AI-Assisted Active Mixed Control for Performance Optimization of an On-Grid PMSG-Based Wind Energy Conversion System

Abstract

Objective: A The integration of wind energy into modern power grids is significantly challenged by the highly stochastic and nonlinear nature of wind speed variations.

Methodology: To address these issues, this paper proposes an artificial intelligence (AI)–assisted advanced active mixed control framework for the optimal operation of an on-grid Permanent Magnet Synchronous Generator (PMSG)–based Wind Energy Conversion System (WECS). The proposed control architecture combines a non-singular terminal sliding mode controller (NT-SMC) with a smooth switching active sliding mode observer (ASMO) to achieve robust regulation and intelligent state estimation under uncertain operating conditions. The NT-SMC is employed to ensure fast convergence and accurate tracking of system states, while the ASMO provides AI-assisted estimation of rotor speed and electromagnetic torque, thereby eliminating the need for physical sensors and reducing system complexity.

To ensure stable grid integration, a dedicated DC-link voltage management strategy is implemented, along with a feedback-based power control mechanism that regulates the active power injected into the utility grid.

Findings: The AI-enabled observer enhances disturbance rejection and adaptive decision-making by continuously learning system behavior in the presence of model uncertainties and external perturbations. MATLAB/Simulink-based simulation studies conducted under variable and rapidly changing wind conditions demonstrate that the proposed control strategy achieves superior set-point tracking, enhanced system stability, and strong robustness compared to conventional control schemes.

Novelty: The results confirm that the AI-assisted mixed controller provides an effective and reliable solution for intelligent, sensor-reduced, and grid-compliant wind energy conversion systems.