Bidirectional Optimisation for Load Shaping within Coupled Microgrids

TL;DR

This paper presents a bidirectional optimization framework for load shaping in coupled microgrids, enhancing flexibility and efficiency through a bilevel approach with proven convergence and real-world data validation.

Contribution

It introduces a novel model where power exchange does not depend on demand, improving flexibility, and adapts a bidirectional scheme with distributed routines and quadratic programming.

Findings

Increased flexibility in power exchange independent of demand

Efficient parallelizable quadratic programming approach

Global convergence of the optimization scheme

Abstract

We address the problem of load shaping within a network of coupled microgrids (MGs) in a bilevel optimisation framework. To this end, we consider the charging/discharging rates of residential energy storage devices within each MG on the lower level and the power exchange among neighbouring MGs on the upper level as optimisation variables. We improve a previously developed model such that the maximal amount of exchanged power does not depend on the power demand, thus, increasing the flexibility within the network, and adapt the corresponding bidirectional optimisation scheme accordingly. For efficiency, standard distributed optimisation routines are used for the optimisation on the lower level; the power exchange problem on the upper level is replaced by parallelisable small-scale quadratic programmings. We prove global convergence of the optimisation scheme and illustrate the potential of the approach in a numerical case study based on real-world data.