Reactive-voltage Partitioning Method for Wind Power Grid Based on Interval Uncertainty Clustering

In the context of wind power connected to the grid, the uncertainty of wind power prediction will lead to inaccurate reactive partition based on voltage reactive sensitivity. A reactive-voltage partitioning method for wind power grid based on interval uncertainty clustering is proposed in this paper. Based on the Jacobian matrix, the full-dimensional sensitivity matrix is obtained by successive recursive method. The strength of electrical coupling relationship among all nodes of the system is reflected by establishing a full-dimensional spatial electrical distance matrix. The concept of interval number is introduced to establish the full-dimensional electrical distance matrix based on interval uncertainty, which replaces the electrical distance matrix with a single value. In order to improve the accuracy of the partitioning algorithm and reduce the computational complexity, the UIDK-means clustering algorithm based on interval uncertainty is used for reactive partitioning. The simulation results of the improved IEEE-39 node system show that compared with the traditional partitioning method, the partitioning method based on interval uncertainty can accurately obtain the reactive partitioning result considering wind power prediction uncertainty. It is conducive to improve the accuracy of day-ahead dispatch.