The European Project EnergyGuard project is building a European Testing and Experimentation Facility (TEF) for trustworthy AI in the energy sector, integrating real infrastructures, high‑fidelity digital twins and regulatory assessment environments.

The project reached its Month 18 milestone: at mid‑project, much of this foundation is now in place.

The EnergyGuard consortium met in Soria, Spain, for its Month‑18 in‑person meeting, hosted by CIEMAT and FORA at the FORA facilities.

The gathering marked a pivotal point in the project: beyond coordination, partners reviewed concrete technical results already delivered and aligned on the next phase, where EnergyGuard moves decisively from framework building to large‑scale experimentation and validation of AI solutions under real operating conditions.

Since the start of the project, R&D Nester has played a central technical role, acting as the reference facility for large‑scale transmission system testing within EnergyGuard, and leading the Work Package 6 on the coordination of the demonstrations.

To date, R&D Nester has been actively involved in four work packages and in the following core technical outputs: 

Digital Twin of Transmission Network with high-share of renewables:

R&D Nester contributes a high‑TRL digital twin of a transmission system with realistic size, capable of offline, real‑time and hardware‑in‑the‑loop simulations. This asset underpins EnergyGuard's smart‑grid pilot and enables testing of AI solutions against realistic network dynamics, protection behaviour and communication constraints. 

Smart grid pilot leadership (WP6 - Pilot 1):

R&D Nester leads the pilot focused on AI services for smart-grids operation, including:

  * Real‑time co‑simulation of transmission systems

  * AI‑based fault detection and classification using COMTRADE data from protection relays

  * Model‑free voltage profile estimation techniques for low‑observability Low Voltage networks

PROJECT'S ACHIEVEMENTS: A TEF TAKING SHAPE

Across the consortium, several tangible achievements were highlighted in Soria: 

  * The EnergyGuard architecture is fully defined, integrating data services, AI development sandboxes, HPC resources and an acceptance environment for trustworthy AI.

  * Digital twins from all five facilities (transmission network, microgrid, hydrogen platforms, buildings and energy communities) are being operationally connected to the platform.

  * Early versions of the AI development testing sandbox are running, enabling end‑to‑end experimentation workflows.

  * Pilot experiments are moving from design to execution in smart grids, microgrids with high-RES penetration and hydrogen systems.

  * The foundations of the common AI risk database for the energy sector are in place, linking technical testing to regulatory and ethical assessment.

The Soria meeting clearly showed that EnergyGuard has moved beyond "architecture on slides" into hands‑on experimentation.

FROM MEETING ROOMS TO FIELD REALITY: TECHNICAL SESSION AT CEDER‑CIEMAT

A highlight of the meeting was the technical visit and working session at the CEDER‑CIEMAT microgrid, a real operational system combining wind, photovoltaics, batteries, EV charging and grid‑connected operation.

Partners discussed how AI models trained in simulation behave when confronted with:

  * Fast renewable intermittency

  * Non‑ideal measurements

  * Storage constraints

  * Network topology changes in radial and ring operation

This on‑site session reinforced one of EnergyGuard's core principles: trustworthy AI cannot be validated without exposure to real‑world complexity.

For R&D Nester, the discussion closely mirrored challenges already seen at transmission level, strengthening the conceptual bridge between microgrids and large‑scale networks.

OBJECTIVES FOR THE REMAINING MONTHS: FROM VALIDATION TO IMPACT

Looking ahead, the EnergyGuard's ambition is clear for the second half of the project: 

  * Deliver fully validated AI use cases across grids, hydrogen, buildings and energy communities.

  * Populate the European AI ecosystem with trustworthy, test‑backed energy AI assets.

  * Demonstrate that rigorous AI testing can reduce time‑to‑market, cost and risk for critical energy applications.

  * Lay the technical and business foundations for a self‑sustaining EnergyGuard TEF beyond the project lifetime.

For more information:

EnergyGuard website

EnergyGuard Project @ R&D Nester website