In the electrical sector, ontologies have long been used to describe transmission and distribution networks. The IEC CIM standard has been proposed as a reference model for representing network elements and their interconnections. One advantage of using these solutions is the ability to integrate different databases into a single model, based on semantic triples of “subject-predicate-object.” This approach overcomes issues related to storing different types of information in multiple relational databases (e.g., separate databases for geographical information, topological data, asset data), which utilities typically manage in disparate and often misaligned databases. The use of ontologies in conjunction with semantic datastores also allows for the inference of information from the underlying ontologies.

Moreover, these technologies facilitate interoperability between different systems and the exchange of information among operators. Given these features of semantic models and anticipating increased interaction between various energy vectors, the three-year research program included the development of a semantic platform for integrating information across different energy vectors: the Multi Energy Semantic Platform (MESP). This platform, based on the IEC CIM standard for electrical networks, also supports different types of energy networks and aims to provide utilities and researchers with an integrated set of tools and functions starting from the description of network infrastructures. Using a semantic model similar to that of electrical networks, MESP can also handle analyses on other energy vectors such as district heating, gas, and water, with the goal of developing a unified integrated system. As an initial example, the district heating network has been addressed with an ontology derived from IEC CIM. In addition to developing the architecture, functionalities such as geographical and topological visualization of networks and network analysis functions (e.g., power flow analysis) have been implemented in MESP. These functions, realized with Docker-based microservices, are accessible via a web application. The platform is easily expandable and allows for the addition of further services of interest to utilities or research.

The use of semantic data models has also been considered for analyzing continuous real-time data streams. To develop predictive models for managing distribution network assets, an architecture was designed and implemented to integrate streaming data (e.g., electrical measurements, faults, weather conditions) with semantic information related to the network (e.g., topology, components). This architecture enables the creation of a unified, semantically enriched data stream from various available data flows, ready for analysis using Streaming Machine Learning tools.