Nonlinear Autoregressive Neural Network Approaches for Managing Active and Reactive Power in DFIG Systems

The effective command of the mechanical and electrical components of a wind turbine is essential to secure optimal efficiency and stability of the system. This article aims to present a novel Nonlinear Autoregressive Neural Network (NARNN) strategy for controlling the electrical aspect of a system employing a Doubly-Fed Induction Generator (DFIG). The control strategy is designed to regulate active and reactive power in order to optimise energy production. To generate a reference power signal, rotor speed control is implemented in the mechanical part of the system. The results provided by the presented NARNN control strategy are then compared with those obtained from the reference Proportional Integral (PI) controller.