Adaptive Model Predictive Control of Dual Output Three-Level Inverter

This paper proposes an adaptive model predictive control (A-MPC) strategy for a three-level dual output voltage (TLDOV) converter. The proposed control addresses the growing demand for robust, multi-objective control solutions in multi-output multilevel inverters. It has a fast transient response, straightforward implementation, and can handle non-linear constraints like finite control set model predictive control (FCS-MPC). The proposed A-MPC is immune to performance degradation due to sensitivity to system model inaccuracies and parameter variations, which impacts the performance of FCS-MPC. A mathematical model of the TLDOV converter is developed, and A-MPC is utilized to control two output voltages independently while maintaining DC-link voltage balance, irrespective of load fluctuations. The system's performance is validated through Matlab/Simulink in constant and dynamic frequency conditions, with its robustness tested against sudden load changes.