Research On Intelligent Line Selection Of Single-phase Grounding Fault In Low Current Grounding System Of Medium Voltage Distribution Network

The neutral point grounding method of China's medium voltage distribution network is neutral point ungrounded or grounded by arc suppression coil.The probability of single-phase grounding among various fault types is the highest,reaching 70%-80%.When a single-phase earth fault occurs in the system,the current flowing through the grounding point is small.At this time,the line voltage of the system remains unchanged.In this case,allow system to run with fault 1-2h.However,as the scale of the distribution network system continues to expand and more and more cables are used,the capacitance of the line to the ground increases,which in turn leads to a large capacitance current of the line to the ground,making it difficult for the grounding arc at the fault point to self-extinguish,which may The scope of the fault is further expanded,which seriously threatens the safety of the system.Therefore,when a singlephase ground fault occurs in the system,it must be quickly identified according to the fault information which circuit has a fault,which is of great significance for ensuring the safe and stable operation of the power grid.This paper first expounds the basic knowledge of the small current grounding system,and analyzes the transient and steady-state characteristics of the neutral point ungrounded system and the neutral point through the arc suppression coil grounding system when single-phase ground fault occurs,and the differences between the old and new transient equivalent circuits are compared,and the new equivalent circuit is more accurate by simulation.Then,EMTDC/PSCAD is used to build a single-phase ground fault model of the neutral point of the 10 kV distribution network through the arc suppression coil.At the same time,the arc ground fault model and the high-resistance ground fault model are built and verified by simulation.In order to verify the difference between the zero-sequence current fault of the faulty line and the zero-sequence current fault characteristic of the non-faulty line when the fault occurs in the system,this paper does a lot of simulation by changing the closing angle,transition resistance,and fault distance when the fault occurs.The simulation results are consistent with the theoretical analysis.Based on the simulation data,this paper introduces a new signal denoising algorithm.The principle of the algorithm is to combine the complementary ensemble empirical mode decomposition algorithm with the permutation entropy algorithm,which is an improved ensemble mode decomposition algorithm.The zero-sequence current of each line is extracted by this algorithm,and it is verified that the algorithm can effectively filter out the noise signal contained in the zero-sequence current.Then,the characteristic intrinsic mode energy of the intrinsic mode function decomposed by the algorithm is obtained,and the zero-order active component of each line and the integral parameter defined when using the transient direction method are obtained.Finally,the probabilistic neural network is studied in detail.Finally,the probabilistic neural network is studied in detail,and the data of the intrinsic mode energy and the zero-order active component and the defined integral parameter are input into the probabilistic neural network as samples to select the faulty line.The result of the line selection proves that the method is equally applicable to the case of high resistance grounding.