Graph Theory-Based Power Routing in Modular Power Converters Considering Efficiency and Reliability

Abstract

Modular power converters offer reduced $dv/dt$ , improved power quality and higher redundancy. Normally, the power is shared equally by the individual cells of a modular system. However, differences in electrical and thermal parameters of the individual cells result in a non-optimal efficiency and reliability characteristics with equal power sharing. Therefore, an unequal internal power sharing strategy, called power routing, has been proposed to improve primarily the system reliability. Nevertheless, a unified approach to route the power in modular converter architectures is missing. This work uses graph theory to represent various modular topologies for easier identification of power paths. Graph theory simplifies the power flow in complex modular architectures to a network flow problem, focusing on the reliability and efficiency optimization. The paper presents a power routing algorithm based on convex cost flow optimization for improving the reliability without sacrificing the efficiency. The proposed method is validated through case studies and the proof of concept is demonstrated experimentally.