H. Benbouhenni - Labo. LAAS, Departement de Génie Electrique, Ecole Nationale Polytechnique d’Oran Maurice Audin, Oran, Algeria.

In this paper, a novel switching table (ST) of the twelve sectors direct power command (DPC) strategy of doubly-fed induction generator (DFIG) based dual rotor wind power (DRWP) is proposed using two-level hysteresis controllers for reactive and active power and feedforward neural networks (FNNs) algorithms. This intelligent technique was used to replace the conventional ST in order to reduce rotor flux ripple, active power ripple, total harmonic distortion (THD) of stator voltage, torque and reactive power undulations. The simulation and modeling of the proposed strategy were carried out in Matlab software. The DFIG is tested in association with a DRWP systems. The simulation results show that the DPC with FNN controller (DPC-FNN) reduced the THD value of stator voltage, rotor flux undulation, active/reactive power undulation, and electromagnetic torque ripple compared to conventional DPC strategy. It was found that the current waveform becomes purely sinusoidal with a reduction in the THD rate to ۰.۶۴%.

DPC, DFIG, THD, FNNs, DRWP