Protection Scheme For UHV Transmission Line Based On Wavelet Transform And Artificial Neural Network

With the development of social economy and smart grid technology,the ultra high voltage(UHV)transmission line has been studied and applied widely to satisfy people's increasing need for electric power.As the UHV transmission system is usually connected with two large power systems,and it has a further transmission distance and a larger transmission capacity,so ultra-high-speed protection is required for UHV transmission line to reduce the fault clearance time in order to improve the transient stability,transmission capacity and reliability of power systems.To meet the requirements of ultra-high speed protection on UHV transmission lines,the starting method and fault phase selection method based on transient signals were studied in this paper by combining wavelet transform with artificial neural network.Firstly,an extraction method of fault featrues is proposed,which uses wavelet transform to detect singularity of transient fault signal.The transient fault signal has the characteristics of mutation and singularity,wavelet transform can denote the local signal characteristics in both time and frequency domain.In this paper,a phase-mode transform is used to the fault component of transient current for eliminating the impact of the various phases' coupling.The mutation of transient fault signal can be detected by wavelet singularity detection,and the wavelet transform modulus maximas are regarded as the fault featrue.The mode fault featrues of different faults are summarized according to the boundary conditions of different faults.It is proved that this method of fault feature extraction is effective by simulating Jin Dongnan-Nanyang-Jingmen UHV transmission line in PSCAD.Secondly,a starting method based on wavelet transform and artificial neural network is researched.In this paper,a artificial neural network for starting is designed to identify the running state(fault or normal)of UHV transmission line,which regards those fault featrues as the input.By building the UHV transmission line model and simulating different running states in PSCAD,it is proved that this starting method can identify the running states accurately,and it is suitable for ultra-high-speed protection of UHV transmission lines due to the fast and sensitive starting result.Finally,two kinds of fault phase selection methods for ultra-high-speed protection on UHV transmission line are studied and compared.One is proposed on the basis of fault feature extraction by wavelet transform,which summarizes the relationships among those fault featrues of different faults and regards them as the phase selection criterion.The other method is an artificial neural network for fault phase selection,which regard the fault featrues as the input and reflects the fault types according to the results of the neural network output.Some simlations of these two methods are done,and the simulation results show that both of them can identify different fault types of UHV transmission lines correctly.Moreover,they are also not affected by fault resistance,fault location,fault inception angle,and it has high reliability.