Analysis Of Fault Characteristics And Research On Diagnosis Technology Of Synchronous Condenser

With the wide application of UHVDC transmission project in China,the reactive power and voltage support of the DC transmission and receiving network are insufficient,resulting in inverter commutation failure and voltage drop..The synchronous condenser has the advantages of large reactive capacity and strong voltage support capability,and can effectively solve the above problems,so it is applied in the power grid.However,the largecapacity synchronous condenser is bulky,complicated in structure,and prone to failure.In order to ensure the safe operation of the synchronous condenser after grid connection,so as to meet the needs of reactive power support of UHV DC transmission system,it is necessary to study its fault characteristics and diagnostic techniques.In this paper,by constructing the finite element model,the mathematical simulation and simulation analysis of the air gap magnetic density,electromagnetic force and vibration characteristics of the synchronous condenser are carried out,and the reference value of the fault characteristics is provided.Then,the air gap magnetic field,the electromagnetic force of the stator and the stator and the rotor vibration characteristics of the synchronous condenser in the air gap static eccentricity,air gap dynamic eccentricity,stator winding short circuit,rotor short circuit and other fault conditions are analyzed,and the faults are obtained.The harmonic component contained in the synchronous condenser stator vibration signal proves that the fault state can be diagnosed by observing the synchronous condenser vibration signal.By comparing the characteristics of fast Fourier transform,windowed Fourier transform,wavelet analysis and wavelet packet analysis on signal feature extraction,wavelet packet energy is used to extract the vibration signal of synchronous condenser,and principal component analysis is used to reduce feature vector..The model structure of RBF neural network is introduced,and the process of neural network learning training is deduced.The clustering algorithm for automatically acquiring K value and hidden layer center is proposed for the problem of K value selection.The fault analysis of RBF neural network is carried out on the dimensionality reduction vector.The fault diagnosis algorithm based on wavelet packet-neural network is completed and the reliability is verified.In this paper,the hardware circuit of the synchronous condenser fault diagnosis device is designed by monitoring the electrical signal of the synchronous condenser and the rotor vibration signal.In order to meet the actual needs of the project,the software development of the synchronous capacitor electrical signal diagnosis and vibration signal diagnosis function is carried out.A synchronous condenser fault diagnosis system and device have been developed and functional tests have been carried out.The test results show that the diagnostic device can realize the state monitoring and fault diagnosis of the electrical signal and vibration signal of the synchronous condenser.