Five degree of freedom modulation parameters for three level dual active bridge converter
Abstract
To improve the performance of isolated bidirectional DC/DC converters, this article investigates an enhanced control strategy for the multilevel (ML) dual active bridge (DAB) topology. Compared with the conventional DAB structure, the ML structure supports the use of lower voltage rating switches and provides more control degrees of freedom. The five degree of freedom modulation strategy used here, in which the optimal duty cycles are determined from the fundamental and third harmonic components of the high frequency bridge voltages by means of vector diagrams and without complex node analysis. The contribution of the present study is the implementation of that established framework: the conventional look-up table based realisation is replaced by an offline trained function fitting neural network that predicts the five duty ratios directly from the measured operating point, avoiding both multi-dimensional tables and iterative online optimisation. The inductor current is significantly reduced at light loads by the proposed modulation strategy and neural network. Moreover, this optimization method does not require complex optimization tools. All results reported in this paper are obtained from MATLAB/Simulink simulations of a 100 kW three-level NPC DAB converter, and no experimental prototype has been built. Within this scope, the simulation results confirm the effectiveness of the proposed method in reducing the inductor current. Compared with single phase shift (SPS) modulation, the proposed strategy reduces the inductor current RMS value by up to 34.8% at light load (10 kW) and by about 22% at 50 kW, while the three strategies converge at the rated 100 kW, which indicates that the proposed method is a promising candidate for next generation power conversion systems, subject to experimental confirmation.
Related
- Effects of Fatigue on Six-Degree-of-Freedom Knee Kinematics in Military Personnel Marching on a Treadmill
- Simulation-based active cell balancing of a four-cell lithium-ion pack using a Cuk converter and fuzzy-PID current control
- A New Degree of Freedom in Stellarator Coil Design
- Rotational elastography: six degree-of-freedom seismology in soft tissues measures local elasticity with diffuse fields
- Chirality as a Structural Degree of Freedom in Halide Perovskites
BibTeX
@article{kumar2026five,
title = {{Five degree of freedom modulation parameters for three level dual active bridge converter}},
author = {Seelam Vasavi Sai Viswanada Prabhu Deva Kumar and Shyam Akashe and Govind Murari Upadhyay},
journal = {Scientific Reports},
year = {2026},
month = oct,
doi = {10.1038/s41598-026-71195-w},
url = {https://www.nature.com/articles/s41598-026-71195-w},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - Five degree of freedom modulation parameters for three level dual active bridge converter
AU - Seelam Vasavi Sai Viswanada Prabhu Deva Kumar
AU - Shyam Akashe
AU - Govind Murari Upadhyay
PY - 2026
DA - 2026/10/04
JO - Scientific Reports
DO - 10.1038/s41598-026-71195-w
UR - https://www.nature.com/articles/s41598-026-71195-w
AB - To improve the performance of isolated bidirectional DC/DC converters, this article investigates an enhanced control strategy for the multilevel (ML) dual active bridge (DAB) topology. Compared with the conventional DAB structure, the ML structure supports the use of lower voltage rating switches and provides more control degrees of freedom. The five degree of freedom modulation strategy used here, in which the optimal duty cycles are determined from the fundamental and third harmonic components of the high frequency bridge voltages by means of vector diagrams and without complex node analysis. The contribution of the present study is the implementation of that established framework: the conventional look-up table based realisation is replaced by an offline trained function fitting neural network that predicts the five duty ratios directly from the measured operating point, avoiding both multi-dimensional tables and iterative online optimisation. The inductor current is significantly reduced at light loads by the proposed modulation strategy and neural network. Moreover, this optimization method does not require complex optimization tools. All results reported in this paper are obtained from MATLAB/Simulink simulations of a 100 kW three-level NPC DAB converter, and no experimental prototype has been built. Within this scope, the simulation results confirm the effectiveness of the proposed method in reducing the inductor current. Compared with single phase shift (SPS) modulation, the proposed strategy reduces the inductor current RMS value by up to 34.8% at light load (10 kW) and by about 22% at 50 kW, while the three strategies converge at the rated 100 kW, which indicates that the proposed method is a promising candidate for next generation power conversion systems, subject to experimental confirmation.
SN - 2045-2322
ER -