Abstract：In view of the problems caused by photovoltaic grid connection to the distribution network, such as voltage fluctuations, increased line losses, and uncertainty in photovoltaic and load output, this paper constructs a linear convex optimization model based on second-order cone programming. By controlling the on-load voltage regulating transformer and the capacitor bank action, photovoltaic inverter and static var generator reactive power compensation capacity constraints are dynamically analyzed on the day-ahead and intra-day dual time scale reactive power optimization model, which simplifies the solution process and increases the possibility of finding the global optimum. An improved gray wolf algorithm based on chaotic learning initialization, nonlinear convergence factor, optimal particle Cauchy perturbation and spider monkey algorithm position update method is proposed to prevent falling into local optima and enhance global search capabilities. Finally, the algorithm is used to model and simulate the IEEE 33 node system containing photo-voltaic. The results show that the algorithm has the advantages of high optimization efficiency and fast con-vergence speed. The feasibility and effect of the proposed algorithm are confirmed.
于惠钧, 马凡烁, 陈刚, 杨驰泽, 李嘉轩. 基于改进灰狼优化算法的含光伏配电网动态无功优化[J]. 电气技术, 2024, 25(4): 7-15.
YU Huijun, MA Fanshuo, CHEN Gang, YANG Chize, LI Jiaxuan. Dynamic reactive power optimization of photovoltaic distribution network based on improved gray wolf optimization algorithm. Electrical Engineering, 2024, 25(4): 7-15.
2024, Vol. 25 
