Feature Extraction And Recognition Of Transmission Line Fault Diagnosis Based On Improving VMD-MSE Methods

With the rapid development of China's economy,in the past 30 years,China's power system has entered a new era of power generation,including large units,large power plants,large power grids,ultra-high voltage and automation.Therefore,further requirements are imposed on the transmission power,the transmission distance and the voltage level.In power systems,AC highvoltage long-distance transmission is widely used because of its flexibility.In the AC transmission system,the transmission line is a very important part of the power system.It plays the role of transmitting power over long distances.Once a short circuit fault occurs in the transmission line,it must respond quickly and remove the fault,otherwise it will affect the security and stability of entire power system.Therefore,it is crucial to study a method to quickly identify short-circuit faults in transmission lines.When a short circuit fault occurs in the transmission line,the fault signal contains a large amount of transient information.However,due to the characteristics of the transmission line itself,the long distance,large transmission capacity and harsh surrounding environment are more susceptible to the surrounding electromagnetic environment,which makes the fault signal feature extraction difficult and incomplete,resulting in lower accuracy of the final recognition result.Therefore,for the short-circuit fault of transmission line,the feature extraction of fault signal and the identification of fault,this thesis proposes an improved variational mode decomposition(VMD)-multi-scale entropy(MSE)method for short-circuit fault feature extraction and identification of transmission lines.Firstly,the Drosophila optimization algorithm is used to determine the number of components of the variational mode decomposition algorithm and the penalty parameters to ensure the best decomposition effect.Then,the variable mode(VMD)decomposition method is used to resolve the current signal of four typical short circuit faults of the transmission line(Ag,ABg,AB,ABC)into several eigenmode components(IMF);Lastly,multi-scale entropy(MSE)is used to calculate the four eigenmode components of the four shortcircuit current signals after decomposition by the improved VMD algorithm.The multi-scale entropy value of(IMF)is composed of eigenvectors and its fault characteristics are identified by generalized regression neural network(GRNN).The experimental results show that VMD can decompose the fault signal better than EMD and suppress the occurrence of modal aliasing.The improved VMD-MSE method can effectively extract the fault characteristics of four short-circuit current signals of transmission lines,which can extract more obvious fault features compared with MSE method.When the generalized regression neural network(GRNN)is used to identify the short-circuit faults in different fault initial angles,different transition resistances and different fault locations,the average recognition rate reaches 96%.The effectiveness of the method was finally proved.There are 35 figures,6 tables and 59 references in this thesis.