Abstract

A sensorless vector-control strategy for an induction generator in a grid-connected wind energy conversion system is presented. The sensorless control system is based on a model reference adaptive system (MRAS) observer to estimate the rotational speed. In order to tune the MRAS observer and compensate for the parameter variation and uncertainties, a separate estimation of the speed is obtained from the rotor slot harmonics using an algorithm for spectral analysis. This algorithm can track fast dynamic changes in the rotational speed, with high accuracy. Two back-to-back pulse-width-modulated (PWM) inverters are used to interface the induction generator with the grid. The front-end converter is also vector controlled. The dc link voltage is regulated using a PI fuzzy controller. The proposed sensorless control strategy has been experimentally verified on a 2.5-kW experimental set up with an induction generator driven by a wind turbine emulator. The emulation of the wind turbine is performed using a novel strategy that allows the emulation of high-order wind turbine models, preserving all of the dynamic characteristics. The experimental results show the high level of performance obtained with the proposed sensorless vector-control method.