、稳定、经济运行以及提高运行效益具有重要意义。 关键词：风力发电；功率预测；AR MA；卡尔曼滤波 中图分类号：F272 ；TK01+8 文献标志码：A Comparison Study on Two Wind Power Forecasting Models 1 2 2,3 SHI Qinghua , GAO Shan , CHEN Hao (1. Rizhao Juxian Power Supply Company, Shandong Electric Power Corporation, Rizhao 276500, China; 2. Southeast University, Nanjing 210096, China; 3. Jiangsu Nanjing Power Supply Company, Nanjing 210008, China) Abstract: Wind power forecasting is very important to wind farm planning and stable operation of the power system. In this paper, wind power is firstly estimated by means of the ARMA (Autoregressive Moving Average) model as built, and then the Kalman filter state equation and measurement equation are derived from the ARMA model to transform the forecast into state space, and finally wind power is forecasted with the Kalman filter method. A comparison on the forecasting performance between the ARMA model and the Kalman filter method suggests that the Kalman filter method effectively inc