Evaluation Method of Wind Turbine Group Classification Based on Calinski Harabasz

At present, the research on clustering of large-scale wind farms is gradually changing from the single-machine equivalence method to the multi-machine equivalence method for modeling, and the Calinski Harabasz principle-based wind farm grouping evaluation method is proposed for the problem of determining the optimal number of equivalents. Firstly, the active power of the system is collected, and the k-means method is applied to cluster, and the optimal number of equivalent machines is determined by comparing the dispersion between groups and the dispersion ratio within groups in various clustering cases. Secondly, the simulation analysis based on Xinjiang power grid confirms that the best equivalent model evaluated by the Calinski Harabasz method is similar to the detailed model results. the wind farm, which shows that the method is correct, and the calculation speed is faster than the conventional method, and it has good universality for complex systems.