Research On Model-Free Adaptive Control Of Wind Turbine Pitch System

With the depletion of traditional fossil energy,wind energy,as one of the main renewable energy sources,has achieved great development in the past decades.However,due to the influence of random wind speed disturbance,wind energy becomes an unstable power source in grids.In addition,there are wind shear,tower shadow,wake,and turbulence in the wind field,which lead to serious asymmetric loads on wind turbines.Thus,variable speed variable pitch technology has become a means to effectively adjust generator power and alleviate wind turbine load.The traditional pitch control strategies mostly depend on the accurate model information of the controlled system.However,wind turbines are faced with random wind speed disturbance and have the characteristics of large inertia,strong nonlinearity,and coupling,which makes it difficult to establish an accurate mathematical model.Meanwhile,most of the traditional methods are based on the linearized model on the specific working point of the pitch system,which causes a significant decline in the control performance when the random wind speed fluctuates in a large range and deviates from the design point.Moreover,a good pitch control effect depends on the frequent action of pitch actuators,which greatly increases the risk of actuator fault.Therefore,further research on the fault-tolerant control strategy of the pitch actuators is also of great significance to improve the robustness of the system.Consequently,this dissertation uses the online I/O data of the pitch system to design the pitch control strategies of the wind turbines and realize the fault-tolerant control scheme of the actuators without relying on any mathematical model information of the controlled system.The main work is as follows:(1)Based on the widely used collective pitch mechanism,a gain scheduled proportional-integral(GSPI)control strategy is designed.The simulation test platform of the controller and NREL 5MW wind turbine are built by using Simulink/FAST,and the comprehensive performance test of the GSPI is carried out.The results can be used as the benchmark for subsequent research.(2)Since the insufficient nonlinear response of the GSPI and the dynamic performance of the existing MFAC based on compact form dynamic linearization(CFDL-MFAC)is unable to meet the control requirements of the pitch system,an improved model-free adaptive collective pitch control strategy(DE-MFAC)is proposed by considering the change rate of tracking error in the cost function.The convergence of the tracking error and the BIBO are proved by using the contraction mapping method.The simulation results show that compared with GSPI and CFDL-MFAC,the average fluctuation amplitude of motor power and speed under DE-MFAC can be controlled within 1% and 2% respectively,which achieves better steady-state performance and dynamic response.(3)Aiming at the collective pitch control(CPC)can not relieve the asymmetric load of wind turbines,and the single variable individual pitch control(IPC)strategy is unable to completely remove the coupling between tilt and yaw components of wind turbines.An IPC strategy based on MIMO-DE-MFAC has further proposed in this paper..The simulation results under different working conditions show that compared with the GSPI CPC strategy,the average short-term damage equivalent load under MIMO-DE-MFAC can be reduced by 17.55%,5.02 percentage points more than DE-MFAC IPC and the optimal control effect is achieved under all verified wind speed conditions.(4)The above pitch control strategies are only designed for normal working conditions,and there are no corresponding fault-tolerant means for sensor/actuator faults.To effectively reduce the risk of controller instability in the case of actuator fault,based on the analysis of the operating characteristics of the pitch system under various actuator faults,the fault-tolerant IPC scheme based on multi-blade coordinate transformation is proposed for the first time.Then,a dual MIMO-DE-MFAC fault-tolerant individual pitch control(FTIPC)strategy is designed based on the scheme without any parameter estimation,fault detection,and isolation process.The simulation results show that under the conditions of actuator additive fault,multiplicative fault,and multiple faults,the load control performance of the main measuring points on the tower base/top decreases by no more than 0.23%,which is almost the same as that under the condition of no-fault and effectively improves the robustness of individual pitch control.(5)To further verify the effectiveness of the proposed methods,the hardware-in-the-loop experiment platform of the 5MW wind turbine controller is built using Simulink Desktop Real-Time and SIMATIC S7-1200 PLC.The developed GSPI CPC,DE-MFAC CPC,MIMO-DE-MFAC IPC,and dual MIMO-DE-MFAC FTIPC are transplanted into PLC for the hardware-in-the-loop experiment.The experimental results not only further verify the reliability of the above simulation results but also prove the practicability of the proposed methods in real hardware conditions.