Applying Model-Free Predictive Current Control With Parallel Cost Functions for Open-End Winding Synchronous Reluctance Motor

This article presents a model-free predictive current control (MFPCC) with parallel cost functions for the common dc-link dual-inverter-fed open-end winding synchronous reluctance motor (OW-SynRM). The MFPCC has a general prediction algorithm that can be implemented in any reference frame. Using this feature, and the fact that each winding of the OW-SynRM is supplied separately by two legs of the dual inverter, the current predictions are achieved independently for each phase in the Cartesian coordinates. Then, three parallel cost functions are defined for three predicted phase currents. Thus, the finite control-set approach is utilized separately for each phase to determine the switching state of the dual inverter. In addition, an adaptive proportional-resonant extended state observer (PR-ESO) is introduced to improve the current predictions regarding the operating frequency of the motor. The proposed MFPCC for OW-SynRM has several advantages. First, the computational burden is decreased significantly because the prediction algorithm utilizes fewer candidates in the Cartesian coordinates. In addition, the weighting factor is not required due to using independent cost functions. Furthermore, the zero sequence current is automatically adjusted without a controller. Finally, the PR-ESO-based model-free approach improves the current predictions of the proposed method. The proposed MFPCC of the OW-SynRM is evaluated using the experimental results.