Intelligent Partition Strategy of Distributed Photovoltaic Cluster in Distribution Network Based on SLM-RBF

Abstract

Access of large-scale distributed power supply to the distribution network brings dimensionality disaster to the optimal dispatching of the distribution network. Therefore, it is necessary to cluster the distributed power supply to reduce the difficulty of regulation and control, and a reasonable division of distributed power supply cluster is very important. However, the incomplete real-time measurement data of the distribution network has caused difficulty and low time efficiency in real-time cluster division of the distribution network. Therefore, this paper proposes a distributed power cluster division strategy based on the smart local moving (SLM) algorithm and the radial basis function (RBF) neural network. First, the range of active power and reactive power regulation and the sensitivity of active power and reactive power to voltage are selected as the indexes of cluster division. By constructing a similarity matrix, the SLM algorithm is used to form the historical strategy library of cluster division of distributed power sources. Then, a voltage fitting model is established offline, which can observe the relationship between the power and voltage of buses in real time. Meanwhile, a voltage-division result model is established offline, and the real-time division result is obtained through the voltage online, which solves the problem that cluster division cannot be performed when the power flow model is missing, and improves the real-time performance of cluster division. Finally, the rationality and superiority of the algorithm are verified by simulation on MATLAB platform.

Keywords

• smart local moving (slm) algorithm
• radial basis function (rbf) neural network
• cluster partition
• voltage fitting