Controllability Techniques for the Multilevel Power Converter Manhattan Topology

Abstract

This paper explores the controllability of different configurations of the Manhattan Topology. The Manhattan Topology is a multilevel power converter topology that is defined by a set of series stacked capacitors where each capacitor establishes a voltage level. The functionality of the converter is built around the transfer of power between these capacitors. The methodology, quantity, and connectivity of the capacitive power transfer scheme is not specific to the Manhattan Topology. Different topology configurations will have different capacitive power transfer connectivities. A completely connected topology is not necessary for a fully controllable converter (where capacitor voltage balance of any arbitrary ratio can be maintained in steady state). For some practical implementations of the Manhattan Topology, it is also not feasible to connect all capacitive power transfer links together. Different link topologies will result in different levels of controllability. This paper shows three different link topologies: a fully controllable topology, a partially controllable topology, and a modification to the partially controllable topology that results in a fully controllable topology. Converter state-space models, controllability theory, and control diagrams are provided. Results are validated through high-fidelity simulation of example Manhattan Topology power converters that use the three different link topologies in DC/DC mode.