Study On Model Predictive Control Of The Double-Fed Wind Turbine

With the development of double-fed wind turbine towards high-capacity and high-performance, the research of wind turbine control focus on developing a reliable and optimal operation of the double-fed wind turbine control system. The most important part of this job is to achieve coordination between speed adjustments and increase the structural damping to achieve optimal load. In this paper, due to double-fed wind turbine to reduce dynamic loading and achieve efficient and stable control, a multi-states model predictive control strategy is proposed, to reduce the drivetrain vibration, out-of-plane deflection of blade and deflection of tower-top fore-aft, design torque and pitch model predictive control system, the paper innovative work carried out as follows:For the influence of uncertain parameters of double-fed wind turbine, this article identified the specific factors that affect the unit operation and quantify it, and selected a different state variable according to the different control objectives based on the state model 3. Give full consideration to the impacts on the whole turbine made by rotor rotational, drivetrain vibration, out-of-plane deflection of blade and deflection of tower-top fore-aft, established 7 states linear model for pitch controller and 3 states linear model for torque controller respectively.In order to restrict the dynamic load caused by wind disturbance to double-fed wind turbine, this article proposed a model predictive control algorithm to suppress wind disturbance and apply it to in the unit control. It considered the influences of wind disturbance in predicting future output and optimal control objective function; increased the calculation of wind disturbance; fully derived the calculation of increasing in wind disturbance rejection of double-fed wind turbine model predictive control algorithm, thus increased in the control system of wind disturbance feed forward compensation. In order to reduce load of the double-fed wind turbine drive train, blade and tower, establish the model of predictive controller with wind disturbance rejection, increase the damping of the following three parts:drivetrain vibration, out-of-plane deflection of blade and deflection of tower-top fore-aft using of pitch control. By selecting appropriate target weight matrix, adjust the emphases of damping for each degree of freedom. Additionally, for the wind speed near the rated wind speed, the problem of blade pitch angle on the drive train torsional vibration modes reduced controllability, adjust drive chain generator torque increases torsional vibration damping.To take control strategy to a single environment simulation technology to achieve union environment simulation, To overcome the problem that a single simulation environment cannot take into account the aerodynamic characteristics of rotor, mechanical properties and control strategy issues, It leaded into co-simulation concept, and for the first time established a double-fed wind turbine integrated environment for the joint motion simulation. Using AeroDyn software to calculate the aerodynamic, FAST software to calculate the mechanical dynamic, MATLAB software to control algorithm, In order to reduce the one-sidedness in simulation of a single environment. Three co-simulation took the development of the double-fed wind turbine into account, made the simulation more realistic, and facilitated the controller design and verification.The 3MW doubly-fed machine controller was tested in the factory. Combined with the introduced multi-state model, compared with variable gain PI control, model predictive control algorithm is effective in reducing the wind turbine drivetrain vibration, out-of-plane deflection of blade and deflection of tower-top fore-aft, verify the feasibility of the method used, improve double-fed wind turbine control performance.