Research On Faults In Position Sensors And Fault Tolerant Control Of Brushless Permanent Motor

With the problems of fuel shortage and environmental pollution being worse and worse,conventional vehicles will be replaced by electric vehicles in the near future.In the applications of electric vehicles,high reliability and fault-tolerant capability of electric drives are required,in which the position sensor of the electric motor plays an important role.Therefore,faults of position sensor were investigated for the brushless permanent magnet motor(BLPMM)drives,and several fault diagnosis and fault tolerant control methods were proposed.Furthermore,the improvement of the fault tolerant control methods was also investigated.Modeling and simulation are essential approaches for fault analysis and fault tolerant research,so the detailed model of the BLPMM drive was investigated comprehensively.First the BLPMM is modeled in the stationary reference frame and the synchronous rotating frame,respectively,the relationship and characteristic of which were also studied.According to the operation states,the three-phase inverter was modeled based on the drive signals and output currents.Then the relationship between the inverter output voltages and the phase voltages was proposed.In the two referecnce frames,the six-step dirve method and the vector control method were analyzed.According to their implementation in the micro-controller,the control methods were modeled based on the timer interrupt and event-triggered interrupt programs.The whole drive system were modeled and simulated in MATLAB/SIMULINK,which is very important for the fault tolerant research.Then the faults of Hall position sensor and the corresponding performance deterioration of the BLPMM drives were investiaged.The rotor position information implied in the Hall signals was analyzed,during which Hall vector state,Hall scalar state and Hall transitions were defined.Besides,the fault type and fault angle were also defined respectively,according to the number of faulty sensor and the fault occurring position.The influence of the fault type,fault angle and final value of faulty sensor on the Hall states and transitions were studied,which was also investigated in the simulation.The simulation results depicted that the steady-state and transient performance of the drive system were directly and respectively affected by the fault type and fault angle,while the final value of the faulty sensor influences the sequences of the motor currents.After that,three fault diagnosis methods were proposed based on the research of Hall signals during the steady-state process of the Hall sensor faults.The unique sequences of Hall states and transitions were summarized for each fault type,based on which Hall-state-sensitive fault diagnosis(FD)method and Hall-transition-sensitive FD method were proposed.To make better use of Hall signals,the Hall states and transitions were combined to produce a rotary direction detection methods.Then the detected results were investigated for all the fault situations,based on which a rotary-direction-sensitive FD method was presented.In the simulation,the three FD methods were investigated and compared,which showed that the efficiency of the latter method was higher than the former one.Additionally,the simulation results also depicted that the fault angle and the fault type were the key factors that influenced the diagnosis process.Based on the fault diagnosis methods,the fault tolerant control of the BLPMM drive system was researched.The angle resolution required by six-step control and vector control methods was analyzed,which is also the requirement of angle estimation in fault tolerant control.By using the Hall signals and time intervals measured,the speed and acceleration of the motor are estimated.Then a novel extrapolation method was proposed for Hall sensor fault situations.To improve the precision of the angle estimation during dynamic process,a novel Hall-vector-tracking obsever was presented.In the observer,the spectrum of Hall vector was analyzed under all the fault situations,and then the harmonics decoupling method were proposed.The the two fault tolerant control(FTC)methods base on the two angle estimation strategies were tested in an BLPMM drive system,where the FTC methods were implemented in both software form and hardware form.The experimental results depicted that the tow FTC methods can recover the drive system from Hall sensor faults and the method based on Hall-vector-tracking observer showed better dynamic performance.It should be noted that the fault diagnosis methods were also verified during the experiment.Although the proposed FTC methods are effective for the BLPMM drive system,large transient currents/torque can occur during fault diagnosis process(FDP)resulting in speed dip and degraded drive performance.During FDP,the fault type of Hall sensor cannot be diagnosed immediately,so the FTC methods cannot be activated appropriately until the end of FDP.The acceleration estimation and fault diagnosis results were analyzed during FDP,and the indication method for FDP was proposed.Then a structure-variable Hall-vector-tracking observer was propsed,which can convert into an open-loop observer to eliminate the influence of the fault during the FDP and convert back to close-loop form thereafter.The experimental results showed that the method eliminated the abnormal currents and speed fluctuation,i.e.the FTC methods were improved significantly.