Research On Single-phase Arc Grounding Fault And Line Selection In Distribution System

At present,80% of the faults occurring in China’s distribution network are single-phase ground faults,and the damage caused by arc ground faults is the most serious.First,the state of the arc in the arc ground fault is dynamically changing,alternating between arcing and extinction,and high frequency components are generated in the system.Due to the presence of inductive components and line-to-ground capacitance in the distribution network system,the system will generate electromagnetic oscillations,and the overvoltage generated by electromagnetic oscillation will cause damage to the distribution network system.Second,the arc ground fault is a small current ground fault.One type,the system does not have high accuracy in selecting such faults,and then may operate with faults,and the generation of arcs will release a large amount of heat,which may cause an accident to expand.In order to solve this problem,the most important thing is to be able to accurately locate the line where the arc ground fault occurs,thereby cutting off the power supply of the faulty line,thereby ensuring the safe and stable operation of the system.In practical power distribution system,it is difficult to collect the characteristic signals corresponding to arc grounding fault,so in order to study the problem of line selection,simulation is an effective method.In the study of single-phase arc grounding fault,the commonly used models are assumed that the resistance value of the grounding arc is zero(gold attribute grounding)or a constant value.But the model can not accurately reflect the dynamic characteristics of the actual arc.The other is the classical Mayr and Cassie black box models,both of which have inherent limitations.Therefore,first of all,a simulation model which can accurately reflect the fault characteristics of arc grounding is established.Up to now,in the research process of arc-light single-phase grounding fault line selection technology in distribution network,although experts and scholars have proposed that the steady-state eigenvalue method and transient characteristic value method can be used to select the line of the system.However,it is still not possible to select the single-phase arc grounding fault exactly.Therefore,combining the advantages and disadvantages of each method,it is a more suitable line selection direction to take advantage of the advantages and disadvantages of each method.In order to accurately simulate the single-phase arc grounding fault,the paper firstly describes the arc theory and the mathematical model of the arc.The traditional arc model is analyzed.Basedon the corresponding limitations,based on the cybernetic model,a An arc model suitable for single-phase arc grounding faults in distribution network,and compares its influence on fault characteristics under different arc lengths and line lengths;to solve the limitation of single line selection method for fault line selection In this paper,several line selection methods that can reflect the characteristics of single-phase arc ground faults are analyzed.The feature selection criteria are used to study the line selection and the limitations are pointed out,On this basis,the Softmax regression algorithm is introduced into the distribution.In the network single-phase ground fault line selection,the technical means of comprehensive utilization of multiple feature quantities is utilized to cope with the problem that the single line selection feature method will fail.In the Tensorflow environment,the method combines a plurality of feature selection features,and iteratively trains the feature data to obtain a cost function.The cost function is used to obtain the probability of the fault occurrence line.By using a single method feature quantity to identify the failed fault data for verification,the method can still achieve a good line selection effect.