Research Of Distribution Network Fault Fast Isolation Based On Multi-terminal

With the development of economy,the demand for power supply reliability and power quality is further improved.As the link between the transmission network and the users,the stable and economical operation of the distribution network has become the focus of the power sector.The fault rate of medium voltage distribution network accounts for more than 70% of the whole power grid,and the fault treatment of medium voltage distribution network plays an important role in improving the index of power supply.In this paper,we study the algorithm of distribution network fault,focus on analyzing and improving distributed feeder automation and genetic algorithms to distribution network reconfiguration.The main contents of this paper are as follows:To improve the real-time topology identification method of distributed feeder automation,and to improve the efficiency of multi terminal identification by replacing the relay query by broadcasting.The terminals use distributed matrix algorithm to generate operation command according to the positioning result.Using the tentative closing operation to the rural and forest power network with poor communication,this operation make the system has ability of completing the operation of fault isolation correctly with communication blocking and packet loss,it also can eliminate the transient fault closing attempt to produce.In order to solve the problem of low voltage and heavy load after the fault isolation,we use the genetic algorithm to reconfigure the distribution network.In order to solve the problem of dynamic change of distribution network loop after fault isolation,we use the breadth first search algorithm to identify the loop of distribution network.According to the identification results,we encode the solutions by loop structure,which can increase the proportion of feasible solutions in the encoding space,reduce the cost for feasibility test of genetic operation,the infeasible solutions bring the operation.To solve the repeat calculation problem in the late of algorithm,we use taboo list to store all the historical solutions and the model calculation results,which can avoid repeated calculation of model.When the algorithm got a new solution,we insert it to the taboo list by order,so the taboo list will be in the solution sequence,and it will improve the search efficiency of the taboo list.To improving the decisions of non-dominated solution,we make the decisions by the degree of the annual indicators of the distribution network,and it can improve the optimization ability of the actual indicators.Simulation has been carried out for the algorithms in this thesis.Simulation results show the effectiveness of the proposed algorithms.And these algorithms have been implemented in the terminal,which shows the feasibility of the algorithm running in the terminal.