Abstract:
Wind energy is a renewable and highly efficient power source, capable of converting
approximately 40–50% of the wind’s kinetic energy into electrical energy.
Wind energy extraction is primarily governed by two control stages: turbine control
and generator control. In the turbine control stage, extraction efficiency
depends on accurate Maximum Power Point Tracking (MPPT) under varying
wind conditions. Conventional MPPT techniques rely heavily on precise turbine
characteristics, which can lead to time delays and reduced dynamic responsiveness.
This work proposes an artificial intelligence (AI)-based MPPT scheme that
eliminates dependence on empirical equations and extensive simulations. The
proposed approach directly predicts the optimal rotor shaft speed using an Artificial
Neural Network (ANN). Furthermore, an AI-based wind forecasting model is
incorporated to enhance predictive control and improve MPPT accuracy. In the
generator control stage, an Interior Permanent Magnet Synchronous Generator
(IPMSG) is employed to achieve high efficiency. The absence of rotor windings
in the IPMSG eliminates rotor copper losses, thereby improving overall efficiency.
Efficiency is further enhanced using a Maximum Torque Per Ampere (MTPA)
control strategy, which minimizes stator current for a given torque demand and
reduces stator copper losses. Field-Oriented Control (FOC) is implemented to
achieve precise regulation of the stator current components in the d-q reference
frame. Clarke and Park transformations are used to convert three-phase stator
currents into d-q components and vice versa, enabling independent proportionalintegral
(PI) controllers for the d-axis and q-axis currents. A slower PI speed
controller, operating at a bandwidth ten times lower than the current control
loop, determines the required torque for speed regulation. The performance of
the complete system is validated through computer simulations. In addition, a
laboratory-scale prototype incorporating an IPMSG and a power electronic converter
is developed to experimentally verify the power conversion in the system.
