Research On Overvoltage Analysis And Identification Method Of Flexible DC Distribution System

The flexible DC power distribution system has flexible structure,complex control,and large number of components.In the long-term operation of the system,the insulation aging of the equipment,bird crossover wires,lightning disasters and other phenomena are inevitable,which may lead to faults such as system short circuit and insulator string flashover.Different fault types have different effects on the system,and the corresponding protection strategies of the system are also different.It is of great significance to analyze the overvoltage characteristics of common short-circuit faults and lightning faults in the system and improve the identification accuracy of system fault types,which is conducive to the operation and maintenance personnel to take accurate protective measures and improve the insulation design of related equipment.Firstly,the advantages and disadvantages of different configurations of the system are compared,and the operation characteristics of the system are analyzed.The equivalent circuit method is applied to establish the mathematical model of the converter,and the control method of the system is designed.The PSCAD software is used to build the simulation model of the operation overvoltage of the system.Based on the fault principle and simulation model,the influence of four typical faults on the voltage of the AC side,the converter side and the DC side of the system is analyzed,considering the action of the circuit breaker.The short circuit fault of the converter bridge arm will cause the converter to be locked.The current of the lower bridge arm is zero,and the upper bridge arm still has a large over current.DC unipolar grounding fault has an impact on the DC side voltage of the system.The short circuit fault between DC poles seriously affects the system operation,and the action of DC circuit breaker affects the voltage at both ends of the DC reactor.Secondly,on the basis of the original system parameters,the lightning overvoltage simulation model of the system is built,and the influence of back flashover and shielding failure on the AC side,converter side and DC side voltage of110 k V AC transmission line is analyzed according to the flashover and without flashover of insulator string.When the insulator string does not flashover,the influence on the system voltage is small.When the insulator string flashovers,the non-fault phase voltage recovers rapidly,the fault phase voltage attenuates gradually,and the converter side generates large overvoltage.When back flashover occurs,the direction of voltage traveling wave of transmission line is the same.When shielding failure occurs,the direction of voltage traveling wave of hit line is opposite to that of other lines.Finally,based on the operating overvoltage and lightning overvoltage simulation model of flexible DC distribution system,six kinds of fault types are selected,and the DC side voltage signals are sampled and processed.A feature extraction method based on discrete wavelet transform and singular value decomposition is proposed,and a multi-level support vector machine model is established for signal recognition and fault classification.Support vector machine is used to identify the fault types.The method is compared with PCA-SVM,Wenergy-SVM,DWT-SVD-MLP and DWT-PCA-SVM recognition methods.The results show that the accuracy of the method is 99.88 %,and the accuracy of PCA-SVM is 90.75 %.