Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
380851 | Engineering Applications of Artificial Intelligence | 2013 | 13 Pages |
The smart grid vision demands both syntactic interoperability in order to physically be able to interchange data and semantic interoperability to properly understand and interpret its meaning. The IEC and the EPRI have backed to this end the harmonization of two widely used industrial standards, the CIM and the IEC 61850, as the global unified ontology in the smart grid scenario. Still, persisting such a huge general ontology in each and every one of the members of a distributed system is neither practical nor feasible.Moreover, the smart grid will be a heterogeneous conglomerate of legacy and upcoming architectures that will require first the possibility of representing all the existing assets in the power network as well as new unknown ones, and second, the collaboration of different entities of the system in order to deploy complex activities. Finally, the smart grid presents diverse time span requirements, such as real-time, and all of them must be addressed efficiently but use resources sparingly.Against this background, we put forward an architecture of intelligent nodes spread all over the smart grid structure. Each intelligent node only has a profile of the global ontology. Moreover, adding reasoning abilities, we achieve simultaneously the required intelligence distribution and local decision making. Furthermore, we address the aforementioned real-time and quasi-real-time requirements by integrating stream data processing tools within the intelligent node. Combined with the knowledge base profile and the reasoning capability, our intelligent architecture supports semantic stream reasoning. We have illustrated the feasibility of this approach with a prototype composed of three substations and the description of several complex activities involving a number of different entities of the smart grid. Moreover, we have also addressed the potential extension of the unified ontology.