Fault Diagnosis Of Transmission Line Based On Neural Network

With the rapid development of China's modernization,the power industry has entered a new stage characterized by "three types and two networks",and the reliability of power supply system becomes more and more important.The efficient operation of power grid depends on the safe operation state of power lines to the greatest extent.As the transmission line is an important link to connect the power,once the line fault occurs,it not only poses a threat to the safe operation of power system,but also causes great economic harm.Therefore,it is necessary to accurately and timely identify faults and fault line selection for transmission lines,so as to restore power supply as soon as possible and reduce economic losses caused by power failure.Firstly,the 110 kv current grounding system model was built in Simulink simulation software to simulate the zero-sequence current and voltage of different fault types and analyze the transient characteristic energy values generated by transmission line faults.Then,using the wavelet packet transform to the fault signal processing,low frequency and high frequency analysis to fault transient characteristics of the energy value of the wavelet packet decomposition of three layers,choose db3 wavelet basis function decomposition,it is concluded that the energy of different frequency entropy will be different energy eigenvalues and neural network with transmission line selection function which can identify each other.The wavelet toolbox in MATLAB simulation software is used to verify that the algorithm can extract fault values efficiently and accurately,and that the energy values of different faults are different.Eventually,the fault diagnosis model of transmission line is established by using neural network algorithm,and the fault characteristic data processed by wavelet is input into the fault diagnosis system to identify the fault type of transmission line.However,in order to optimize the improvement of fault recognition in the network,Elman neural network is cited to establish a diagnosis recognition model.Using the same criterion,the two networks are simulated one by one,the MATLAB simulation software test,simulation experiment of Elman neural network in fault recognition rate and convergence steps will take precedence over the BP neural network,and can better,faster,more accurately identify the fault type,to complete the task of transmission line fault diagnosis.