Design and Analysis of an Artificial Intelligence-based Rotor Current Control Strategy for a Doubly Fed Induction Generator

Abstract

A 2 MW doubly fed induction generator (DFIG) wind turbine is used in this thesis as the plant of the overall system, and the mathematical modeling of the wind speed, aerodynamics, mechanical component, and electrical component of the wind turbine along with the appropriate control mechanisms are incorporated. The dynamic features of DFIG mainly depend upon rotor current, stator current, and stator flux. These three parameters of DFIG increase the overall system complexity. One of the DFIG parameters is rotor current, which needs to be controlled because the output current of the wind turbine has greater overshoot, very large settling time, and large steady-state error, which reduce the output power efficiency and affect the system’s stability. To reduce such problems the design of a reliable and high-performance controller is required. Hence, the design and analysis of rotor current control strategy have been applied to the rotor side converter of a doubly fed induction generator in this thesis work as a solution to the shortcomings of the DFIG parameter. The conventional control design, such as a proportional-integral (PI) controller has many shortcomings, such as difficulties in setting parameters and reduced robustness. Thus, the fuzzy logic, fuzzy-PI, and artificial neuro-fuzzy inference system (ANFIS) controllers are designed to regulate the rotor current loop of DFIG in the rotor side voltage source converter. By applying a vector control strategy, the rotor-side converter (RSC) is controlled to regulate the generator rotor speed, torque, and rotor currents, since the conventional controller could not satisfy the required system performance of the rotor currents. The designed systems are simulated in MATLAB/Simulink. The performance of the system subject to PI, fuzzy, fuzzy-PI and ANFIS are presented, and the result shows that the system used with artificial intelligent controllers show the best performance compared to the conventional PI controller in some performance index. When the objective of control is steady state error (SSE) of the system’s rotor current, the SSE of the system used with; PI is 2.9084 A, fuzzy,0.8668 A, fuzzy-PI, 7.654A, ANFIS, 11.5472. The fuzzy controller based system is found to be the best candidate for SSE control of rotor current. The ANFIS controller based system has the best settling time for both quadrature axis and direct axis rotor currents, and the fuzzy-PI controller experiences the best performance from the perspective of steady-state error and settling time for torque control purposes.