Aggregation modeling of large wind farms using an improved K-means algorithm

Large wind farms require modeling and simulation in their planning and construction process to provide a basis for grid planning. As a large wind farm is comprised of hundreds or even thousands of wind generators, the detailed-model building of wind farms is difficult, which are associated with slow simulation speeds and heavy workloads. In this paper, an improved K-means algorithm-based aggregation modeling method is proposed for large wind farms. A model is built for a typical 300 MW wind farm of the northern Jiangsu Province in China and its simulations are performed. The results based on a comparison between a single-fan aggregation model and a multi-fan aggregation model are obtained. The aggregation method proposed in this paper proves to be accurate and fast, which is able to meet the requirements for output simulation. Key-words: wind farm, power grid, wind generator, aggregation modeling, output simulation, K-means clustering algorithm.