Search in the site by keyword

reports - Deliverable

Design of digital twins for energy networks and models development for the characterization of physical phenomena from videos

reports - Deliverable

Design of digital twins for energy networks and models development for the characterization of physical phenomena from videos

The report describes the new functionalities of the MESP-DT platform for managing Digital Twins in the energy sector: the management of geographic and temporal aspects, the addition of the simulation component, and the support for early warning systems and predictive maintenance management of assets. The analysis of images and videos continues with artificial intelligence techniques for the recognition of surface discharge on insulators and monitoring of extreme weather events.

In the context of designing Digital Twin platforms for the analysis and management of multi-energy networks, this deliverable discusses the new requirements for the Multi Energy Semantic Platform (MESP) and the definition of its Digital Twin architecture, defined MESP-DT.

 

The platform will support the entire asset lifecycle, the integration of real-time measurements from the field, the storage of network information in knowledge graphs with version management and the integration with network components simulators. To support real-time data management, features have been introduced in the MESP-DT to handle streaming data using Apache Kafka-based solutions.

 

The MESP-DT has been equipped with knowledge graph versioning capabilities that allow you to retrieve and reanalyze grid configurations in the past. For the description and analysis of geo-spatial aspects on grids, the management of geographical aspects has been integrated using the GeoSPARQL language. A service for the management of simulations on network assets has also been integrated into the MESP-DT and the case of a Digital Twin of a Storage System has been developed, operating in both learning and simulation modes. In addition, planning has been started for the integration into the MESP-DT of early warning algorithms for underground cable failures using grid information from the knowledge graph.

 

Then, functions for predictive maintenance of network assets have been integrated with the use of machine learning techniques. As an example, predictive maintenance for backup batteries of secondary substations is presented. Finally, artificial intelligence techniques were used for the processing of images and videos to support the diagnostics of energy networks.

 

Specifically, the results of the study are presented for the characterization and autonomous monitoring of two phenomena harmful to the infrastructures of the electricity system: the surface discharges that occur on the surfaces of the insulators due to the passage of leakage currents due to the accumulation of polluting deposits and the formation of snow caps on the conductors of overhead line following the occurrence of wet snowfall.

Projects

Comments