Study On The Treatment Technology Of The Single-Phase Fault In The Neutral Resonant Grounded System

China's distribution network uses the neutral resonant grounded system(NES)widely now,When the line in the system occurs single-phase grounding fault,due to the failure of fault steady character,the difficulty of obtaining transient signals,the signals is easy to be affected by external environment,the failure certerion is relatively single,the signal acquisition accurate synchronization is difficult and other issues,the field application effect of line selection and section location is poor.It is badly in need of carry on more research and improvement to the fault line selection and section location.First,the main causes of single-phase grounding fault are summarized from fault history datas;the paper using the transient equivalent circuit to analyse the components of the transient zero-mode current and their influencing factors under different transition resistance,and the simulation model of the distribution network is established by ATP-EMTP to analyse the concrete influence of transition resistance,fault initial phase angle,line material,fault position on the transient zero-mode current;And the paper analyses the distribution of the transient zero-mode current in fault point upstream and downstream.Aim at line selection,based on the idea of making full use of the fault transient charaxteristics,expanding the differences between the healthy line and fault line,the paper proposes the multi-criteria fused synthetical fault line selection method base on k-medoids clustering algorithm.The way uses generalized S-transform(GST)which has excellent ability in time-frequency analysis,to get the time-frequency matrixes which contains rich fault characteristics,from the transient zero-sequence current of each line;Then extracting indirect characteristic quantity which are complementary:singular value comprehensive euclidean distance,energy torque of the characteristic frequency,similarity coefficient of the characteristic frequency;Finally,the k-medoids clustering algorithm is used to fuse the feature quantity to get the selection result.Through verified by the simulation model data,this method can overcome the shortcoming of single fault-detection method,reduce the blind area of line selection and improve accuracy.Then,Aim at section location,according to the characteristics of the transient zero-mode current in the upstream and downstream of the fault point,consided the characteristic of section location,the paper follows the three feature quantities mentioned above and do a little change to make them fit in section location,puts forward a new multi-criteria fused fault section location base on fuzzy synthetic decision,The way extracts a cycle zero-mode current which starts in the fault time and from each testing points,obtains the characteristic quantity in each section,calculate the value of fault membership function of each characteristic quantity,and multiply the comprehensive weighting coefficient to get the value of comprehensive fault membership function,compares with threshold to do section located.Of which,the way adopts GST to demarcate the fault moment to avoid the deternine errors from sampling out of sync.The faults relative membership degrees is decided by the characteristic quantity.And improved weight to the comprehensive weights which is fused by objective weight and subjective weight.The objective weight from the historical data,based the rough set theory.The subjective weight is setting by expert experience.Through the simulation model and the actual data validation,the positioning method is reliable,effective,and has higher accuracy in the section location on the case of fault with less initial phase angle,high transition resistance,discontinuous transient process and other difficult fault condition.On the basis of introducing the distribution automation system with single-phase grounding fault processing function,the paper uses the actual fault data to verify the methods,and expounded the concrete process of verifying the selected line and location method with typical fault cases.It is further proved that the proposed methods have strong reliability and adaptability.