GAT-GRU Based Reactive Power Optimization for Distribution Networks With High Penetration of Distributed New Energy

Reactive power optimization plays an important role in voltage control, power flow distribution and overall stability of the distribution network. At present, the distributed grid connection of highly penetrated new energy and the increasing diversification of load have brought great challenges to the stable operation of the power grid. The timeliness and accuracy of traditional reactive power compensation methods can no longer meet the requirements of low-cost and high-quality power supply under the current complex power grid background. In view of this, in this paper we employ graph attention networks (GAT) and gated cycle unit (GRU) neural network to make optimal decisions on the reactive power of the distribution network. Based on the GAT-GRU network, we capture the correlation characteristics among nodes and acquire the temporal dependency of distribution network properties. The reactive power optimization problem of the distribution network is solved through the cooperation of reactive power regulation equipment and intelligent flexible switch (SOP). Finally, an improved IEEE 33-bus distribution model is used to validate the proposed method. The results indicate that the GAT-GRU network has good effects in voltage control, network loss optimization, and other aspects, thereby confirming the effectiveness and superiority of the proposed method in reactive power optimization.