Fault Diagnosis Of Traction AC Drive System Of Contactless Network Urban Rail Vehicles

The electric traction AC drive system is the core of the contactless network urban rail vehicles power system.The traction inverter and the permanent magnet synchronous motor are the key components of the traction AC drive system,and have very strict requirements on reliability.From a simplified perspective of quantitative analysis,their typical faults can represent the fault conditions of the entire AC drive system.By monitoring the inverter and the motor in real time and responding to the fault,the normal operation of the system is ensured.Permanent magnet synchronous motor for the contactless network urban rail vehicles has large capacity and poor working environment,which is prone to insulation and demagnetization faults.The power module of the converter is also a weak link which is prone to failure.Therefore,the fault diagnosis methods of traction inverter open circuit and permanent magnet synchronous motor inter-turn short circuit and demagnetization are mainly studied.The main contents are below:(1)In order to enhance the immunity of the traction inverter open circuit fault diagnosis method to uncertain factors,a information fusion diagnosis method based on Bayesian network is proposed.Taking the current and voltage on the output side of the traction inverter as the signal source,the wavelet packet entropy feature is extracted,and the original feature is reduced by principal component analysis.The Bayesian parameter estimation method is used to fuse the reduced feature quantity to obtain new feature vector which contains complementary information.The Bayesian network is used to identify the new eigenvectors after fusion,and the maximum posterior probability estimation results under different observation values are fused to make a decision,so as to complete fault diagnosis.The simulation model is established and compared with the fault classification using K-means algorithm to verify the effectiveness of the proposed fault diagnosis method under different speed and white noise conditions.(2)Aiming at the problem that the fault signal of the permanent magnet synchronous motor is weak and difficult to detect,considering the interference factors such as working environment,the method of extracting the harmonics of the inverter as the excitation source and extracting the fault characteristic quantity from the negative sequence voltage is proposed.Firstly,the fault model of permanent magnet synchronous motor is established,and the fault negative sequence voltage is introduced to calculate the fault index.Then the high frequency harmonic excitation in the voltage is extracted,which is injected in the original voltage signal,and the high frequency negative sequence component is separated and detected by the generalized second order integrator.The simulation results show that the proposed method can effectively identify the inter-turn short-circuit fault of permanent magnet synchronous motor and has strong anti-interference ability.When the fault severity is relatively small and the torque is relatively large,it is more conducive to check the inter-turn short circuit fault.(3)The model reference adaptive algorithm is used to identify the magnetic flux linkage of the permanent magnet synchronous motor and complete the uniform demagnetization fault diagnosis of the permanent magnet synchronous motor.Considering the working condition of vehicle,the fault characteristic harmonics are extracted under non-stationary state,and the stator current harmonics are detected by HHT to complete the local demagnetization fault diagnosis of PMSM.The simulation results show that the methods are effective.