Abstract
This paper proposes a computationally efficient sensorless control framework for a permanent magnet synchronous generator (PMSG)-based wind energy conversion system (WECS) employing three-level neutral-point-clamped (3L-NPC) back-to-back converters. A high-gain observer is developed to estimate grid voltages, frequency, and electrical angle, with its estimation error analyzed using Lyapunov theory to establish uniform ultimate boundedness. An integral sliding mode control (ISMC) combined with finite-control-set model-predictive control (FCS-MPC) is employed to regulate the system. To reduce the computational burden, a two-stage search algorithm is proposed, reducing the number of evaluated switching combinations from 27 to 7 per sampling period while maintaining comparable control performance. The proposed strategy is evaluated through Processor-in-the-Loop (PIL) simulations using MATLAB/Simulink and the LAUNCHXL-F28379D DSP platform. The observer is tested under sudden variations in grid voltage amplitude, abrupt frequency changes, frequency ramps, and measurement noise. The results show that the estimated variables converge toward their actual values, with increased fluctuations under measurement noise. The control strategy is evaluated under wind-speed variations and parametric uncertainties with a stochastic wind profile. Under nominal conditions, both FCS-MPC methods show comparable control performance, while the proposed method requires less execution time. Under parametric uncertainties, the proposed method improves DC-link voltage regulation and neutral-point voltage balancing, achieving a grid-current THD of 0.99%, compared with 2.43% for conventional FCS-MPC.