Fault Diagnosis Of Commutation Failure In High-voltage AC-DC Hybrid Transmission System

High voltage DC transmission technology has been developed rapidly because of its obvious superiority in power transmission and other aspects.Phase change failure is one of the most common faults in high-voltage AC/DC hybrid transmission systems and can cause serious consequences such as DC blocking and power oscillations.A correct and rapid diagnosis of phase change failure is an important prerequisite for the control of the entire system and for the timely adoption of corresponding control measures.However,the current fault diagnosis of phase change failure is not comprehensive and the accuracy of fault diagnosis is not high due to a single variable.Therefore,this thesis carries out relevant research on the diagnosis of phase change failure faults in high-voltage AC-DC hybrid transmission systems,with the following main work and contributions.(1)The phase change failure fault diagnosis methods are sorted and summarized.The fault diagnosis methods of high-voltage AC-DC hybrid transmission systems are summarized and sorted out,and a new fault diagnosis model for phase change failure is established.(2)The influence of faults occurring on the AC-DC side on phase change failure is analyzed.A model of a 500 k V high-voltage AC-DC hybrid transmission system is built in the PSCAD simulation software,and the effects of faults occurring on the AC system in different environments on phase change failure are analyzed through simulation.(3)PSO-RBF diagnostic model and GS-RF diagnostic model were established.The AC voltage,DC current and rectifier-side trigger angle command values at the commutation bus when different faults occur in the high-voltage AC-DC hybrid transmission system under different fault conditions are first collected as a fault feature database.Particle swarm algorithm(PSO)and grid search algorithm(GS)are used to optimize the RBF neural network and random forest,respectively,to establish the PSORBF phase change failure fault diagnosis model and GS-RF phase change failure fault diagnosis model.The two models are used to determine whether the faults occurring on the AC and DC sides cause phase change failure,the type of faults causing phase change failure and the location of the faults,and the best fault diagnosis model is selected by comparing the two models to provide a reference for accurate and fast fault isolation,a basis for the implementation of DC control and protection,and a technical means and support for the construction of new power systems.The experiments demonstrate that the accuracy of the algorithm-optimized fault diagnosis model is better than the original fault diagnosis method,in which the PSO-RBF phase change failure diagnosis model is more accurate than the GS-RF phase change failure diagnosis model for the diagnosis of whether the phase change failure is caused,and the GS-RF phase change failure diagnosis model is more effective for the diagnosis of the fault type and fault location causing the phase change failure.