Study On State Pattern Recognition Of Transmission Lines Based On Port Information

Nowadays automatic reclosure is widely used in EHV transmission lines, and isone of the important methods for improving secure and reliable operation of powersystem. Operation experiences of power system show that the faults occurred onEHV transmission lines are mostly transient and single phase to ground fault typemajority. But conventional autoreclosure’s shortcoming is that it acts blindly entirelybefore reclosing and can’t distinguish transient fault from permanent fault. It willcause serious damage to power system. If fault property could be judged advancedbefore the operation of reclosure, that is to say, the reclosure could operationadaptively according to the characteristic of the fault, power system could keepworking stable and improve the success rate of reclosure because of the avoidance ofreclosing on permanent fault and the subsequent impact on system.Considering that there are so many shunt reactors with neutral small reactors inEHV transmission lines, this paper detailed analysis the characteristics of recoveryvoltage at open-phase during the time of adaptive single-phase reclosing, includingthe component, the relationship between the size and influence factors. Because ofthe existing shunt reactors with initial energy, the terminal voltage at open-phaseshows difference between transient fault and permanent fault. For transient fault, theterminal voltage at open-phase exists the beat frequency phenomenon. But forpermanent fault, the terminal voltage at open-phase contains only fundamentalfrequency component, doesn’t exist the beat frequency phenomenon. Use Matlab tosimulate the two kinds of fault respectively and obtain the signals of the terminalvoltage at open-phase. This paper applies stochastic subspace identification(SSI)method to distinguish the nature of fault, and proposes a decision spaceidentification method for distinguishing the properties of fault automatically.The scheme uses the recovery voltage of fault phase as characteristic quantity, and achieves system eigenvalues through EMD, stabilization diagram and SSI.Based on the distribution characteristics of system eigenvalues in the complex plane,define a decision space of fault characteristics, using the system eigenvalues as areference point and permissible deviation of frequency as the radius. If theeigenvalues of fault signal all locate in the decision space, then the fault type ispermanent fault, otherwise the fault type is transient fault. The quantitative indexusing to distinguish fault type is given in the paper in order to achieve automaticrecognition of fault properties. Simulation results show that the proposed scheme canjudge the fault property correctly and reliably without cross-regional discriminationresults.