Perturbance Observer-based Nonlinear Adaptive Control Of DFIG-based Wind Turbine Systems

In this thesis,three adaptive control strategies are designed to solve the problem that the wind power system(WECS)based on DFIG has poor control performance when the working point is offset.The main work of this theses can be summarized as the following aspects:This thesis proposes an adaptive control strategy(PI-FLC),which combines proportional integral feedback control and feedback linearization control(FLC)based on differential geometry theory.The proposed control strategy for rotor-side converter is studied,in which current inner-loop controller is designed by feedback linearization control,while power outer-loop controller is designed using proportional integral feedback control.This control strategy can provide accurate and satisfactory control performance when the working point deviates,meanwhile also reduces the number of state variables needed in the feedback control.The simulation results show that compared with the traditional vector control strategy,the proposed PI-FLC strategy can better track the maximum power of the doubly-fed wind power generation system under slope wind speed changes and random wind speed changes,and has stronger robustness to parameter changes.This thesis presents a disturbance observer-based multi-loop adaptive control(POMAC)for doubly fed wind power system.Aiming at the disadvantage of feedback linearization control based on differential geometry theory whose performance depends on the exact value of state variables.At the same time,aiming at reduce the influence of modeling error and external disturbance on the system model,a multi-loop adaptive control method based on disturbance observer is proposed and applied to doubly fed wind power system.First of all,during the input/output linearization process of the grid side converter and the rotor side converter of DFIG-WT,the high-order system is decoupled into four low-order subsystem according to four corresponding control target.In each subsystem,a disturbance variables is introduced to describe the influence of the system model uncertainties and external disturbances on the system model.Then,the state variables and disturbance variables of each subsystem are observed by using the extended state and disturbance high gain observer.Finally,the linear optimal control is realized using the estimated value of the state variable and the disturbance variable instead of the real value,so as to realize the adaptive control of the whole system.This thesis proposes a disturbance observer-based flux compensation control method(OFCC)to improve DFIG-WT’s low-voltage ride through capability.Firstly,the influence of stator flux’s dynamics on the system is analyzed,and the importance of stator flux’s dynamics compensation in improving the capability of low-voltage ride through is illustrated.On the basis of POMAC,disturbance variables is used to describe the influence of stator flux’s dynamics on the system model,and then the disturbance variables are observed by using the extended state and disturbance high-gain observer.Finally,the estimated value of perturbation variables is used to compensate the stator flux’s dynamics in the feedback control.Simulation results show that the proposed OFCC can achieve the same low-voltage ride through performance as FFTCC,far better than the low-voltage ride through control performance of vector control.Moreover,different from FFTCC,which requires accurate flux calculation based on accurate model,OFCC uses observers to obtain the dynamic change of the flux,so it is strongly robust to the change of system parameters.This thesis builds a semi-physical testing platform based on dSPACE and RTDS.In order to verify the control algorithm,we built a semi-physical test platform by combining digital signal processing and control engineering(dSPACE)1006 processor board with the real-time digital simulator(RTDS).The wind power generation system model composed of DFIG,transformer,load and power grid is established on RTDS,while the traditional vector control and the proposed adaptive control strategies are realized on dSPACE,and they interact with each other through A/D and D/A.The experimental results show that the proposed adaptive control methods can achieve the control objectives well,and has a faster adjustment speed and a smaller fluctuation.In addition,because the POMAC method use observers to estimate the disturbance and compensate it in the feedback control,it does not need accurate system model and has strong robustness to system parameters.The proposed OFCC method can compensate the system disturbance caused by stator flux’s dynamics,so as to suppress the rotor over-current after fault and improve the low-voltage ride through ability of DFIG-WT system.