基于t Location-Scale分布的风电功率概率预测研究

风电功率特有的随机波动性,导致风电功率点预测方法的预测精度不高,增加了风电并网的难度,致使风电场弃风现象严重。基于风电功率点预测的基础上,风电功率概率预测可以预测出风电功率的波动范围,为电力系统的安全运行以及电网调度运行给出不确定信息和可靠性评估依据。提出了一种基于t locationscale分布的风电功率概率预测方法,即采用t location-scale函数来描述风电功率预测误差概率分布,并以此建立误差分布,基于已建立的误差分布可以进行概率预测。并引进了覆盖率和平均带宽来评价预测区间的优劣程度。利用吉林省西部某风电场历史数据验证了该方法的可靠性。

Wind Power Probability Prediction Based on t Location-Scale Distrabution

Wind power forecasting is difficult because of its stochastic variance characteristic. Based on t location-scale distribution function, a wind power forecasting method is proposed. The t location-scale function is adopted to describe probabilistic distribution of wind power prediction error. The prediction is then performed based on the distribution model. The coverage rate and average bandwidth are selected to evaluate prediction accuracy. The historic data of power fluctuation of a wind farm in Jinlin province proves the effectiveness of proposed method.