Sliding Mode Technique Based Fault Diagnosis And Fault-tolerant Control With Application To High-Speed Railway Systems

CRHs are products China pays great attention to.Reliability and safety,the main research directions of CRHs and are also two important factors to improve power of CRHs,are affected by incipient faults significantly.However,there are still few mature theories and approaches to diagnose and tolerate incipient faults,and few results applied to CRHs.This dissertation mainly concerns incipient fault diagnosis and fault tolerant control issues.Based on sliding mode techniques and combining with interval estimation techniques,robust control theories and adaptive estimation techniques,this thesis develops new incipient fault diagnosis and fault-tolerant control theories to improve incipient fault diagnosis ability and fault tolerant ability.Then,these theories are applied to the traction devices of CRHs,including single-phase rectifiers,three-phase inverters and traction motors.These works are described as follows:(1)Incipient sensor fault detection issue for the traction inverter nonlinear dynamical systems with "observer unmatched" uncertainties is considered.A particular fault detection sliding mode observer is designed for the augmented system formed by the original system and incipient sensor faults.The designed parameters are obtained using linear matnx equality(LMI)and linear filter techniques to guarantee that the generated residuals are robust to uncertainties and that sliding motion is not destroyed by faults.Then,three levels of novel adaptive thresholds are proposed based on the reduced order sliding mode dynamics,which effectively improve incipient sensor faults detectability.Case study of the traction system is presented to demonstrate the effectiveness of the proposed incipient senor fault detection schemes.(2)Based on adaptive estimation technique and sliding mode technique,two fault accommodation schemes are proposed for incipient sensor faults and actuator faults of traction inverters respectively.In the first fault accommodation scheme,an incipient sensor fault estimation and accommodation methods for three-phase PWM inverter devices in electric railway traction systems are proposed.First,the dynamics of inverters and incipient voltage sensor faults are modeled.Then,for the augmented system formed by original inverter system and incipient sensor faults,an optimal adaptive unknown input observer is proposed to estimate the inverter voltages,currents and the incipient sensor faults.The designed observer guarantees that the estimation errors converge to the minimal invariant ellipsoid.Moreover,based on the output regulator via internal model principle,the fault accommodation controller is proposed to ensure that the output voltages track the desired reference voltages with the tracking error converging to the minimal invariant ellipsoid.Finally,simulations based on the traction system in CRHs are presented to verify the effectiveness of the proposed method.In second fault accommodation scheme,an ISMO based sliding mode actuator fault accommodation framework is proposed for non-minimum phase LPV systems with considering online control allocation problem.Firstly,an ISMO is proposed to estimate the set of admissible values for the states of the faulty LPV systems.Then,a FTC law with an online control allocation scheme is designed via stabilizing the proposed ISMO instead of the original faulty LPV system,which can guarantee that the unmeasurable states of the original LPV system is asymptotically stable and the measurable outputs converge to zero in finite time and the actual control efforts allocated to all actuators are optimized to satisfy prescribed performance.Finally,a simulation based on the inverter used in CRHs is presented to illustrate the effectiveness of the proposed framework(3)For the nonlinear dynamic model of traction rectifiers,an incipient sensor fault detection scheme and a multi incipient sensor fault isolation scheme are proposed.In fault detection scheme,an ISMO based incipient sensor fault detection method for a class of nonlinear control systems with observer unmatched uncertainties are proposed.The interval bounds for continuous nonlinear functions and new injection functions are constructed to design ISMOs.An incipient fault detection framework with newly designed residual and threshold generators is proposed.The detectability is then studied,and a set of sufficient detectable conditions are presented.An application to an electrical traction device used in CRHs is presented to verify the eflfectiveness of the proposed incipient sensor fault detection methodology.In fault isolation scheme,a dc voltage incipient sensor fault isolation method for single-phase three-level rectifier devices in high-speed railway electrical traction systems is proposed.Different incipient fault modes characterizing locations and incipient fault types are parameterized nonlinearly by unknown fault parameters.A new incipient fault isolation method is developed by combining sliding mode technique with nonlinear parametrization adaptive estimation technique.A bank of particular adaptive sliding mode estimators is proposed,which facilitates to derive new isolation residuals and adaptive threshold intervals.The isolability is studied,and the isolable sufficient condition is derived using new functions.For the practical electrical traction system in CRHs,simulation and experiment based on TDCS-FIB(a software)are presented to verify the effectiveness and feasibility of the proposed method(4)A stator-winding incipient shorted-turn fault detection method for the traction motors used in CRHs is proposed.An interval sliding mode observer is proposed to deal with uncertainties caused by measuring errors from motor speed sensors.The active robust residual generator and the corresponding passive robust threshold generator are proposed based on this particularly designed observer.Furthermore,design parameters are optimized to satisfy the fault detectability requirement This developed technique is applied to an electrical traction motor to verify its eflfectiveness and practicability(5)A class of nonlinear systems with unmatched disturbances,which does not possess actuator redundancy,is considered.An actuator fault-tolerant control(FTC)scheme is proposed based on sliding mode control(SMC),which is able to well accommodate unmatched disturbances and actuator stuck fault modes not satisfying redundancy condition.A novel approach for designing smooth control laws and smooth update laws is developed using the equivalent output injections and hyperbolic functions,which can stabilize the system asymptotically,and simultaneously,reconstruct time-varying parameters with reconstruction error converging to zero asymptotically.Based on this approach,a method for sliding motion analysis and sliding mode control design is also proposed using backstepping design procedure.A set of FTC laws and update laws are then designed such that under the proposed FTC strategy,the closed-loop system is uniformly bounded,the output tracking error converges to zero asymptotically,and part of time-varying fault and disturbance signals are reconstructed.The simulation example is presented at last to illustrate the effectiveness of the proposed FTC method.