Abstract：In order to further improve the accuracy of fault identification of local abnormal factors in

distribution network，a digital twin based fault identification simulation of local abnormal factors in distribution

network was proposed. Through real-time acquisition of electrical parameters of distribution network operation and

preprocessing，the fault feature matrix based on time series was extracted，and the multidimensional scaling（MDS）method was used to detect the abnormal physical nodes of distribution network from the reduced dimension fault  features. Then，the fault section corresponding to the abnormal physical nodes was obtained according to the distribution network topology. Finally，the local abnormal factor value corresponding to each physical node was calculated with local outlier factor（LOF）algorithm，so as to obtain the fault diagnosis results and complete the accurate identification of distribution network faults. The simulation results show that the proposed method can achieve accurate identification of distribution network faults.

      Key words：digital twins；distribution network；fault identification；local anomaly factor（LOF）；operation

characteristic analysis