Research On Control Method Of Offshore Floating Wind Turbine

Compared with onshore wind energy,offshore wind energy has many advantages,but because the fixed wind turbine support structure can only be installed in shallow coastal areas,the technology is limited by the depth of the water.By installing wind turbines on floating structures,offshore floating wind turbines can be formed.Floating offshore wind turbines(FOWT)are a new technology,which allows the A large number of high-quality wind energy resources are explored in deep offshore areas.Early studies have shown that the platform negative damping effect of offshore floating wind turbines may increase the load of the unit.Therefore,the control method plays an important role in the dynamic behavior and performance of the wind power system.At present,the goal of control for high wind speed regions of floating wind turbines has begun to achieve multi-objective control while satisfying all the constraints of the actuators.A common multiple goal is to stabilize the output power of the unit while reducing the fatigue load.At this time,PI controllers commonly used in the industry may become invalid or difficult to tune.Model predictive control(MPC)provides a possibility for the realization of multi-objective control.This paper takes the floating wind turbine of the tension leg platform as the object,and studies how to use MPC control to achieve effective balance of power generation and reduce the structural load of the floating wind turbine,which are two different potential conflicting control methods.In the article,firstly,the combination of Simulink-FAST was realized by using professional software,and the simulation environment of the tension leg type floating wind turbine was built.And designed a unified pitch and independent pitch controller.Among them,the independent pitch controller is designed based on the fuzzy PID principle and can output three different pitch angle commands.In the simulation environment,the effects of the two controllers are compared,and the performance of the two controllers in stabilizing power and reducing tower load is analyzed.Then,a wind speed prediction study was conducted.Because,in order to achieve model predictive control of floating wind turbines,at least wind speed information at multiple times in the future needs to be obtained.In this paper,long and short-term memory networks are used,and two wind speed multi-step prediction models are established according to different training rules: h-1 and h-N prediction models.And through experiments,the prediction accuracy and error trend of the two prediction models under different prediction steps are studied.Finally,using the selected multi-step forecasting model of wind speed,a forecasting controller for floating wind turbines is designed.Using the floating wind turbine simulation software,linearized calculation is performed to obtain the unit state space model as the predictive model of predictive control.Then the model predictive controller was established,and its performance in reducing unit load and stabilizing output power was verified by comparison.It further explores the influence of the predicted step length on the predictive control performance of the model,and obtains the best wind speed prediction step length.