Improved Modulated Model Predictive Control for Three-Level NPC Inverter-Fed PMSM Drive With Ternary Partial Derivative-Based Cost Function

In the three-level neutral point clamped (NPC) inverter-fed PMSM drive, the modulated model predictive control (M2PC) possesses the advantages of fixed switching frequency, lower torque ripple, and higher current quality compared with the traditional finite control set model predictive control (MPC). However, there are some defects with it such as complicated calculation, limited accuracy, and neutral point (NP) voltage fluctuation. To address these problems, this article proposes an improved M2PC with the ternary partial derivative-based cost function. First, according to the principle of deadbeat, the sector of the reference vector is quickly identified, and the corresponding voltage vectors (VVs) are selected for subsequent steps. Then, both the motor current control and the neutral voltage balance are fully considered when establishing the cost function. Finally, by performing ternary partial derivative operation on the designed cost function, the duty cycles of the selected VVs that minimizes the cost function value are obtained. The results of simulations and experiments confirm the effectiveness of the proposed strategy in suppressing NP voltage fluctuation, reducing current total harmonic distortion, and improving the system steady-state performance.