以预测区间边界为约束的风电功率场景预测方法

A Wind Power Scenario Prediction Method with Prediction Interval Boundary as Constraints

针对现有不确定性预测方法存在保守性大而造成对系统调节能力要求较高的问题,提出一种区间估计和场景分析相结合的风电功率预测方法,首先通过区间估计得到以给定置信度包含未来风电功率真实值的短时波动区间,然后筛选出历史日中各时段落在区间包络线内的风电场景样本,再对筛选出的样本基于均值聚类算法进行缩减从而得到各时段的预测场景。最后通过风电预测算例和风火联合优化算例,证明了方法的可行性和有效性。

To address the problem that the existing uncertainty prediction method has high conservatism which leads to a high demand for the system adjustment ability,this paper presents a method of wind power forecasting with combination of interval estimation and scenario analysis.First,the short-time fluctuation of the wind power true value is obtained by interval prediction.Second,the wind power scenario samples within the interval envelope of each time are selected,and the selected samples are reduced based on the mean clustering algorithm to get the prediction scenarios in each time period.Finally,through simulation examples,the feasibility and effectiveness of the method are proved.