Semi-active Structural Control For Load Mitigation Of Floating Wind Turbine Based On ARL Algorithm

With the rapid development of wind power industry in recent years,wind power technology has gradually shifted from land to sea where wind resources are more abundant.Floating wind turbines are designed to capture deep-sea wind power,which is developed by land-based wind turbines.However,the environment on the sea is relatively complex.Besides the influence of wind,the wave also has a great impact on the turbines.This poses a high challenge to the installation of wind turbines.At present,floating wind turbines mainly rely on the floating foundation and anchoring system to ensure the floating in the fixed sea.However,the structure of floating wind turbine is flexible.That makes it produce large structural loads under the action of wind and waves.Therefore,advanced structural load control theory and technology are needed to support the development of floating wind turbine.In this view,domestic and foreign scholars have conducted a lot of related research.The research on the structural control of the wind turbine is also progressing steadily,and the research results are published continuously.However,most of the structure control research are done for onshore wind turbines or stationary ones,and the research results on the structural control of floating wind turbine are relatively few.Moreover,in the study of the structural control of floating wind turbine,it is mainly concentrated in passive structural control and a few active structural control.Semi-active structural control is rare.Therefore,this paper studies the semi-active structural control of floating wind turbine based on magnetorheological damper.The main research contents are as follows.(1)This paper introduces the structural control,including passive,semi-active and active structural control.Moreover,the applications in wind turbine also are introduced and compared.This paper introduces the principle of different types of structural control and displays the specific mechanical control parts.(2)The magnetorheological(MR)damper is adopted in the tuned mass damper for the semi-active structural control of the wind turbine.The magnetorheological damper can play an important role in the structural control of floating wind turbine because its response is fast and it’s easy to control.The damping force can be controlled in a certain range by the voltage or current added to it.In this paper,the working principle of the magnetorheological damper is explained.Several typical mathematical model of MR damper are selected for introduction and the nonlinear characteristic of the magnetorheological damper is presented.(3)This paper presents a fixed batch adaptive regularized learning(ARL)algorithm.This is a machine learning algorithm which can fit the nonlinear characteristics well without knowing the specific parameters of MR damper.In this way,the inverse analysis system of the magnetorheological damper is established,and its hysteresis characteristics can be accurately described by it.This article has carried on the improvement on the basis of this algorithm.The weighted fixed batch ARL algorithm is proposed.The data of fixed batch are weighted and historical data which are too early,are abandoned.That embodies the thought that the newer the data the better it reflect the current situation.(4)In this paper,a linear model of a barge type floating wind turbine in the pitch plane is established.Based on this model,we use the model following control method of model reference adaptive control to make the real wind turbine tracking the reference model.Then the adaptive algorithm is adopted to obtain the damping force for structural control and the degeneration controller is designed to obtain the control signal that can be realized on MR damper.In this paper,5MW wind turbine of the national renewable energy laboratory are used to simulate the control objects and good results are obtained.