The Research Of Transformer Relay Protection Based On SVM

At present,Power system plays an important role in the process of the development of national economy and has a profound impact on the development of the society from all walks of life.Meanwhile,the transformer protection has become a top priority in the whole power system protection.At the same time,in addition to the complexity of the transformer's own system,the artificial damage,natural factors and Force majeure,We focus on how to boost sensitivity of transformer and check out transformer's fault type and fault location so that the reliability and accuracy.of transformer can be improved.And it will be the important direction of current transformer protection research.The key to diagnose fault problem of Transformer relay protection is to identify the fault type quickly and determine the fault location and fault components.As an important section of power system,the current transformer protection mainly adopts the traditional method of differential protection.For refusing action and error action of transformer caused by magnetizing inrush current,it is still difficult to solve.During recent years,the problems of low accuracy of transformer protection remains unsolved on account of magnetizing inrush current.So,in terms of these reasons above,this paper puts forward a kind of support vector machine(SVM)model of transformer fault diagnosis,which is Based on current situation of the study of transformer relay protection from home and abroad.First of all,according to the basic principle of support vector machine(SVM),we make analysis of the kernel function parameter by optimization algorithm(such as The grid search method,GA,PSO,etc.)so that we can find out the optimal value of related parameters and the optimal value of regularization of support vector machine(SVM).Then,we can establish a model of transformer fault diagnosis on the basis of PSO-SVM algorithm.meanwhile,we can establish a simulation model of magnetizing inrush current to analyze the reason why it have an impact on fault discrimination of transformer.By electromagnetic transient simulation software PSCAD/EMTDC simulation for fault diagnosis model of transformer,we can get the sample data of transformer.And then sample data can be trained and tested using LIBSVM toolkit in MATLAB.Finally,The results of fault diagnosis can be achieved.This system can be more simple and effective to predict the fault types and enhance the validity of fault diagnosis of the transformer,which greatly reduces the probabilities of refusing action and error action of transformer caused by magnetizing inrush current.What we do can not only make model of the transformer protection more simple and intelligent,but also point out a new direction for the development of transformer protection in the future.