Real-Time Validation of Quadratic Regression Model-Based Predictive Control of Grid-Side Converters

Inhalt:
A quadratic regression model-based predictive control method is presented for grid-side converters. The control scheme calculates the cost function for seven voltage vectors of the inverter and extends it to the entire modulation region using regression. Then a constrained optimization problem must be solved to find the most feasible control action in this continuous control space. The found optimal voltage vector can be generated using Pulse-Width Modulation methods. The control method is validated in a custom Hardware-in-the-Loop environment using model-based rapid-prototyping tools of Matlab/Simulink. Measurement results show the viability of the method.