摘要
分别利用数学建模方法中的BP神经网络、灰色预测GM(1,1)、简单移动平均法和自回归滑动平均(ARMA)方法进行风功率预测。首先简单介绍各方法的原理,其次对实测数据进行测试,最后对各方法的优缺点进行分析,从而为风电并网以及电网的合理调度提供科学依据。
With respect to wind power forecast, we used prediction methods in mathematical modeling to predict the output respectively, including BP neural network, grey model GM (1, 1), simple moving average method and ARMA (Auto-Regressive and Moving Average Model) method. Firstly, the principle of each method was introduced concisely and then the results were examined by real-measured data. Finally, we analyzed the advantages and disadvantages of each method, thus provided scientific basis for wind power grid-connection and reasonable scheduling.
出处
《太阳能学报》
EI
CAS
CSCD
北大核心
2015年第5期1081-1087,共7页
Acta Energiae Solaris Sinica
基金
国家自然科学基金(11101120)
关键词
数学建模
风功率
预测
仿真
mathematical modeling
wind power
prediction
simulation