采用谱聚类的风电典型出力场景选取方法

为有效地减少原始场景数量,以较小的工作量和较高的精度表征风电出力的全时空特性,本文提出了一种基于谱聚类算法的风电日出力典型场景生成方法。首先采用谱聚类算法对风电日出力场景进行聚类分析,得到能有效反映样本亲疏关系的聚类簇。随后考虑风电的反调峰特性,从聚类后的风电日出力场景中选出能够很好的反映各类场景中原出力曲线调峰效益的出力曲线,将其作为典型场景。最后,以云南某地区2017年风电实际出力数据为算例进行聚类分析,验证所提方法的正确性及有效性。

Generation of Typical Wind Power Scenario Based on Spectral ClusteringMethod

In order to effectively reduce the number of original scenes and characterize the full-time and spatial characteristics of wind power output with less workload and higher accuracy,a typical scene generation method of wind power sunrise based on spectral clustering algorithm is proposed in this paper.Firstly,the spectral clustering algorithm is used to cluster the wind power sunrise scenarios,and the clustering cluster which can effectively reflect the sample sparseness is obtained.Then,considering the inverse peak shaving characteristics of wind power,the output curve which can well reflect the peak shaving efficiency of the original output curve in various scenarios is selected from the clustered wind power sunrise scenarios,and it is taken as a typical scenario.Finally,taking the actual wind power output data of a certain area in Yunnan Province in 2017 as an example,cluster analysis is carried out to verify the correctness and effectiveness of the proposed method.