Research On The Control Strategy Of The Doubly Fed Induction Generator

This paper aims to research the double-fed wind generators grid-connected control strategy on rotor side,which made the stator voltage to reach the requirements of grid-connected and eliminate system disturbances.In order to improve the doubly-fed Wind Power Generator Control Strategy,the following work is done in this paper:1)Firstly,the basic principle of DFIG is introduced.The equivalent circuit of DFIG and energy flows of DFIG when it is in running states of super synchronous and synchronous are analyzed.The state equations are built according to the mathematical model of DFIG.3s/2s and 2s/2r are introduced to simplify mathematical models that are high order,strong coupling and nonlinear.2)By deducting the double-fed generator mathematical model and analyzing mathematical model before and after grid-connected,not only can we come to influences of the system steady state when the motor parameters were changed,but also can we know that the DFIG system structure would not change before and after grid-connected and only increase a large disturbance which would bring adversely affect on grid-connected moment.This paper designed a variable structure control systems and proof its stability by Lyapunov.In order to eliminate the vibration of variable structure system,this paper introduces the extended state observer.Through simulation results,in an ideal model conditions,stator voltage track on the grid voltage at 0.2s under the traditional PI control strategy and the tracking error is about 20 V.But under the variable structure control strategy based on extended state observer the stator voltage track on the grid voltage at 0.06 s and the tracking error is about 16 V,which almost realized the smooth current grid-connected in 0.5s.When the double-fed induction generator resistance parameter increased 60% and mutual inductance reduced 5% and other model parameters were kept constant,stator voltage track on the grid voltage at 0.2s under the traditional PI control strategy and the tracking error is about 34 V.But stator voltage track on the grid voltage at 0.07 s under the variable structure control strategy based on extended state observer and the tracking error is about 25.5V.It is proven that compared to the traditional PI control strategy,based on extended state observer variable structure control strategy not only reduces system regulation time,which made the system sate reach stability,but also eliminate the sudden disturbance.Even in the changes of motor parameters,based on extended state observer variable structure control strategy is also robust.3)This paper present the BP neural network self-tuning PI control of inner current loop and PI control of outer voltage on the grid-side converter.The simulation shows that DC bus voltage achieves stability in 0.07 seconds under BP neural network tuning PI control strategy.But in conventional PI control method,DC bus voltage achieves stability in 0.09 seconds and the overshoot is larger.When the load conditions was changed in 0.5s,compared to the conventional PI control,active power can track given power more fast under BP neural network tuning PI control strategy.It is proven that BP neural network self-tuning PI control can reduce the adjustment time and enhance system stability.