Research On Predictive Variable Pitch Control Technology Of Wind Turbine

The excessive exploitation of fossil energy in modern society has caused a series of environmental pollution problems.Wind energy,as a clean energy,is becoming more and more popular in various countries with its remarkable social benefits.Due to the fluctuation and intermittency of wind speed,and the characteristics of large inertia and non-linearity of wind turbine,the output power control of wind turbine has brought challenges.With the rapid development of wind power industry in recent years,great progress has been made in wind power equipment manufacturing and control technology.Compared with fixed-speed and fixed-pitch wind turbines,variable-speed and variable-pitch wind turbines have incomparable advantages in wind energy conversion efficiency and power control.It has gradually become the mainstream of the market.In this thesis,the variable speed and variable pitch wind turbine is taken as the research object,and an effective power control strategy for wind turbine is proposed in order to achieve the goal of suppressing the power output fluctuation without adding additional equipment.Firstly,this thesis analyzes the working principle of wind speed and each subsystem of wind turbine and builds a simulation experimental model,which lays a foundation for the wind speed prediction and variable pitch control.Due to the difficulty in predicting the variation of wind speed and the complexity of wind turbines,it is difficult for traditional control strategies to achieve better control effects.As a meta-heuristic algorithm,firefly algorithm has obvious advantages in solving nonlinear system optimization problems.Therefore,this thesis introduces and improves firefly algorithm,designs parallel structure and four communication strategies,and proposes distributed parallel firefly algorithm(DPFA)with faster convergence speed and higher convergence accuracy.The algorithm proposed in this thesis is compared with other algorithms under several test functions to verify the convergence performance of the improved algorithm and provide effective algorithm support for the wind speed prediction model and controller design.Secondly,in view of the back propagation neural network(BPNN)of wind speed time series prediction of problem,this thesis introduce empirical mode decomposition was carried out on the wind speed time series decomposition.DPFA algorithm was used to optimize the BP neural network’s initial weights and thresholds,finally the superposition of the components in the heavy constitute the final prediction results.For large variations in the wind,the variable pitch actuator is not timely due to its large inertia,and feedback control signal has the lag problem,this thesis puts forward predictive feedforward-feedback control strategy.The predicted value as reference of feedforward controller,using predictive feedforward controller to compensate the control signal of the feedback controller,so that the variable pitch actuator can change the pitch angle one step in advance and achieve the goal of smooth output of wind turbine power.Finally,the algorithm and control strategy proposed in this thesis are applied to megawatt wind turbine and compared with the traditional control strategy.The experimental results show that the predictive feedforward and feedback control strategy proposed in this thesis has a good control effect,which can effectively smooth the power output of wind turbine,improve the stability of wind turbine and reduce the impact on power grid.