Research On Reliability Improvement Of Distribution Network Fault Line Selection Based On Data And Model Dual Drive

As a vital part of the task of power transmission,the distribution network plays an extremely important role,and it has the closest connection with users.Most of the distribution network in my country adopts the neutral point operation mode in which the neutral point is not directly grounded,that is,the small current grounding mode.When a single-phase grounding fault occurs in a small current grounding system,the small current fault line selection device shall be selected in time.It can quickly isolate faults,which plays an important role in the power supply reliability of the distribution network.However,due to the complex topology of the distribution network,variable operation modes,numerous lines,and difficulty in fault detection,various fault line selection devices currently in use have low line selection accuracy and a single line selection principle,which leads to small current faults in many places.The line selection device lost trust and returned to the old road of pulling the line selection,which caused the technical development of the small current fault line selection device to stagnate,and there was a big gap between theoretical simulation and actual engineering application.Traditional distribution network fault identification and line selection algorithms mostly rely on inherent simulation models.This paper comprehensively considers the historical fault data stored in the small current fault line selection device,and integrates these existing historical fault data.Perform comprehensive analysis,extract useful information from massive data,grasp the overall operation status of the distribution network and real fault characteristics,and dig out the fault characteristic information of the actual operation of the distribution network.First,summarize the characteristics of the fault signal when a single-phase ground fault occurs in a small current grounding system,and analyze the principle of single-phase ground fault from the steady state and the transient state.Combining simulation data to explain,explain the subtle differences between recorded historical fault data and simulation data.Secondly,a large amount of historical fault data is preprocessed.Based on the difference between the physical model of the faulty line and the non-faulty line,the data is cleaned with the idea of post-test simulation to retain valid data and eliminate invalid data.Then analyze the occurrence of single-phase ground faults based on historical fault data,and analyze the distribution of instantaneous and permanent ground faults,high resistance and low resistance ground faults,small fault angles,and large fault angle ground faults in the historical fault data.Analyze fault characteristics from different dimensions,and recognize the complex operating behavior of the distribution network from the data.Then conduct empirical analysis on the line selection function,verify the correctness of line selection,estimate line parameters based on historical fault data,and correct engineering wiring errors.Finally,based on historical fault data and simulation model,a fault data set is established,and a new method of fault line selection based on PSO-Kmeans particle swarm optimization clustering is proposed.The fault data set is used to verify the correctness of the proposed line selection algorithm.At the same time,for complex faults that are difficult to complete line selection,an intelligent open-circuit selection fault line strategy based on historical fault distribution rules and risk assessment is proposed.