Research On Fault Diagnosis And Fault Tolerance For Electrical Vehicle Electrical Drive System Based On Permanent Magnet Synchronous Motor

Safety and reliability are important to vehicle.As electric vehicles(EV)are gradually replacing conventional cars,the safety and reliability of electric vehicles have a significant research value.Electric drive system is the core of EV powertrain,fault diagnosis and fault-tolerance technology can improve the reliability of electric drive system.In this paper,the discussion focuses on electrical drive system based on permanent magnet synchronous motor(PMSM),its position sensor fault,irreversible demagnetization fault and interturn short circuit fault are the theoretical and experimental researched.Compared with classical motor fault diagnosis system,the electric drive system is limited by cost and space,and therefore,its fault diagnosis system cannot add hardware.Meanwhile,there is plenty of intervals in the EV operation,the offline fault diagnositic method can be introduced to improve diagnositic results.In this paper,considering EV work cycle,the relavent design of fault diagnosis and fault-tolerance methods for the position sensor fault,irreversible demagnetization fault and interturn short circuit fault are designed based on electric drive system hardware platform.For the position sensor fault,a sensorless vector control algorithm is proposed as a fault-tolerance method.This algortithm is a mixed method based on linear anti disturbance rejecton control(LADRC).For high speed region,linear extended state obsevers(LESO)are used to obsever motor electromotive force(EMF),and the position is estimated by a phase locking loop(PLL).For low speed region,a metod based on high frequency sinusoidal current injection is proposed.A sinusoidal harmonic current is injected into the stator,and the position estimation is achieved by detecting the relevant harmonic voltage output.Compared with the classical sensorless control method,the high speed and low speed methods are integrated into the same control scheme.Moreover,the current control and positon estimation are integrated based on the advantage of LADRC.The propsed method is validated by the simulation and experiments.For interturn short circuit fault,a diagnostic stragety integrating online and offline fault detection is proposed.During the electric drive system op eration,the fault diagnostic system mainly relies on the online detetion.Durting the interval of halt,the offline detection is executed as the effective supplement.The sensitivity of diagnostic stragety is optimized by the integration of the two methods based on EV work cycle.For online detection,the model of faulty motor is constructed to analyze the motor operating characteristic.Based on the analysis,the positive sequence third harmonic in motor EMF is used as the fault characteristic,and the relevant diagnostic algorithm is proposed based on LESO and second order generalized integrator(SOGI).This algorithm can reduce computation cost without large data storage.Meanwhile,the offline detection based on high frequency voltage excitation is propsed.Based on the anlaysis of fautly motor model,the positive sequence high frequency component of reponse current is applied as the fault characteristic.In the method,the high frequency is used to amplify the asymmetry of the windings to improve diagnostic sensitivity.The propsed method is validated by the simulation and experiments.For demagnetization fault,the relevant fault diagnosis and fault-tolerance methods are proposed.First,for fault diagnostic method,SOGI is constructed to estimate the fundamental frequency component of motor flux and abstract the EMF harmonics.Second,the characteristic harmonic is calculated by a special ratoting reference frame.The fault diagnosis is achieved by the amplitude of fundamental frequency component of motor flux and characteristic harmonic.For fault-tolerance method,a maximum torque per amp stragety based on motor parameter identification is proposed.This method can eliminate the output torque decrease caused by demagnetization fault.The experimental platform using a DSP TMS320F28335 as control chip is constructed to validate the proposed methods.The relevant experimental results demonstrate the feasibility of the fault diagnosis and fault-tolerance motheds.