No abstract

Summary This paper presents results from an ongoing research project (SINTEF 2011) on pressure-tolerant power electronics. The main goal of the research is to provide and demonstrate solutions that enable power-electronic converters to operate in pressurized environments. Oil companies have plans for subsea processing of oil and gas. Today's concept considers power-electronic converters in 1-bar vessels. As the depth and the converter power rating increase, these vessels become increasingly bulky and heavy. Pressure-balanced converters would allow lower vessel-wall thickness, thereby giving lower weight, and simpler cooling owing to improved heat conduction through the vessel walls. The new concept considers power-electronic converters placed in vessels completely filled with appropriate dielectric liquid able to operate at pressures from 1 bar up to several hundred bar. Dielectric fluids are required to prove their properties, especially those related to insulation, incompressibility, and heat transport. The methodology presented here considers a mechanical adaptation of components followed by various laboratory experiments for verifying or correcting the proposed solutions (Pittini et al. 2010; Hernes and Pittini 2009; Petterteig et al. 2009). Standard off-the-shelf components and special pressure-adapted components have been subjected to various provocative pressurization tests up to 300 bar. Tests clarified the need for special adaptation of some components, while others could be used without any modification. Subsequently, a full converter phase leg has been built, submerged in dielectric liquid, and tested in full operation up to 300 bar. This phase leg is based on a press-pack IGBT modified for operation at high pressure. Measurements performed at different pressures demonstrate that there is no relevant difference in terms of electrical parameters between this modified IGBT and a standard IGBT. Long-term tests proved the concept's validity. The next step will be to test bonded IGBT technology in a high-pressure environment, and finally, we will realize a test demonstration to be located at a suitable subsea site.

Silicon Carbide (SiC) power devices can provide a significant improvement of power density and efficiency in power converters. The switching performances of SiC power devices are often a trade-off between the gate driver complexity and the desired performance; this is especially true for SiC BJTs and JFETs.

Energy production from renewable energy sources is continuously varying, for this reason energy storage is becoming more and more important as the percentage of green energy increases. Newly developed fuel cells can operate in reverse mode as electrolyzer cells; therefore, they are becoming an attractive technology for energy storage grid-tie applications. In this application dc-dc converter optimization is very challenging due to the large voltage range that the converter is expected to operate. Moreover, the fuel-electrolyzer cell side of the converter is characterized by low voltage and high current. Dc-dc converter efficiency plays a fundamental role in the overall system efficiency since processed energy is always flowing through the converter; for this reason, loss analysis and optimization are a key component of the converter design. The paper presents an isolated full bridge boost dc-dc converter (IFBBC) designed for this new application focusing on losses analysis. The system topology is briefly discussed and the major concerns related to the system, cells stacks and converter operating points are analyzed. The dc-dc converter losses are modeled and presented in detail; the analysis is validated on adc-dc converter prototype rated at 6 kW 30-80 V 0-80 A on the low voltage side and 700-800 V on the high voltage side (for a grid-tie application). The prototype is based on fully planar magnetic, Si MOSFETs, Si IGBTs and SiC diodes; efficiencies up to ~96.5% and ~97.8% were demonstrated depending on the converter operating point.