Key node-based Observable and Predictable Technology for Distributed PV Clusters Research

To address the problem of new energy regulation "blind zone" caused by the high-density and dispersed distribution of large-scale distributed photovoltaic power systems connecting to the distribution grid, firstly, the community detection algorithm and tabu search algorithm are used to divide the PV clusters effectively based on the electrical distance and regional voltage regulation ability. Then, the voltage sensitivity matrix calculation is used to determine the key nodes within each cluster to reduce the cost of data collection. Next, the genetic algorithm (GA) and BP neural network are combined to predict the PV power output of the cluster using the data of the key nodes, achieving the goal of observable and predictable. Finally, the validity of the proposed method is verified by MATLAB software simulation.