Semi-active Control For Wind Turbine Tower-blade Coupling Structure

Since the 21st century,china’s wind turbine industry has developed rapidly.Due to the impact of the working environment of wind turbines,the blades of large wind turbines are prone to bend and fracture due to excessive wind loads.In severe cases,the blades are bend and collide with the tower,resulting in wind turbines,causing serious economic losses and casualties.Therefore,reducing blade damage and prolonging the service life of the wind turbines become an important part of ensuring the performance and safety of the wind turbine.Under the condition that the stiffness and flexibility of the blade material are limited,reducing the vibration of the blade has becomes an important direction of our research.The tower-blade structure of wind turbine will produce coupling vibration,which has a greater impact on the overall stability of the wind turbine.In order to study the coupled vibration and coupled vibration control technology,a variety of software joint modeling is used to conduct a parametric simulation analysis of the large wind turbine tower-blade coupling structure to determine the weak parts of the blade vibration.The structure-oriented control method and BP neural network are introduced to control the wind turbine tower-blade coupling vibration,and the new type of damper is taken as the key equipment of blade vibration control,which can give consideration to both economy and practicability.NREL 5MW turbine wind turbine is a relatively mature wind power equipment in both technology and application at present.Therefore,this paper establishes a 5MW wind turbine model example,and uses LQR,quasi-sliding mode and BP neural network control strategies to simulate vibration control.The main work content and effective results are as follows:1)Solidworks software is used to draw the sketch of airfoil and complete the assembly of the wind turbine tower-blades coupling structure.The low-order vibration modes of the model were analyzed by finite element software.The simulation results show that the blade vibration mainly includes in-place vibration and out-of-place vibration.According to the mode diagram,the structural weakness of the wind turbine blade is determined to be 0.25L from the tip of the blade,which is determined as the installation position of the damper.The mathematical model of the coupling structure of the wind turbine is established by Euler-Lagrange method;2)According to the working principle and structure of the new damper,the corresponding parameters are selected.the composition structure and transmission mode of the damper are introduced,the mechanical properties of the new damper are analyzed and calculated,and the general mechanical model of the damper is established;3)Because LQR and quasi-sliding mode control algorithms are commonly used classic control algorithms,and they have achieved relatively good control effects in practical applications.The LQR control and quasi-sliding mode control are programmed through MATLAB software,and the dynamic model and semi-active control algorithm of the wind turbine tower-blade coupling structure are established.The simulation analysis obtains the vibration of the coupling structure under uncontrolled and two control strategies.Response result.The comparison shows that both LQR and quasi-sliding mode control have good control effects,and the corresponding semi-active control can also effectively suppress blade vibration.4)Due to the uncertainty of the wind turbine tower-blade coupling structure parameters and the complexity of the model to be simplified,etc.,the assumed modal method is used to model and simulate the wind turbine.The calculated data is somewhat different from the data measured in the actual project.error.Considering that the BP neural network establishes an Actual coupling structure model through actual expected data,the BP neural network algorithm is introduced to control the vibration of the wind turbine tower-blade coupling structure.Use MATLAB software to program the BP neural network,and train the BP neural network with the expected wind-induced vibration response data of the wind turbine tower-blade coupling structure to obtain the neural network model.5)Davenport wind load,wind speed and other related influence parameters were loaded into the BP neural network uncontrolled model to calculate the blade vibration control results.The simulation analysis shows that the vibration control effect of the BP neural network is better than that of the LQR and quasi-sliding mode control strategies,and the semi-active control effect is also relatively good.The robustness analysis found that the robustness of the BP neural network is better than the other two control strategies.When the stiffness changes,only the BP neural network control strategy recognizes the stiffness change of the coupling structure and adjusts the control,so that the vibration control rate of the wind turbine tower-blade coupling structure is still dynamically improved when the stiffness decreases.