Simulation Study of Dynamic Reactive Power Optimization in Distribution Network with DG Based on Improved Lion Swarm Algorithm

This paper proposes a spectral clustering method in consideration of the operational issues of distribution networks like load fluctuation, intermittent power output, reactive power flow, and daily switching frequency of reactive power compensation. We divide the daily load curve of the distribution network with distributed generation units (DG) into time periods, and set the minimum network loss and voltage offset of each time period as the objective function. Then we use this method to establish a time-divided dynamic reactive power optimization (RPO) mathematical model of DG. Since the traditional random lion swarm optimization (LSO) can hardly escape a local optimum, a random black hole mechanism is introduced to improve the LSO algorithm, and to formulate a random black hole based lion swarm optimization (RBH-LSO) algorithm. This paper takes the improved IEEE 33-node system as the sample object. The RBH-LSO algorithm, the LSO algorithm and the particle swarm optimization (PSO) algorithm are mutually used to realize the optimization of this system. After the simulation results of the optimization are analyzed, this paper demonstrates, as a summary, that the RBH-LSO algorithm has exceeding excellence in performance and proves to be an effective mechanism for dynamic RPO of distribution networks with DG.