Novel Approach for Converter Systems Modeling, Design, and Comparisons

Abstract

In the wake of increasing demand for DC-DC converters with high power density and high efficiency, the design of modern power converters has evolved into a multi-objective optimization problem with multiple uncertainties associated with parameter and component selection. The iterative design procedure, manual component selection, and the diverse nature of passive and active components complicate optimum DC-DC converter design and comparison. Existing methods suffer from excessive computation time, inaccurate component modeling, and inadequate design variables, resulting in sub-optimal solutions. In this thesis, the design of a non-isolated DC-DC power converter is formulated as a multi-objective optimization problem, which seeks to reduce the total converter volume and power loss. A new component modeling method is proposed, that incorporates the impact of major converter components, including the inductor, capacitor, MOSFET, and heat sink, to overcome the uphill challenge of automating the optimization process. This model allows for the mathematical expression of design objectives in terms of technology-specific parameters and converter operating conditions without selecting or assuming specific components. After the optimal solutions are determined, all components can be selected in one step at the end of the process. Experimental and data-driven analyses are conducted to confirm the validity of the component modeling and design method. The automated optimum design procedure is then used to study the effects of various parameters on design objectives and to compare the efficiency and power density of different converter topologies for particular applications.