Loss minimization in distribution networks by using particle swarm optimization

In distribution networks, each distribution feeder consist residental, commercial and industrial loads and changes of these loads are different from each others. Daily changes of the load, result with peak loads on main transformator and feeder and this case cause problem on reliable and quality power flow of distribution networks. To relieve feeders and make the network operate reliable and stable conditions, distribution network reconfiguration is an effective method. Network reconfiguration provides the transfer of the loads from heavily loaded feeder to lightly loaded feeder and as well it is also effective on voltage profile improvement and power loss reduction. Transfer of the loads between feeders is provided by changing of the open/closed cases of the switches in the distribution network. When we think n is the number of the lines in the distribution network and there is one switch for each line, 2n switching cases must be considered to find a new optimal network topology which provides minimum power loss and maximum voltage magnitudes. In this thesis, particle swarm optimization algorithm is used for the distribution network reconfiguration with the objectives of power losses minimization and voltage profile improvement. Particle swarm optimization algorithm is a swarm-based solution method developed by Kennedy and Eberhart. Algorithm is inspired by social behaviours of bird and fish. Unlike genetic algorithm, particle swarm optimization algorithm has no evolution operations. Advantages of the algorithm are easy implementation to the optimization problems and algorithm has few parameters to adjust. Good choice of the parameters has important impact on the performance of the particle swarm optimization algorithm. Because of the network reconfiguration problem is a complex combinational optimization problem, the particle swarm optimization algorithm is used for determination of the optimal switching combination that provides objective functions of the network reconfiguration problem. By choosing of the proper switching, network topology has changed in comprasion with the base case and a new network topology which gave the minimum power loss and the best voltage profile has found. Proposed algorithm in this thesis is tried on two frequently used test networks with three different load levels (nominal, light and heavy) and results are compared with the other methods in literature. In comprasion, it is seen that the particle swarm optimization algorithm choosed the proper switching combination and gave minimum loss and the best voltage profile. Voltage stability index is used for the study to determine critical buses in the network before and after the reconfiguration thus voltage stability of the new network topology is examined. Voltage stability indexes are compared the voltage profiles and it is seen that indexes are successful to determine voltage stability. Distributed generation is a popular issue in the literature nowadays. As the growing demand on energy with the developing technology. And as conventional energy sources are exhaustible, new energy resources are gain importance. The new energy resorces can be given as wind, photovoltaic, fuel cell power plants and they are also enviromentally friendly resources. Solution of the reconfiguration problem is tackled again to examine the effect of distributed generation systems are located on feeders with optimal sizes and allocations in distribution networks. And results showed that distributed generation is an effective method to minimize power losses and improve voltage profile of the distribution network. Capacitor placement is a one of the loss minimization technique in the literature. Optimally located and proper sized capacitor banks have benefits for the power systems. To show the effects of the capacitor placement, network with optimally located capacitor banks are considered. And solution of reconfiguration problem is analysed. 33 and 69 bus radial distribution test networks are used for the study and loss minimization techniques are tried on the networks. The active and reactive power supplies connected to the network are studied.