Robust Control Of Permanent Magnet Synchoronous Machine-Based Railway Traction System Considering Multiple Perturbations

Permanent magnet synchronous machines(PMSMs)outperform conventional induction machines with compact structure,high power density and high efficiency.Such advantages of PMSM accord with the environmentally-friendly and sustainable development goals of domestic high-speed railway traction systems.Therefore,design of high-performance and high-efficiency control strategies for PMSM-based railway traction applications have become an active area of research.Single-phase grid-connected converters and the traction machines with their inverters are the key components of PMSM-based railway traction systems.These converters and machines play an important role in transferring and transforming the energy that comes from the power grid.However,due to the complexity of the operating environment,potential temperature increase in the rotor and demagnetization effects of the permanent magnet,unexpected parametric perturbations(such as perturbations in resistance,inductance,flux linkage,and etc.)and unmodeled dynamics(such as friction torque,cogging torque,errors in sensors,and etc.)widely exist in PMSM-based railway traction systems.Such perturbations deteriorate the transient and steady-state performance of the control of grid-connected converter and PMSM,and degrade the control accuracy and robustness of the whole railway traction application,which has a negative influence on the operating efficiency and driving comfort of the high-speed electric multiple units(EMU).Therefore,this thesis conducts research on the parameter estimation and robust control of the PMSM-based railway traction system considering multiple perturbations.The main contributions are summarized as follows.A robust control strategy for the grid current and dc-link voltage of the single-phase grid-connected converter is developed to suppress the harmonics in grid current.A grid inductance estimation method using dual-harmonics injection and an adaptive quasi-proportional-resonant controller are designed for grid current control.For dc-link voltage control,a super-twisting sliding-mode controller is designed based on the state average model of the converter.Extensive simulation and experimental studies are presented to verify that the proposed parameter identification method accurately estimates the grid inductance,and the proposed robust control strategies greatly reduce the grid harmonics and improve the stability and robustness of the grid-connected converter control.A fault-tolerant control strategy based on Walcott-Zak sliding-mode observer(WZ-SMO)is proposed to improve the robustness of the single-phase grid-connected converter control under sensor faults.WZ-SMO for both grid current and dc-link voltage is designed to produce analytical redundancy.Normalized residuals are generated using measured and observed values.The fault diagnosis unit can detect and isolate sensor faults by comparing residuals with thresholds.The fault-tolerant control is realized by substituting the observed values for the information of faulty sensors.Simulation and experimental results indicate that the proposed fault-tolerant control can be implemented within 10 sampling periods of the control system and the stability of the converter under sensor faults can be guaranteed.Therefore,the proposed fault-tolerant control strategy is proved as an effective method to improve the robustness of converter control under sensor faults.A disturbance-observer-based control method and an adaptive robust control(ARC)method are designed for the control of rotor speed and stator current considering multiple perturbations.Therein,a lumped disturbance is synthesized to include all disturbances in speed control and estimated to compensate the control errors attributed to system perturbations.To improve the robustness of stator current control,this thesis develops an ARC algorithm,where relevant machine parameters are estimated by an integrated estimator to achieve adaptive model compensation,and the control errors induced by parameter mismatches are calibrated by a robust feedback regulation.The proposed controllers have the capability of achieving robust control of the speed and currents considering system perturbations,while simultaneously estimating several useful parameters of the PMSM.It can be concluded that the proposed algorithms greatly enhance the stability and robustness of the speed and current control under multiple system perturbations.A signal-injection-based robust MTPA operation strategy is developed for the PMSM drives.This method works by injecting a high-frequency low-magnitude signal into the current vector angle and observing the response in the current magnitude,followed by a proportional-integral controller that achieves the adaptive control of the optimal current angle.Extensive simulation and experimental results are presented to validate that the proposed method could achieve the tracking of the optimal current angle within 1s and the robustness and operating efficiency of the machine under multiple perturbations can be greatly improved.In this thesis,the parameter estimation and robust control of the single-phase grid-connected converter and PMSM are proved to be the key issues to improve the control performance of the whole PMSM-based railway traction system,and have been carried out with in-depth theoretical and experimental studies.The proposed control methods have the advantages of simple structure,easy adjustment,flexible design and low CPU occupancy rate,and can be easily implemented in commonly-used digital processors in electric traction applications.This research has a guiding significance for improving the robustness of PMSM-based railway traction systems,and a reference value for achieving high-quality high-efficiency operation of the high-speed EMU.