Exact aggregate models for optimal management of heterogeneous fleets of storage devices

TL;DR

This paper introduces a novel aggregate modeling approach for heterogeneous storage device fleets, enabling efficient, scalable optimization for power grid management by treating large fleets as single units under certain conditions.

Contribution

It presents an exact aggregate model that simplifies large-scale fleet optimization, making it independent of fleet size and scalable with the number of time-slots.

Findings

Model scales linearly with time-slots, independent of fleet size

Enables treating heterogeneous fleets as single storage units

Facilitates optimal management in power grid applications

Abstract

Future power grids will entail large fleets of storage devices capable of scheduling their charging/discharging profiles so as to achieve lower peak demand and reduce energy bills, by shifting absorption times in sync with the availability of renewable energy sources. Optimal management of such fleets entails large scale optimisation problems which are better dealt with in a hierarchical manner, by clustering together individual devices into fleets. Leveraging on recent results characterizing the set of aggregate demand profiles of a heterogeneous fleet of charging (or, respectively, discharging) devices we propose a way to achieve optimality, in a unit commitment problem, by adopting a simplified formulation with a number of constraints for the fleet that scales linearly in the number of time-slots considered and is independent of the size of the fleet. This is remarkable, as it shows that, under suitable conditions, a heterogeneous fleet of any size can effectively be treated as a single storage unit.