Finding Efficient and Lower Capacitance Paths for the Transfer of Energy in a Digital Microgrid

Abstract

In a digital microgrid (DMG), different from an analogous microgrid, energy is transmitted in well-defined amounts and in a store-and-forward fashion. Nodes of a DMG network, or energy packet switches (EPSs), use supercapacitors as temporary energy storage units to control the amount of energy supplied to a load. An EPS aggregates energy coming from different inputs or sources and forwards it to other EPSs or to a load. Rather than referring to electrical power, we measure the delivery of it as energy. An EPS is built with many supercapacitors to be able to provide significant amounts of energy to one or multiple loads. An EPS dedicates a configurable number of supercapacitors to an energy flow. In this paper, we find the conditions to achieve the smallest energy loss in the supply of energy from energy sources to loads in a DMG and propose a routing algorithm to find a path with small capacitance in a DMG network built with store-and-forward energy nodes. In addition, because an EPS has a finite amount of capacitance, the number of flows that the DMG can sustain may be limited. Exacerbating this problem, the passive transfer of energy between energy units may suffer losses as a result of the capacitance used and energy transmitted between supercapacitors. Therefore, the path between a source and a load has to be carefully selected. To solve this problem, our proposed routing algorithm finds the smallest capacitance paths to enable the scalability of the DMG. We analyze a path of supercapacitor-based networks and underscore the conditions to achieve minimal energy losses, to minimize the path capacitance, and to balance these two conflicting objectives. We analyze these approaches and show numerical examples on a small power network. Results show that total energy loss in this DMG is path independent as this loss depends on only the voltage of the capacitors at Node 1; the node connected to the source. In addition, results show that by adopting the proposed algorithm, the scalability of the DMG can be increased by finding the smallest capacitance paths to transfer energy between the sources and the loads. We show how store-and-forward transfer works on an actual DMG testbed with two EPSs and two loads.