Model-Free Predictive Control of Power Converters With Cascade-Parallel Extended State Observers

The sensitivity of conventional model-predictive control to parametric uncertainties has made model-free predictive control (MFPC) necessary. MFPC with disturbance rejection features requires a high-gain extended state observer (ESO) for state and total disturbance estimation. The standard ESO is limited in performance when high-frequency measurement noise is fed into it. Hence, this article proposes a novel hybrid cascade-parallel ESO, which has good measurement noise suppression and does not diminish the outstanding dynamic performance of predictive control. The proposed ESO also has better disturbance rejection than that of the cascade ESO. Comprehensive theoretical analyses of the cascade-parallel ESO and its alternative structures provide further insights for design and application. Finally, the practicality of the proposed method is demonstrated experimentally by the MFPC of a grid-connected converter.