Research On Model Predictive Direct Power Control Strategy For Wind Turbine

Wind energy is the renewable energy source with the best potential and application prospect,therefore,the rapid development of wind power industry has become the consensus of the Chinese government,enterprises and research institutions.China is rich in land-based wind power resources,and double-fed induction generators(DFIGs)are the mainstream of the world's recognized land-based types of generators.The dissertation focuses on the operation control technologies of DFIG.The rotor-side and grid-side converters of DFIG are investigated.The direct power control techniques of the converters are studied in depth from theory to practice and from simulations to experiments.The main work and innovations of the dissertation are as follows.1.The mathematical models of instantaneous direct power control of DFIG and grid-side current converter are established,which lay the foundation for the study of control strategy of DFIG.2.Aimed to the poor stability of traditional hysteresis direct power control,a direct power control technique for finite model prediction of DFIG based on rotor coordinate system is proposed.The PARK transformation for the rotor voltage,the measurement for the rotor current and the sensor with rotor current are not required in the method.Hence,the design of the system can be highly simplified and the hardware cost can be saved.Simulations and experiments are implemented and a much better control performance can be obtained compared with traditional hysteresis direct power control.3.Aimed to the hysteretic response to perturbation mutation in model-predictive control method,combined with the traditional sliding control,self-adaptive control and model-predictive control,the self-adaptive model-predictive method for direct power control is presented.With the conservation of the robustness of traditional sliding prediction control,the dynamic performance,steady-state performance and the antiinterference ability of the control system are enhanced in the recommended method.The effectiveness of the method is verified in the simulations.4.Aimed to the heavy computation burden of the conventional model-predictive control,integrated with explicit model-predictive control algorithm,a single-step explicit model-predictive method for direct power control with perturbation resistance is proposed.Based on the derivation of an optimal explicit control law for state feedback on offline computational state partitioning,calculations for the modelpredictive control method can be decreased greatly and perturbation resistance of the system can be enhanced as well.Comparison is made in the simulations from which the validity of the method is confirmed.5.To fulfill the control of DFIG without speed sensor,aimed to the parameter perturbation in the operation of DFIG and the high sensitivity of ensemble Kalman filtering control algorithm to uncertain model parameters,a new desensitized and integrated filtering control algorithm is proposed.Applied to DFIG,the state observation accuracy of DFIG is promoted and the self-adaptive model-predictive direct power control for DFIG without speed sensor is realized.Simulations are conducted to prove the availability of the method.6.The experimental platform of DFIG is built,the control program of rotor-side and grid-side converters for DFIG is accomplished and the experimental verification of the part control strategy is achieved on the presented experimental platform.