Optimized MPPT and transient suppression for DFIG-WECS outer-loop speed regulation using hybrid ANN-backstepping

This paper focuses on maximizing energy harvesting in variable-speed wind turbines by improving maximum power point tracking (MPPT). A hybrid control strategy combining an artificial neural network (ANN) with nonlinear backstepping control (BS) is proposed for the outer-loop speed regulation of a 1.5 MW doubly fed induction generator (DFIG). The novelty lies in coupling the fast nonlinear mapping capability of the ANN with a Lyapunov-based backstepping corrective mechanism to improve transient tracking. Simulations show that the ANN–BS controller achieves a settling time of 1.2 ms, a speed-tracking error of 0.08%, and a dynamic overshoot of 1.09%, outperforming the PI and standalone ANN controllers. Under a measured stochastic wind profile, it achieves a 0.70 ms startup settling time, reduces error-chattering intensity to 0.324, and maintains the power coefficient close to 0.48, confirming accurate MPPT tracking under variable wind conditions.