基于改进K-means聚类和SBR算法的风电场景缩减方法研究

Wind Power Scenario Reduction Based on Improved K-means Clustering and SBR Algorithm

场景法是适应风电高占比电力系统优化调度的重要方法。作为场景分析方法的研究热点,场景缩减的意义在于用少量代表性场景描述大量复杂性场景特征,达到降低计算复杂度的目的。针对风电出力提出一种基于改进的K-means聚类和同步回代消除算法(simultaneous backward reduction,SBR)相结合的场景缩减方法。首先基于改进的K-means聚类算法对原始场景进行快速分类,其次针对每一类簇中的场景集合采用基于Kantorovich距离的SBR算法进行缩减。该方法可以在保证计算精度的同时,提高规模较大场景集合缩减的计算效率。最后采用我国西北某省网风功率实际数据开展实证分析,通过布莱尔分数(Brier score,BS)指标和风功率波动的高斯混合模型验证了所提场景缩减方法的有效性和优越性。

The scenario method is important in the adaptation of the optimal dispatch of the power system with a high proportion of wind power.As a research hotspot of scenario analysis methods,the significance of scenario reduction is to describe a large number of complex scenario features with a small number of representative scenarios to achieve the purpose of reducing computational complexity.Aiming at the wind power output,a scenario reduction method based on the combination of the improved K-means clustering and the Simultaneous Backward Reduction(SBR)is proposed.Firstly,the original scenarios are quickly classified based on the improved K-means clustering algorithm.Secondly,the SBR algorithm considering Kantorovich distance is used to reduce the scenario sets in each cluster.Finally,an empirical analysis is carried out using the actual data from a certain province in Northwest China.The effectiveness and superiority of the proposed scenario reduction method are verified with the Brier Score(BS)indicator and the Gaussian mixture model of wind power fluctuations.