Model Predictive Control with Parameter Identification of DC-Transformer (DCX) in Class-D Audio Amplifier

A dual-bridge series resonant converter (DBSRC) has the characteristics of bidirectional power flowing capability, high power density and high efficiency. Model predictive control has been widely applied to power electronics and motion control, and the control performance relies heavily on the accuracy of the predictive model. Both the error introduced by the fundamental harmonic analysis and the error introduced by the mismatch of resonant tank parameters could affect the steady-state output voltage of a DBSRC. To address the aforementioned issues, a parameter identification method based on Kalman Filter is proposed, which thoroughly eliminates the control target deviation caused by the application of FHA and the mismatch of resonant tank parameters. Finally, the effectiveness of the proposed method is validated through a 150W experimental prototype.