Accurate estimation of losses in high-power traction converters is essential for an effective design. Precise estimation of switching and conduction losses is crucial for this purpose. In this paper, the widely recognized Double Pulse Test (DPT) is employed to determine these losses. However, time-shift errors and misalignments in measurements can lead to significant deviations in loss estimation of the actual setup. This paper introduces a postprocessing method aimed at mitigating time-shift and misalignment issues in voltage and current waveforms. The proposed method is validated through simulation, demonstrating its effectiveness in improving the accuracy of loss estimation for high-power traction converters.

This paper evaluates traction inverters for heavy-duty electric vehicles, focusing on key criteria such as raised power ratings with improved efficiency and power densities. Boosted voltage and current levels are required to achieve higher power levels and provide megawatt charging system solutions, which results in the need to utilize new semiconductors and topologies. In this study, 3-Level neutral point clamped (3L- NPC) and 2-Level 6-phase (2L-6Ph) voltage source inverters (VSIs) are evaluated and compared to conventional 2-Level 3-phase (2L-3Ph). The comparison uses figure-of-merit parameters and a virtual prototyping method based on several performance indices, such as efficiency, power density, output harmonic quality, and reliability. Then, efficiency maps are acquired to find out the sweet operating points, minimizing losses. Results show that the 3L-NPC VSI system provides a higher switching frequency, which also shrinks the size of the passive elements and cooling system. Although the 3L-NPC inverter requires additional power switches and isolated gate drivers, its estimated performance outweighs such reliability and cost-dependent issues. Therefore, this study concludes that multi-level inverter topologies hold promise for high-voltage, high-power traction drives.

The semiconductor industry plays a critical role in numerous sectors, yet faces vulnerability in its supply chains. The recent global semiconductor shortage highlighted the risks of relying on a single supplier. To mitigate this, companies adopt dualsourcing strategies, but power devices like silicon carbide (SiC) metal-oxide semiconductor field-effect transistors (MOSFETs) pose challenges due to manufacturing nuances. Configurable gate-drive units (GDUs) offer flexibility but often require external input for device recognition. This paper introduces a method to achieve a self-configurable gate-drive unit based on measuring the gate step-response for power device identification. The proposed method enhances safety, ensures seamless integration, and offers adaptability in full-bridge or multi-phase systems. Experimental results demonstrate component uniformity, emphasize the importance of interval selection, and showcase the impact of external gate resistors on rise and fall times. Estimations of input capacitance using different methods highlight their effectiveness in distinguishing among devices. The practical implementation of the proposed method contributes to the efficiency, reliability, and cost-effectiveness of self-configurable GDUs.