An integrated neural approach for maximum power point tracking and electrical losses minimization of wind generators with induction machines

This paper presents an Induction Machine (IM) based wind generation unit with an integrated maximum power point tracking (MPPT) and Electrical Losses Minimization Technique (ELMT) based on the Growing Neural Gas (GNG) network and Discontinuous PWM (D-PWM). The target is the development of a highly efficient wind generator with high dynamic performance, able from one side to quickly track the maximum generable power, according to any variation of the wind speed, and from the other side to minimize the IM´s the and inverter´s losses. The proposed wind generator is based on a back-to-back power converter topology with two IGBT based VSIs. It improves a previously developed neural based MPPT, integrating in a unique algorithm both the MPPT the ELMT features. The proposed wind generation unit has been tested experimentally on a suitably developed test set-up. Results clearly show that the integration of the GNG MPPT and ELMT+D-PWM into the wind generator control permits the active power injected to the power grid to be increased from 9 % at high wind speeds up to 163 % at low wind speeds and, in general, to increase the average active power in a low wind speed profile of almost 50 % with respect to the previosly developed GNG MPPT.