The Research Of Distribution Network Fault Location Method Based On Different Information Source

Distribution network fault location means fast fault location with full use of real-time fault information, when short circuit or wire break occurs in distribution network. The method of fault location directly affects interruption outage duration of power supply users, and to the society, economy, life etc. Therefore, reliability and accuracy for fault location are both needed, which is the precondition for fault elimination and power restoration in distribution network. Thus, in order to improve the rapidity and tolerance of fault location, this paper separately carried on the thorough research for non-supervise area and supervise area of distribution network,according to the difference of fault information source.For the area where user's electric energy data data cannot be collected in non-supervise area, this paper proposes fuzzy fusion of telephone complaints information and a small amount of equipment alarm information to locate the fault position. We firstly define the membership function based on hierarchical model of distribution network, then improve the fault weight of the branch lines corresponding to alarm devices according to the equipment alarm information, thereby reduce the uncertainty of telephone complaints information and improve the fault tolerance performance of fault location.Then, for the area where user's electric energy data data can be collected in non-supervise area, the user of electricity information is used as the main fault information, supplemented by the user telephone complaint information equipment alarm information to realize fault location in this paper. Based on rough set theory, we firstly establish the fault decision table with this three kinds of information as the condition attributes of fault classification, then calculate the minimal reduction and match the fault, thus we can get the fault section. Owing to the priority of user electric energy data acquisition and the use of telephone complaints as a supplement, it makes the fault location number of telephone complaints been severely depleted, and to further shorten the time for fault location.Besides, for the supervise area of distribution network where feeder automation had been achieved, the gravitational search algorithm is applied to fault location in view of its better optimization effect in discrete domain. This paper establishes uniform function model, and uses gravitational search algorithm to sample and calculate the model, then gradually increases the probability of optimal fault section vector in the solution space, finally we get the optimal fault section vector. The simulation results verify the validity of the proposed fault location method, and And show that it overcomes the disadvantage of common search optimization algorithm in the easiness of falling into local optimum and slow convergence speed and so on.