Cluster-based distributed algorithm for energy management in smart grids

Abstract

In recent years the share of renewable energy sources has risen massively. This leads to for example supply fluctuations, thus forcing power grids to balance their energy demand and generation in order to guarantee operability. This paper proposes a cluster-based and privacy-friendly distributed algorithm for energy management in smart grids. Two clustering approaches are shown. The first one balances the energy consumption and generation of each cluster in a first stage, while adhering to a load limit. After that remaining capacities are distributed among unsatisfied clusters. The second one firstly reports the adaption potential and distributes it among all clusters. Simulation results show that both clustering approaches can improve the energy management, while establishing scalability. This is reached without drawbacks for the communication parameters and allows for an energy management interval of 1 min and a cluster size of 100 participants with a variable amount of clusters.