Are you an EPFL student looking for a semester project?
Work with us on data science and visualisation projects, and deploy your project as an app on top of Graph Search.
In a context in which economy and ecology follow historically opposed trajectories, renewable energies appear as an interesting compromise. However, the energy production from these emerging technologies is well below that of their ancestors, and their costs remain high. For this reason, the "Solar-Impulse" solar airplane project was initiated to show the virtues and the potential of renewable energies. This aircraft, the first of its kind in the world of aviation, is also an incredible vector for information. In fact, designed to fly day and night and propelled exclusively by solar energy, this aeroplane displays the performance that renewable energy can achieve when combined with high efficient systems. The subject of this thesis focuses on the challenges of the Solar-Impulse, and that is why it presents a methodology for designing an interleaved DC/DC converter adapted to applications in which efficiency and weight are the essential criteria. Using given specifications, this methodology can determine the optimal topology and the components to be used in its creation. Within the context of the Solar-Impulse, the converter is used to control the flow of energy between the photovoltaic generator and the battery module that regulates voltage from the continuous bus. In this type of application, the most appropriate topology is a boost converter, as it steps up the voltage of the continuous bus in order to increase efficiency and reduce the weight of the overall system. In order to identify the optimal converter topology for an application, in terms of efficiency and power density, a model of each component within the converter must be created. Models of the components of the commutation cell were developed in order to be able to estimate losses within the cell and compare various converter operating modes from an energetic perspective. Having modelled the commutation cell, a method for the optimal dimensioning of the inductance of each channel was then developed with the goal of obtaining the dimensioning that offers the best compromise between weight and total loss of inductance. As the converter was designed for an application that operates under high altitude (10,000 m) climate conditions, the influence of altitude on the dimensioning of the heat sink was studied in order to determine the most adverse conditions. Following this analysis, a thermal model was proposed. After completing the modelling, the sensitivity of the various design parameters was studied. Based on this study, the influence of these parameters on efficiency and power density was represented using a Pareto front. Based on this analysis, two dimensioning methodologies are proposed. The first is a multi-criterion method able to trace the dependency between the converter's efficiency and its power density through the intermediary of the Pareto front. The second method, a mono-criterion approach, maximises efficiency while respecting the constraint imposed on power density. The mono-criterion method applied to maximising the European efficiency criterion showed that an increase in the number of channels increased the quantity of energy collected over a day by increasing the power density of the converter. Lastly, the prototype created based on the theory developed in this work is presented, and its performance quantified, with the goal of validating the elaborated dimensioning methodology.