Sensor Fault-Tolerant Control Of Permanent Magnet Traction Drive System Based On Equivalent-Input-Disturbance

As the power source of rail transit traction equipment,motors and its drive systems have undergone three generations of changes from “DC” to“AC”,and then to “permanent magnets”.With the first full-line permanent magnet traction system in China,Changsha Metro Line 5 officially carries passengers,rail transit traction has truly ushered in the “permanent magnet era”.However,the on-site application environment of rail traction control systems is often very harsh.In addition,the motors need to be frequently started,accelerated,decelerated and braked,which can easily induce the fault of the motor drive system.It is very likely to cause catastrophic consequences if the fault is not detected and corrected in time.At present,there are many researches related to fault-tolerance control of permanent magnet traction motors,but few researches on sensor faults of the systems are carried out.At the same time,due to the harsh field application environment of the system,unknown external disturbance is easily affected the diagnosis and fault-tolerance control of the faults.Therefore,it is necessary to require the fault diagnosis system to be insensitive to unknown disturbance and to maintain high sensitivity to small faults.However,the existing fault diagnosis methods difficult to achieve both,and often resort to compromises.In terms of mechanism,faults and disturbances are often coupled with each other and act on the system together.If the decoupling and reconfiguration of faults and disturbances can be realized,this contradiction will be easily solved.In addition,the existing fault tolerance methods often used the inverse model of the system to reconstruct the fault-tolerant controller,or to reconstruct different fault-tolerant structures for different fault modes.This greatly reduces the stability of the system and the timeliness of fault-tolerance control.The equivalent-input-disturbance theory just does not need the prior information of the fault model and the inverse model of the system.Which makes up for the shortcomings of the existing fault-tolerance methods.In the context of locomotive traction control system object.This article carries out research on fault diagnosis and tolerant of current sensors based on equivalent-input-disturbance theory and sliding mode variable structure technology.The main research contents are as follows:First of all,for the problem the coupling of disturbance and sensor fault,a method for decoupling sensor faults and unknown disturbances and a strategy to estimate the decoupling faults and disturbances are presented.Based on the coordinate transformation method,the decoupling of the fault and the unknown input disturbance is realized.The sliding mode equivalence principle is used and the sliding mode observer is designed to realize the reconstruction of the fault and the disturbance;At the same time,the detailed system-stability-analyze is carried out.which verified the feasibility and effectiveness of the method.Secondly,in view of reconstructing fault-tolerant controller or different fault-tolerant structure for different fault modes in the existing fault-tolerant control methods,the strategy of estimating and tolerant of current sensor faults for permanent magnet traction drive systems is presented.By rebuilding the system,the sensor faults in the current measurement circuit are regarded as system disturbances.Based on the equivalent-input-disturbance theory and sliding mode variable structure technology,a sliding mode observer and filter are designed to form an improved equivalent-input-disturbance estimator,which estimated the equivalent-input-fault.The fault-tolerance control of the sensor fault is realized by subtracting the equivalent-input-fault from the control input.The RT-LAB platform hardware-in-the-loop experiment was performed on the sensor fault example,which verified the feasibility and effectiveness of the method.