IN Brief:
- Researchers compared bidirectional full-bridge, unidirectional Vienna rectifier and bidirectional T-type SiC converters using a common 25kW experimental platform.
- Measured grid-current distortion was 3.46% for the T-type, 6.48% for the full-bridge and 7.42% for the Vienna rectifier under the reported test conditions.
- Simulations of up to six parallel 100kW modules showed that carrier interleaving reduced distortion substantially in the full-bridge and T-type configurations.
Researchers at Portugal’s University of Minho have compared three silicon-carbide AC-DC converter topologies for modular high-power electric vehicle charging, combining simulations of parallel converter arrangements with measurements from a common 25kW experimental platform. The study identifies differences in grid-current distortion and ripple, and shows how those differences change when converters operate together with coordinated switching phases.
Published on 20 October 2026 in Frontiers in Future Transportation, the research examines a bidirectional two-level full-bridge converter, a unidirectional three-level Vienna rectifier and a bidirectional three-level T-type converter. All three use silicon-carbide power devices and were assessed with a common current-control approach to distinguish the effects of converter structure and modulation from differences in the basic controller.
The experimental comparison found grid-side current total harmonic distortion of 6.48% for the full-bridge, 7.42% for the Vienna rectifier and 3.46% for the T-type converter. The corresponding measured current ripple was 231mA, 212mA and 183mA. These figures were obtained from the study’s single-module laboratory setup under the stated operating conditions; they are not universal performance ratings for each circuit architecture.
High-power charging equipment normally contains an AC-DC stage connected to the electricity supply and a downstream DC-DC stage that establishes the electrical conditions required by the vehicle. The grid-facing converter must transfer energy while controlling the current drawn from the supply, maintaining the required DC-link voltage and limiting unwanted harmonic components.
Harmonics arise when the current waveform contains frequencies beyond its fundamental component. Switching converters can produce these components through their operating sequences and control behaviour. Excessive distortion may increase losses, affect other connected equipment and create difficulties meeting the requirements applicable at the installation’s point of common coupling.
The full-bridge arrangement uses six actively controlled semiconductor switches, with two devices assigned to each phase. It can support power flow in either direction when the wider electrical system and controls are configured accordingly. Its relatively simple two-level switching structure reduces the number of active devices compared with some multilevel alternatives, but it also produces larger voltage transitions at the converter output.
Those transitions influence the current ripple that the coupling inductors and control system must manage. An inductor limits the rate at which current changes in response to an applied voltage, so the voltage steps created by the power stage affect the filtering needed to obtain a suitable grid-current waveform.
The Vienna rectifier uses a three-level arrangement incorporating controlled switches and diodes. Its intermediate voltage levels can reduce selected ripple components, but the diode conduction paths restrict the available operating states during parts of the electrical cycle. It is normally used for unidirectional conversion and cannot provide the same reverse-power-flow functionality as the two bidirectional topologies assessed in the study.
The T-type circuit also produces three voltage levels but uses a greater number of actively controlled semiconductor devices. Its switching arrangement allows more flexible synthesis of the converter voltage, helping the controller track the desired current waveform. That improvement carries additional requirements for gate driving, switching coordination and DC-link voltage management.
To maintain a consistent comparison, the researchers adopted the same deadbeat predictive current controller for all three circuits. The method uses a discrete-time electrical model to calculate the converter voltage required to bring the predicted current towards its reference during the next sampling interval. It is implemented as a continuous-control-set model predictive control scheme, with topology-specific modulation translating the calculated voltage into switching commands.
Using the same control law helps isolate the effects of the power circuit, although the three converters still have different switching states and modulation constraints. The comparison therefore considers both the quality of the resulting electrical waveform and the implementation complexity associated with each architecture.
In the laboratory, the team used a reduced-scale 25kW SiC converter platform to test one module at a time. The common hardware arrangement, current sensors, coupling filter, DC-link conditions and switching frequency provided a consistent basis for the measured comparison. The work used a 50kHz switching frequency selected to balance ripple, switching losses and digital control requirements.
The T-type converter produced the lowest measured current distortion and ripple in those single-module experiments. The Vienna rectifier achieved less ripple than the full-bridge, but had the highest distortion of the three designs. Its diode conduction intervals prevent the controller from freely selecting every desired voltage state, which can affect current tracking even when the circuit’s multilevel structure reduces individual switching steps.
A separate computational investigation examined systems containing between one and six 100kW AC-DC modules connected in parallel. These were simulation cases representing aggregate ratings from 100kW to 600kW, not physical charger installations built or tested at those power levels. The researchers compared converters operating with aligned switching carriers against arrangements with a phase offset of 360 degrees divided by the number of modules.
Carrier interleaving changes the timing of ripple-producing switching events. When identical modules operate with appropriately distributed switching phases, parts of their individual current ripple can cancel in the combined waveform. The resulting effect depends on the number of modules, their electrical characteristics, the modulation strategy and whether their contributions remain sufficiently balanced.
In the reported simulations, two or more interleaved full-bridge and T-type modules achieved similar grid-side current distortion of approximately 0.50%. Without that coordinated phase shift, their distortion remained relatively insensitive to the number of modules. The Vienna rectifier retained substantially higher simulated distortion, around 1.53%, because the diode-controlled operating intervals limited the improvement available from interleaving.
For a single simulated module, the T-type achieved approximately 0.75% distortion, compared with 1.13% for the full-bridge and 1.64% for the Vienna topology. These computational results must be distinguished from the higher values measured on the physical 25kW platform. The simulations were based on idealised models and operating conditions that did not reproduce every source of error present in the laboratory.
The researchers attribute the differences in part to semiconductor dead time, digital delays, measurement noise, hardware parasitic effects and distortion already present in the laboratory supply. The experimental and simulated systems were also operated at different scales. The measured values therefore support comparison of the three hardware implementations without validating the simulated performance of a complete multi-module charging station.
The findings change the balance between circuit complexity and electrical performance. A T-type converter may offer the best grid-current quality when one module operates alone, while a full-bridge arrangement can become competitive when several modules are interleaved. The simpler full-bridge power stage may require fewer active switching devices, although total installation cost and reliability also depend on filtering, cooling, control and other equipment.
Bidirectional capability remains a separate design requirement. The full-bridge and T-type circuits can support reverse power flow, whereas the Vienna rectifier is fundamentally intended for unidirectional operation. A charging installation designed for vehicle-to-grid functionality therefore cannot select a topology solely on the basis of current distortion under forward-power operation.
The study’s measured distortion values also cannot be compared directly with one universal grid-compliance limit. Standards such as IEEE 519 assess harmonic behaviour at a defined point of common coupling, with relevant limits depending on the installation and electrical network. EN 50160 concerns supply-voltage quality, which is not the same measurement as converter-side current distortion.
The University of Minho results provide an experimental comparison of individual converter architectures and a separate computational assessment of their behaviour in parallel. Further physical testing of interleaved multi-module systems would be needed to determine whether the predicted cancellation effects and relative advantages persist under real charging loads, component tolerances and network conditions.
