Study On Key Technologies Of Model Predictive Control For Doubly-fed Induction Generator

In recent years,in order to actively respond to the increasingly enviromental problems and energy crisis,the debelopment and utilization of renewable energy have been rapidly debeloped.As a kind of renewable energy,wind energy has many advantages such as being clean,non-polluting,and renewable.It has received extensive attention from most countries in the world.The installed capacity of wind power continues to increase.As the mainstream generator used in wind power systems,doubly-fed induction generator(DFIG)occupies an extremely important position in the entire wind power industry.Therefore,the development of high-performance DFIG control algorithms is particularly important.In this paper,the high-performance control of DFIG is studied,deeply.The basic operation principle under both balanced and unbalanced power grids are systematically analyzed and introduced.The simulation and experiments verify the effectiveness of the proposed algorithms.Although the traditional vector control(VC)algorithm has very good steady-state performance.However,its dynamic performance is slightly affected by the bandwidth of current loop and other delay factors.At the same time,the tuning work of the controller parameters is also difficult.This paper firstly analyzes the basic principle of direct power control(DPC).Then the double-vector-based method and space vector modulation technology are introduced into DPC,which greatly improves the steady-state performance of the system.Second,this paper focuses on the model predictive control(MPC).The MPC based on the traditional enumeration method is introduced.Then the problem of heavier calculation burden for the enumeration method is analyzed.After that,the vector selection and optimal application time calculation process are greatly simplified by the method based on the fast vector selection.On the basis of analysis above,this paper proposes a universal multi-vector-based MPC.This method profoundly reveals the intrinsic relationship between multi-vector-based MPC methods.Finally,the control of DFIG under unbalanced power grids is studied.The influence of grid imbalance on the DFIG is analyzed.Based on the traditional theory of instantaneous power,a simple power compensation strategy is proposed to avoid the separation of the positive and negative voltages and currents.The extended power theory is introduced to improve the traditional MPC method.The improved MPC and DPC-SVM methods successfully achieve the desired control target.Final,simulation and experimental platforms are built,separately.And the proposed algorithm is verified by the simulation and experiment results.Setting up a dSPACE-based DFIG experimental platform.According to the process of algorithm development,the simulation framework of the system is first established in Matlab/Simulink,and then the required code is automatically generated and directly downloaded to the MicroLabBox.This method is used for fast algorithm verification to avoid the repeated compilation and debugging of the algorithm in different environments,the efficiency of algorithm development is highly improved.