Research On Fault Location Of Transmission Line Based On Double-ended Nature Frequency Of Travelling Waves

With the continuous expansion of the power system,the structure of transmission network becomes increasingly complex,and the extremely harsh natural conditions make the transmission lines exposed in the air extremely prone to various failures.In the event of a failure of the transmission line,it will cause a large-scale blackout,affecting the industrial and commercial production and people’s normal life,and causing significant losses of the national economy.Therefore,when the line fails,how to quickly and accurately determine the type of fault and realize fault location is of great significance to ensure the stability and economic operation of the power system.The traveling wave methods are most widely used in fault location.However,the accuracy of the A-type method and the traditional frequency domain method is affected by the uncertainty of the traveling wave velocity,and the D-type method is affected by the double-ended clock synchronization,and both the A-type method and the D-type wave head are difficult to extract;This thesis mainly completed the following aspects:(1)Through the analysis of the mathematical relationship between the natural frequency of the traveling wave and the fault distance and boundary conditions,it is determined that the natural frequency of the traveling wave is used for fault location.After comparing the advantages and disadvantages of FFT transform,HHT transform and wavelet packet transform,the wavelet packet based is determined.Transforming the fault location method of the transmission line that extracts the natural frequency of the traveling wave.(2)Aiming at the problem that it is difficult to accurately classify faults of transmission lines based on single electrical characteristics,the thesis proposes a fault classification method based on wavelet packet transform and convolutional neural network.The method uses threephase current and zero-sequence current transient components as two-dimensional.Feature vectors that clearly characterize different fault types and are unaffected by fault distance,fault resistance,and double-ended system angle.The simulation results show that the proposed method improves the classification accuracy compared with BP neural network,and has good adaptability.It provides basic theoretical support for the selection of traveling wave modes for fault location.(3)Aiming at the problem of extract for traveling wave head and the influence of traveling wave velocity,resulting in low location accuracy,the thesis proposed a fault location method for transmission line based on double-end traveling wave natural frequency.This method does not require wave speed information and clock synchronization information.It also does not require too high sampling frequency,which is more suitable for practical applications than Atype and D-type traveling wave ranging.Simulations under different fault types,different fault grounding resistances,different fault distances,show it has a higher location accuracy.This thesis finally formed a complete method from fault classification to fault location of transmission lines,which has certain application prospects.