Topology-aware distributed smart grid control using a clustering-based utility maximization approach

Abstract

Efficiently balancing supply and demand in the electrical distribution grid is an important Smart Grid service. The development of applicable agent-based concepts for match-making energy generators with consumers is an active research field. Implementing Smart Grids as a system of Micro Grid systems, we propose an efficient decentralized optimization scheme, which avoids central components in order to increase dependability, robustness against communication failures, and denial-of-service attacks. The concept utilizes traditional peer-to-peer functionality in combination with locality-aware routing: Two explicit clusterings are evaluated and compared to the fully-meshed communication approach. We investigate the impact of node clustering on the optimality and locality of the distributed Smart Grid control algorithm and show that our approach significantly reduces the electrical losses in the examined power distribution grid compared to the centralized approach.