A High-Frequency Transformer(HFT) design for a Dual Active Bridge(DAB) converter

Technical Summary: A High-Frequency Transformer (HFT) Design for a Dual Active Bridge (DAB) Converter

Problem Statement
The Dual Active Bridge (DAB) converter is a critical topology for high-power-density applications such as electric vehicles and renewable energy systems, offering bidirectional power flow, galvanic isolation, and soft-switching capabilities. However, the performance of DAB converters is heavily dependent on the design of the High-Frequency Transformer (HFT). Unlike conventional transformers where leakage inductance is minimized, DAB operation relies on specific leakage inductance values to facilitate power transfer and achieve soft-switching. Designing an HFT for DAB applications presents unique challenges, including the management of core and copper losses at high switching frequencies (e.g., 100 kHz), the optimization of leakage inductance for varying load conditions, and the mitigation of skin and proximity effects in windings. Furthermore, the design must balance thermal stability, efficiency, and power density without compromising the converter's operational limits.

Methodology
The paper outlines a systematic design methodology for a 1 kW HFT operating at a 100 kHz switching frequency with a 400 V/48 V conversion ratio. The process involves four primary stages:

1. Specification and Core Selection: The design targets a maximum leakage inductance of 200 µH and limits magnetizing current to 2–5% of the nominal current. Various magnetic materials (Ferrite, Nanocrystalline, Amorphous, Si-iron, Permalloy, Powdered Iron) were evaluated based on core loss, permeability, and Curie temperature. Ferrite (Epcos N87) was selected for its low core loss (684 mW/cm³) and suitability for high-frequency operation. Three core geometries (ETD, UI, and Toroidal) were compared to determine the optimal shape based on window space factor ($W_x \leq 0.3$) and maximum flux density ($B_m \leq 0.2$ T).
2. Winding Design: To address skin and proximity effects at 100 kHz, multi-stranded Litz wire (AWG 33) was selected. The individual strand radius was kept below the calculated skin depth (~0.2 mm). The design utilized 40 strands for the primary winding and 200 strands for the secondary to achieve current densities of 4 A/mm², resulting in cross-sectional areas of 1 mm² and 5 mm², respectively.
3. Leakage Inductance Control: The leakage inductance was treated as a critical design parameter rather than a parasitic element to be minimized. Various winding arrangements were evaluated to achieve the target inductance.
4. Simulation and Validation: The design was modeled and analyzed using ANSYS Maxwell (PEmag) for 2D Finite Element Analysis (FEA) to determine inductance, losses, and magnetic field distribution. Circuit-level validation was performed using LTspice to simulate the DAB converter under Single-Phase Shift (SPS) modulation.

Key Contributions

• Optimized Core and Winding Selection: The study identifies the ETD 39/20/13 core with N87 ferrite material as the optimal choice, balancing low core loss and thermal stability. It demonstrates the efficacy of multi-stranded Litz wire in mitigating high-frequency losses.
• Leakage Inductance Management: The paper presents a design approach where leakage inductance is precisely controlled (targeting ~131.92 µH referred to the HV side) to ensure stable bidirectional power transfer without the need for complex external inductors or dynamic adjustment mechanisms found in other studies.
• Comprehensive Loss Analysis: The work provides a detailed breakdown of losses, calculating a total copper loss of 2.17 W and an average core loss of 1.8 W for the ETD configuration, resulting in a total transformer loss of approximately 3.97 W.
• Validation of Design Methodology: The research validates the proposed design through a combination of FEA and circuit simulation, confirming that the transformer operates safely within saturation limits and maintains efficient coupling.

Results

• Inductance Values: The FEA analysis yielded an equivalent leakage inductance of 2.06 µH (referred to the LV side) and 131.92 µH (referred to the HV side). The primary and secondary self-inductances were calculated at 22.42 mH and 350.36 µH, respectively.
• Loss and Efficiency: The total transformer loss was determined to be 3.97 W. The copper loss was 2.17 W, and the core loss was 1.8 W.
• Magnetic Performance: The maximum flux density was observed to be approximately 0.18–0.19 T, well within the 0.2 T design constraint to prevent saturation. The maximum flux density in the core was 33,008 µT.
• Current Density: The use of Litz wire resulted in a safe current density of $8.746 \times 10^6$ A/m², effectively eliminating skin and proximity effects.
• Circuit Simulation: LTspice simulations confirmed that the designed transformer enables stable power transfer under SPS modulation, with waveforms showing effective phase-shifting operation and current flow consistent with theoretical expectations.

Significance and Claims
The authors claim that this paper provides a reliable, easily predictable, and practical design methodology for HFTs in DAB converters. Unlike other approaches that rely on fully integrated designs with complex dynamic adjustments or wide-input-current-fed configurations, this study demonstrates that a standard ETD 39/20/13 core with optimized winding arrangements can achieve high performance. The design addresses common issues such as high-voltage insulation challenges and mechanical fatigue associated with moving parts in other designs. By maintaining magnetic flux density below saturation limits and utilizing multi-stranded Litz wire, the proposed transformer ensures thermal stability and high efficiency. The authors assert that this work offers a solid foundation for engineers and researchers developing high-power-density converter architectures, particularly for applications requiring robust and efficient galvanic isolation.