Key points are not available for this paper at this time.
The stability of grid-connected active front-end converters (AFEs) is often compromised by the intricate interplay between inherent converter nonlinearities and grid dynamics. This article presents a novel approach to dynamic model predictive current control (MPC) by leveraging recursive least squares (RLS) to accurately estimate the properties of a physical model. This estimation mitigates stability issues, setting the stage for improved control. Unlike conventional methods, our proposed RLS-based MPC, enriched with an auto-tuning feature, empowers controller development without the upfront need for precise external dynamics. This not only ensures high disturbance rejection but also maintains a high-fidelity control performance, rendering the approach versatile for applications where obtaining or predicting precise external dynamics is challenging. At each sampling interval, a cost function is applied to predicted variables to discern optimal switching states. Through extensive simulation studies covering diverse grid impedance changes and system nonlinearities, we evaluate the controller's effectiveness. To underscore its superiority over traditional controls, simulation results are validated on a laboratory hardware platform equipped with Typhoon HIL and dSPACE real-time emulators, further attesting to the robustness and practical viability of our proposed approach.
Debnath et al. (Thu,) studied this question.