An Effective Model-Free Predictive Control for LC-Filter Voltage Source Inverters

The parametric uncertainties have critical effects on output voltage performance of LC-filtered voltage source inverters (VSIs) under model predictive control (MPC). To overcome this known problem, an effective model-free predictive control (MFPC) is proposed. The proposed MFPC realizes robust and accurate voltage predictions by using the capacitor voltage gradients and inverter-side current gradients. Then, the state-space equations of different voltage vectors are established to calculate all the gradients in real time, which can ensure the accuracy of voltage and current gradients. Finally, the practicality of the proposed scheme is demonstrated under ideal and mismatched model. In comparison with conventional MPC, the proposed MFPC not only achieves a comparable voltage performance, but also realizes better robustness.