Cluster Division in Wind Farm Based on Ensemble Modeling of Machine Learning

In order to improve the equivalence accuracy of wind farms and widen the applicability of multiple operating conditions, a method of using Blending to merge extreme gradient boosting (XGBoost) for clustering index dimensiona-lity reduction and dynamic time warping (DTW) optimization is proposed. Density-based spatial clustering of applications with noise (DBSCAN) is a method of clustering clusters and clustering results fusion. It is used to process wind turbines' multi-dimensional time-series feature operation data to obtain accurate and effective wind farm plans for the division of aircraft clusters on site. Firstly, we leverage XGBoost-Blending to select clustering indicators for dimensionality reduction; Secondly, we propose a clustering method based on DBSCAN-DTW for cluster division and perform ensemble clustering; Finally, we employ Matlab / Simulink to build a simulation model. Thus, a three-phase short-circuit at the grid-connection point of a wind farm is constructed, and a case study is performed under a variety of wind speed scenarios. The results verify the accuracy and wide applicability of the equivalence model after implementing this method.