Study on the Spatial and Temporal Distribution Characteristics of Large-Scale Wind-PV Base by Using Two-Layer Clustering Approach

As the installed capacity of large-scale renewable energy power base including wind farms and PV generations increases, the stochasticity, intermittent and fluctuating characteristics of new energy output have a violent impact on grid operation, planning and scheduling. In order to solve these problems, it is necessary to extract typical output scenario through clustering methods from a large amount of output data, which is used to portray the characteristics of renewable energy power bases output. Two steps clustering strategy is proposed in the paper for the purpose of withdrawing the typical scenario and further improving accuracy of classification combined with Euclidean distance and shape-based distance. First, the similarity criterion is adopted through shape-based distance, which identifies the trend difference among all output. Second, the proximity is further evaluated through Euclidean distance, which reflects the magnitude changes the time sequence. Based on typical scenarios, complementarity index and correlation index are defined to analyze the characteristics of the base. Eventually, combining with the historical output data of Zhangbei region, typical scenarios are extracted to verify the effectiveness of the algorithm used in this paper. Therefore, it is used to analyze the wind-PV base output characteristics and conclude that wind-PV base in Zhangbei region has obvious complementarity.