摘要
风电场准确的风速预测可以减轻或避免风电对电网的不利影响,有利于在开放的电力市场环境下正确制定电能交换计划,提高风电竞争力。基于风速序列的时序性,使用极大似然法对风速序列进行了Box-Cox最优变换,建立了ARMA(p,q)风速预测模型。为检验时间序列模型的有效性,利用最小信息准则中的BIC(Bayesian Information Criterion)函数对ARMA(p,q)模型进行识别,并通过风速频率曲线对预测结果进行了修正。仿真结果和算例验证了该方法在风电场风速预测中的适用性,具有一定的实用价值。
Accurate wind speed forecasting of wind power plants can relieve or avoid the disadvantageous impact on the electric network, and it is useful to make electric energy exchange plans and enhance competitive ability of wind power plants in electricity markets. Based on the wind speed sequences, the maximum likelihood method is used to confirm optimal Box-Cox transformation, and the ARMA （p, q） model is established. In order to check the validity of timeseries model, BIC （Bayesian Information Criterion） function as one of minimum information criterion methods is used for model identification. The forecasting result is amended according to wind frequency curve. The result of simulation verifies that the method is suitable for wind speed forecasting.
出处
《现代电力》
2008年第4期35-39,共5页
Modern Electric Power
