Work

Two-Stage DC-Link Voltage Regulation Integrated With Model Predictive Torque Control for Torque Ripple Suppression and Efficiency Improvement in IPMSM Drives

Year: 2026

Type: article

Abstract: This paper proposes a system-level, hardware-side approach to suppressing torque ripple and improving the efficiency of interior permanent magnet synchronous motor (IPMSM) drives in the low-to-medium speed range, where a low modulation index increases the relative influence of inverter dead time. The proposed two-stage architecture comprises a closed-loop Buck converter cascaded with a three-phase inverter and is integrated with an existing adjustable-duty-cycle model predictive torque control (MPTC) framework. Regulating the DC-link voltage from 220 V to 100 V increases the modulation index and reduces the relative influence of dead time during active switching, thereby suppressing torque ripple and reducing current total harmonic distortion (THD) without increasing the computational load of the MPTC routine. The relationships among the DC-link voltage, torque ripple, and flux-linkage ripple are derived, the influence of DC-link voltage reduction on the active-vector duration is analyzed, and an impedance-matching analysis is performed to verify the stability of the cascaded system. The proposed mechanism is quantitatively supported by the results: reducing the DC-link voltage increases the active-vector duty ratio from 0.6173 to 0.9523 and decreases the zero-vector duration from 38.27% to 4.77%. In addition, the measured per-switching-cycle ripple ratio of 2.23 is in close agreement with the theoretical value of 2.20. Experiments conducted on a 1.4 kW, 220 V, 3500 rpm IPMSM over 250–1000 rpm and 50%–150% rated torque demonstrate torque-ripple reductions of 16.8%–46.5%. The maximum system-efficiency improvements are 8.75 percentage points at half load, 4.79 percentage points at rated load, and 3.79 percentage points at 150% rated load.

Source: IEEE Access

DOI: 10.1109/access.2026.3745252

Authors: Tsung-Hsun Wu, David Lwo, Jun-Lin Wu, Min-Fu Hsieh

Institutions: National Kaohsiung University of Science and Technology, National Cheng Kung University

Language: English

Topic: Sensorless Control of Electric Motors

Subfield: Electrical and Electronic Engineering

Field: Engineering

Domain: Physical Sciences

SDGs: Industry, innovation and infrastructure, Affordable and clean energy

Open Access status: gold

Estimated APC paid: $2,075