A neural network approach for the direct power control of a doubly fed induction generator

This paper presents a direct power control strategy for a doubly fed induction generator using a Multilayer Perceptron controller. The controller generates the direct- and quadrature-axis rotor voltage signals without the necessity of current controllers. The proposed control strategy allows the converter that is connected to the rotor terminals to operate with constant switching frequency. Digital simulation results are performed for a 2MW DFIG controller. With this controller, experimental and digital simulation tests for a 2.2 kW machine running in a subsynchronous speed are presented to validate the proposed control.