Research Of Traveling Wave Fault Location Method For Transmission Line Based On WNN

At present,traveling wave ranging method is widely used in transmission lines as a research hotspot.Traveling wave ranging is mainly divided into single terminal method and double terminal method.The two terminal method of traveling wave fault location have higher requirements for time synchronization and communication performance at both ends of the line.If the two terminals of the line are subject to different power grid control departments,affected by the simultaneous start-up factors of the traveling wave data transmission system at both ends and the hardware system at both ends,the use of the two terminal method may not be timely,with low accuracy,or even fail.Compared with the two terminal method,the single terminal method is less limited by external factors.Traveling wave head recognition can affect the reliability of the single terminal method,and then affect the ranging accuracy.Therefore,it is very important and necessary to solve the reliability of traveling wave ranging single terminal method.In this paper,a fault location method based on intelligent method is proposed for the identification of traveling wave head in single end traveling wave ranging method.The main contents are as follows:1.First of all,this paper analyzes the current technical status of fault location methods for transmission lines,including fault analysis method,traveling wave method and artificial intelligence method,and analyzes the advantages of traveling wave method and the technical difficulties of the current traveling wave method.Aiming at the identification problem of traveling wave property in the single terminal method,this paper proposes an artificial intelligence method to solve the problem.2.This paper summarizes the theoretical basis of single terminal traveling wave ranging method and wavelet neural network,analyzes the propagation process of traveling wave in single conductor transmission line,and analyzes the preprocessing work of traveling wave,including the phase mode transformation of traveling wave,the application of wavelet transform and the modulus maximum result of wavelet transform.3.Based on the analysis of traveling wave characteristics of transmission lines and the working principle of wavelet neural network,the sample selection principle of wavelet neural network is proposed.4.The traveling wave signal is extracted,and the functional relationship between the time value of traveling wave arriving at the protection installation,the wave speed and the line length is analyzed;the relationship between the singularity of fault traveling wave and the Li’s index is analyzed.The traveling wave time value and Li’s index are combined to form the network sample value.5.This paper constructs the transmission line model of wavelet neural network,analyzes the important factors that affect the performance of wavelet neural network,and puts forward the optimization method of wavelet neural network,which mainly includes the optimization of learning algorithm,training algorithm,the selection of wavelet basis function and the construction of network structure.6.In the MATLAB / Simulink environment,the transmission line simulation model is built,the wavelet neural network program is completed,and the ranging effect of the wavelet neural network is verified in the simulation.At the same time,the performance of the wavelet neural network under the traditional training method and the wavelet neural network after the optimized training algorithm is compared.The simulation results show that the wavelet neural network trained by the improved PSO algorithm has higher precision,smaller error and faster convergence speed than the traditional wavelet neural network,and the ranging results are not affected by the transition resistance and fault type.