Optimization of Reactive Power Compensation in Low Voltage Networks Using Particle Swarm Algorithm

The paper explores the optimization of single-phase reactive power compensation (RPC) in low-voltage networks using a particle swarm algorithm. Three optimization options are explored using the particle swarm algorithm - with and without sensitivity coefficients. In the first option, potential nodes for RPC are determined using sensitivity coefficients for reactive power and voltage. In the second, they are determined only using sensitivity factors for reactive power. Next, the nodes and capacities of the required devices are selected from the potential ones using a particle swarm algorithm. In the second option, the selection of connection points for compensation devices and their power is carried out using a particle swarm algorithm. In all cases, the same objective function is used, including costs and voltage deviation. The study of all RPC options was carried out on a model network. All calculations are implemented in the Python programming language. According to the results obtained, the most effective is the third optimization option - using the particle swarm algorithm, without using sensitivity coefficients. With such optimization, it is possible to reduce active power losses by 13.5% of the existing ones, improve the voltage values and zero-sequence asymmetry coefficient at network nodes.