D. Boroyevich’s research while affiliated with Virginia Tech and other places

We discuss herein several cooling strategies applicable to shipboard Power Electronics Building Blocks (PEBBs), namely the PEBB 1000 and a notional PEBB 6000 currently under development, subject to the following design goals and constraints: (1) small and lightweight package (i.e., high power density); (2) easy swapability achieved by minimizing any connections requiring sailor intervention for connecting or disconnecting ; and (3) any thermal solution must be able to cool not only a single PEBB unit, but also multiple units placed in close proximity to one another when combined to make up a converter. For the PEBB 1000 application, air cooling is most likely sufficient to meet the needs, and several arrangements of fin structures were investigated. On the other hand, air cooling, water cooling, and water cooling with a dry interface all have potential for meeting the cooling demand of an envisioned PEBB 6000 and warrant further investigation. In any case, the localized heat generated within the PEBB will require significant spreading to a larger surface area for subsequent transfer out of the PEBB.

While wide-bandgap devices offer many benefits, they also bring new challenges for designers. In particular, the new 10 kV silicon carbide (SiC) MOSFETs can switch higher voltages faster and with lower losses than silicon devices while also being smaller in size. These features can result in premature dielectric breakdown, higher voltage overshoots, high-frequency current and voltage oscillations, and greater electromagnetic interference. In order to mitigate these side effects and thus fully utilize the benefits of these unique devices, advanced module packaging is needed. This work proposes a power module package with a small footprint (68 mm × 83 mm), low gate- and power-loop inductances (4 nH), increased partial discharge inception voltage (53 %), and reduced common-mode current (90 %).

The common, and with some exceptions the only method to interconnect high power renewable energy sources and energy storage systems to the grid has been using power electronics converters that operate as current sources to the grid for the purpose of achieving the maximum primary-source power tracking. When grid is not available, such sources would not be allowed to continue operating and will be shut-down after anti-islanded algorithms recognize the loss of the grid. Although existing standards and requirements still limit grid-interface converters to regulate voltage in the grid, this functionality will inevitably be the part of the power system operation in the future. This paper addresses physical and mathematical equivalency between power electronics converters and synchronous generators, and emphasizes how an inherent synchronization feature of the synchronous generators can be used to improve the performance of the grid-interface converters. It has also been shown that if operated as a voltage source, grid-interface converter could have significant stabilizing effect on the system dynamics due to the non-delayed power delivery.

Terminal electromagnetic interference (EMI) models of power converters have been shown to successfully predict conducted emissions up to at least 30 MHz. These models can accurately predict the input side EMI of converters for changes in the source impedance but only for a fixed load. In other words these models cannot predict the input side EMI for changes in the load side parameters. Such terminated EMI models have limited utility as they cannot be used for converters that have a strong coupling between the input and output side EMI. Moreover, in aerospace applications, the conducted EMI standards (DO- 160) dictate limits on the noise currents in all power lines and cable bundles, thus compelling development of models that can predict the EMI on both sides of the power converters. The proposed un-terminated EMI terminal model in this paper directly predicts the input and output side common-mode (CM) noise currents in a motor drive system for changes in both source and load side impedances. Such models can be very useful in understanding how EMI filters and harnesses affect the overall EMI in a motor drive system. The model runs in the frequency-domain and is validated by experiments.

This paper presents complete Verification, Validation, and Uncertainty Quantification (VV&UQ) procedures that are applied to a two-level boost rectifier. The goal of this VV&UQ study is the improvement of the modeling procedure for power electronics systems, and the full assessment of the boost rectifier model (as an example) predictive capabilities. Modeling and simulation of large systems is always the preferred solution to avoid repetitive hardware, and to minimize the cost. However, developing the accurate models and making sure it matches the hardware is always a challenge. This paper provides a set of procedures that if followed one can claim the model is validated.