Nonlinear adaptive control of permanent magnet synchronous generator based wind turbine: a perturbation estimation approach

This thesis deals with the modeling and control of permanent magnet synchronous generator based wind turbines (PMSG-WTs). The PMSG-WTs are widely used in wind energy conversion systems(WECSs), due to their merits such as high reliability, high efficiency, low noise, high torque to weight ratio and fast dynamic response. Usually, a PMSG-WT is connected to the power grid via an AC-DC-AC converter system. The PMSG-WT can rotate at varying speed based on variable wind power input and thus achieve high efficiency as it dose not need to synchronise its rotational speed with the grid frequency. An overview of the modeling of the PMSG-WT is give at first, with conventional vector control (VC) strategies applied for machine-side and grid-side converter. The VC strategy is a popular method widely used in industry due to its decoupled control of active/reactive power, but it may not provide satisfactory performance for the PMSG-WT as it is required to operate at varying speed in an operation envelope with wide operating range rather than one operation point. The feedback linearisation control (FLC) strategy can improve the performance of the VC with a global optimal controller crossing a wide region and variable operation points, but it has weak robustness against parameter uncertainties and external disturbances, and requires full state measurements. To improve performance of the VC and the FLC, nonlinear adaptive controllers (NACs) designed based on FLC and perturbation estimation and their applications in machine-side and grid-side converter control of the PMSG-WT, and speed control of a permanent magnet synchronous motor (PMSM) have been studied. In the design of the proposed NAC, by defining a lumped perturbation term to present coupling nonlinear dynamics, parameter uncertainties, and other unknown disturbance, then a perturbation observer is designed to estimate the perturbation which is used to compensate the real perturbation and realise an adaptive linearising of the original nonlinear system, without requiring the accurate system model and parameters and full state measurements, and still considering all system nonlinearities and unknown time-varying dynamics, such as tower shadow, grid faults and intermittent wind power inputs. In this thesis, the proposed control schemes are applied for control of PMSGWT in Region 2, Region 3 and integration with the grid. A NAC is developed for a PMSG-WT to extract maximum wind power in Region 2. Simulation and experiment studies are carried out to verify the design and results show that the proposed NAC can provide better performance in MPPT and robustness against parameter uncertainties and time-varying wind power inputs, in comparison with a convention VC and FLC. NACs are designed for control of the pitch angle and generator control of a PMSG-WT to limit the extracted power from time varying wind in Region 3. Simulation results of the proposed NACs are compared to a conventional VC and FLC. The fault ride-through capability (FRTC) of the PMSG-WT at different voltage dip’s levels has been enhanced by a novel NAC applied at the grid-side converter. Simulation results have shown that the proposed NAC can provide satisfactory performances with smaller inrush current and voltage overshoots during grid fault and better robustness against uncertainties. A coordinated nonlinear adaptive control (CNAC) of the machine-side and grid-side converter in the PMSG-WT were studied. The NACs are designed based on state and perturbation observers for control of subsystems. Simulation results show that the CNAC can coordinate each other to achieve the objectives of different operating regions and enhance the FRTC of the PMSG-WT. Finally, the proposed control schemes are applied for control of PMSM. NAC is developed for PMSM to track mechanical rotation speed and provide high robustness against system parameter uncertainties and unknown time-varying load disturbances. Simulation results show that the proposed NAC provides better performance and robustness against system parameter uncertainties and unknown time-varying load disturbances, in comparison with a nonlinear controller with an extended nonlinear observer and a conventional VC