Research On Fault Diagnosis Of Transmission Line Based On Wavelet Neural Network

It is the primary task of power system to provide high-quality energy supply services to the vast number of power users.When power system fails,relay protection is needed to shorten the time of failure and reduce the harm caused by failure.The equipment used in traditional relay protection has a single function,which is easily interfered by factors such as environment,low accuracy and consumes more manpower.Therefore,how to save costs while ensuring the safe operation of power system has become a research focus of modern relay protection technology.The rapid development of artificial neural network technology provides a new development direction for relay protection technology.Applying artificial neural network technology to relay protection can improve the intelligence of the system,effectively save costs and improve the anti-interference ability of relay protection devices.BP neural network is the most mature and widely used neural network at present,so people mostly take BP neural network as the research object in the research of related application schemes.However,BP neural network is easy to fall into local optimum,and the training time is too long,resulting in a large error in diagnosis results.Therefore,in order to improve the performance of neural network in fault diagnosis of transmission lines,this paper takes the 110KV neutral grounding power system as the research object,and applies the method of combining neural network and wavelet theory to fault diagnosis of power system.First of all,this paper analyzes the short-circuit faults,and applies a large number of measured data to establish short-circuit fault type discrimination,and then it screens out correct different fault types.Based on the analysis of neural network and wavelet theory,the wavelet neural network is constructed,and the normalized conjugate gradient method is used to optimize the wavelet neural network,which improves the accuracy of transmission line fault diagnosis.In this paper,the wavelet neural network is established by Matlab software,and the fault data obtained from a 110KV power system is used to complete the training and testing of the wavelet neural network.Training BP neural network with the same fault data.By comparing the training results of BP neural network and wavelet neural network,it shows that wavelet neural network can meet the requirements of transmission line fault diagnosis,and its performance is better than traditional BP neural network.Finally,on the basis of the constructed wavelet neural network,from the point of view of considering factors such as iteration times and errors,several improvement schemes are proposed for the wavelet neural network.By comparing their training results,the wavelet neural network is optimized by using the normalized conjugate gradient method,which improves the performance of wavelet neural network in transmission line fault diagnosis.