Research On Fault Diagnosis And Self-healing Control Method Of Smart Distribution Network

Being the latest trend of energy industry revolution in today’s world, Smart Grid shows the development of society and the developing direction of new grid technique. The intelligentialization of distribution network is an essential part of unified strong and smart grid and self-healing is the main approach towards intelligentialization. Therefore, researches on fault diagnosis and self-healing capabilities of the intelligent distribution network carry significant practical values.Currently, although various intelligent algorithms have been introduced to fault diagnosis of distribution system, its application is still restricted. With respect to self-healing control, situations such as long coverage time and low power quality happen due to the complication of actual grid topology and the large volume of the protective information. To deal with the deficiency described above, three approaches are proposed in this thesis, i. e. fault diagnosis approach based on FMCN, flow calculation methods with distributed power, and MAS-PSO algorithm. The switching with the MAS-PSO algorithm is reconstructed based on accuract fault location to enhance the reliability and robustness of the intelligent distribution network, reduce the time needed by switch reconstruction, and meet high level of self-healing.Firstly, distribution network topology, types of grid failure, power distribution protection, the smart distribution network automation terminal equipment and distribution automation platform are introduced and the relationship between different parts of the Smart Grid is elaborated to back up the following discussion.Secondly, fuzzy neural network (FMNN) is composed through the combination between artificial neural networks and fuzzy theory, which further leads to the universal fuzzy min-max neural network classifier (GFMN) and the fuzzy minimum-maximum neural network classifier (FMCN). Simulation results show that FMCN works more precisely than GFMN in intelligent distribution network fault diagnosis.Finally, the calculation of power flow of the network with distributed generation is studied to determine the load of each node. The reconstruction method of distribution network base on (MAS-PSO) is put forward on step further. Both practical data and simulation experiments verified this method’s advantage in converging to the global optimal solution faster with more precision. The self-healing is accomplished and the level of self-healing is increased.