The conventional model predictive control (MPC) applied to five-level active neutral point clamped (5L-ANPC) inverters faces a few major challenges, including a significantly heavy computational burden and an intricate tuning of the weighting factors to balance the DC-link and flying capacitor voltages. This paper presents a simplified MPC without weighting factors to leverage the desirable features of MPC within a distributed electric propulsion (DEP) system consisting of a 5L-ANPC inverter-fed axial flux permanent magnet (AFPM) motor-drive system. The proposed method compensates for the one-step delay between the commanded and applied voltage vectors in calculating the reference voltage vector. Afterwards, the optimum voltage vector is selected based on the sign of the calculated reference voltage vector using a deadbeat approach and its updated mapping position in smaller sections within the space vector of the inverter. Finally, a simplified cost function without weighting factors is presented to select the optimum switching state. The proposed method takes remarkably shorter execution time than the conventional MPC does, leading to higher sampling frequency and better steady-state performance. The performance of the proposed method is evaluated in the Acceleration and Cruise Modes of a DEP system’s flight mission profile.
Simplified Model Predictive Control of 5L-ANPC Inverter-Fed AFPM Motor Drives in Electric Aircraft Propulsion Systems
2024-06-19
2024310 byte
Conference paper
Electronic Resource
English
European Patent Office | 2025