Search

A new strategy for controlling voltage and frequency of a self-excited induction generator (SEIG) is presented. An external excitation circuit, including a power-switched inductance/capacitor, is used to compensate for the reactive demand. The conventional dynamic modeling of this system is enhanced by using an artificial neural network (ANN) to model the induction generator. The proposed ANN model is used to obtain the inverse model of the SEIG. The network is trained using the output voltage, the load and the wind turbine speed as an input vector and the required capacitor bank as the target output. The obtained inverse model is then cascaded with the SEIG. The active filter is used to replace variable capacitors and plays an important role and gives good dynamic response and robust behavior upon changes in load and generator parameters. Thus, the network combined with the active filter acts directly as a controller. Computer simulations are used to demonstrate the validity of the proposed scheme and the behavior of the SEIG for both voltage buildup and wind speed disturbance conditions.