Contrary to the rapid evolution experienced in the last decade of Information and Communication Technologies and particularly the Internet of Things, electric power distribution systems have remained exceptionally steady for a long time. Energy users are no longer passive actors; the prosumer is expected to be the primary agent in the Future Grid. Demand Side Management refers to the management of energy production and consumption at the demand side, and there seems to be an increasing concern about the scalability of Demand Side Management services. The creation of prosumer communities leveraging the Smart Grid to improve energy production and consumption patterns has been proposed in the literature, and several works concerned with scalability of Demand Side Management services group prosumers to improve Demand Side Management services scalability. In our previous work, we coin the term Social Internet of Energy to refer to the integration between devices, prosumers and groups of prosumers via social relationships. In this work, we develop an algorithm to coordinate the different clusters we create using the clustering method by load profile compatibility (instead of similarity). Our objective is to explore the possibilities of the cluster-by-compatibility heuristic we proposed in our previous work. We perform experiments using synthetic and real datasets. Results show that we can obtain a global reduction in Peak-to-Average Ratio with datasets containing up to 200 rosumers and creating up to 6 Prosumer Community Groups, and imply that those Prosumer Community Groups can perform load rescheduling semi-autonomously and in parallel with each other.
Keywords: clustering; prosumer community groups; smart grid; social internet of energy; social internet of things.