A Distribution Network Fault Branch Identification And Location Method Based On Distributed Multi-electric Quantity Information

Customer Service(CS)is a critical task in Distribution Network(NT).Unfortunately,for these reasons,such as the complexity of network topology,the diversity of load types and small distance in environment between aerial conductors and human activities,the Distribution Network fails frequently.Usually due to a multitude of single-phase ground fault in all the types of failure,the fault produces low current in grounding modes of distribution network in China and has a small effect on normal operation of the system.However,once it becomes the phase-to-phase fault or arbitrary two points grounding fault arising from the puncture in over-voltage suppression of intermittent arc,the fault give rise to a multitude of negative effects on operation of the system.At present,most applications to locate faults have focused on centralization,which is imperfect with slow processing speed and a higher probability of on transformation for fault types.It is an essential ability with detecting quickly and accurately the position of failures.This paper presents a scheme for faults location based on the method of multi-source information in Distribution Network that offers the rules for isolation.It consists of four parts.Firstly,analying the topology of the multi-branch distribution network and compare the advantages and disadvantages of the two different modes based on the feeder automation modles.Undoubtedly,for the neutral grounding and neutral point operation modes,there are a lot more discussions as follow: the unite start-up feeder,the interactive time in FTU between collects and delivers,communication technology and so on.Secondly,the aim of this paper is that presents a methodology based on distributed automation location-based service for locating the failures in monitoring the local and actual information.Furthermore,the proposals for selecting the best denosing wavelet algorithm,rather than several normal kinds of adaptive denosing algorithms and extraction of pure failt components are introduced.Thirdly,this section highlights the upper and lower distribution and features in transient state component of fault points.In consideration of ungrounded-neutral mode,simulations of various faults under different conditions and positions are performed by using the topological distribution network mode on PSCAD.The results show the method is feasible and effective.Fourthly,according to the transitionin neural-point grounding method,fault section is obtained through the power frequency components between front and back in same point.The performance of the proposed methodology was also proved through simulation on the topology model with cable hybrid via software PSCAD.These results are very encouraging,indicating that the methodology is applicable and robust.