Optimal Voltage Level-Based Sequential Predictive Current Control With Reduced Complexity for Multilevel Inverters

The conventional Euler-based sequential predictive current control (SPCC) methods for multilevel inverters (MLIs) have aimed to minimize computational complexity, resulting in higher flying capacitor (FC) voltage ripples and poor transient response. These methods typically employ either cost function with weighting factors or offline selection of voltage vectors to reduce the common-mode voltage (CMV). However, improper weighting factor selection and the presence of CMV term in ac current models negatively impact MLI harmonic performance. To address these issues, a new SPCC formulation is proposed, enabling direct estimation of the optimal voltage level based on reference ac currents. This approach eliminates the need for a cost function in ac currents control and simultaneously reduces both computational complexity and CMV. By removing ac current model's dependency on CMV, the proposed formulation further enhances MLI harmonic performance. The mathematical formulation of the proposed SPCC is presented for a four-level inverter (FLI) with both passive and motor drive loads in this study. During the formulation, the Heun's integration method is employed to develop FLI's FC voltage discrete-time models. The performance of the proposed SPCC is demonstrated experimentally on a dSPACE-DS1103 controlled FLI with passive load laboratory prototype. It is further compared with the conventional SPCC methods and their performances are evaluated comprehensively on the laboratory prototype.