Parametrically Robust Dynamic Speed Estimation Based Control for Doubly Fed Induction Generator

Abstract

This paper presents a speed estimation based vector control architecture for doubly fed induction generator (DFIG). The main advantage of the proposed architecture is that with this methodology the generator can be operated without a speed sensor and position encoder. The method calculates the machine parameters online using a recursive least square (RLS) technique based on identifying the transfer function relating to rotor speed and position error. Minimum variance controller (MVC) ensures that the position error is minimum by proper estimation of the speed. For illustration, first, the small signal model of the machine and the regulator design is discussed. Then, a methodology is proposed, in which machine's mutual inductance is estimated online, so that the estimation approach is robust to changes in the machine parameters. Second, the control architecture with the speed estimation technique is discussed. The proposed approach is validated using a real-time simulation platform for a GE 1.5 MW wind turbine (both for steady-state operations and grid disturbance conditions) and with hardware in the loop (HIL) experimental setup for a 2-kW doubly fed induction machine (DFIM). Simulation results demonstrate desired steady-state and dynamic performance of this sensorless control approach for DFIG-based variable-speed wind turbines.