Design of optimal PI controllers for doubly fed induction generators in wind turbines using genetic algorithm

This paper presents a design procedure based on evolutionary computation, more specifically on genetic algorithm (GA) to obtain optimal PI controllers to the static converter connected to the rotor of doubly fed induction generators (DFIGpsilas), in variable speed wind generation systems connected to the electrical grid. The converter of the DFIG has a protection system that monitors continuously the machine operation, emitting a blocking command when the limit of the rotor current is violated, due to transient disturbances occurring in the electrical grid. That implies in deactivating the DFIG control loops which affect negatively the system global controllability. This control action of the DFIG converter is accomplished by PI controllers, which gain adjustments are not a trivial task, due to the nonlinearities and the high complexity of the system. In this way an appropriate fitness function is derived to express the time domain evaluation of the DFIG with the objective to assure the DFIG continuous operation even under a fault condition and improve at the same time its transient behavior as compared with the formal methodology to design PI controllers using poles placement. The results obtained, confirm the efficiency of the proposed control design.