Context Challenge Results News Consortium

Firmware for Resilient Embedded Algorithm Microarchitectures

FIRMAR

Context

The digitalization of the electrical grid requires integrating automation, protection, and monitoring systems into key infrastructure and assets.

Technologies such as Federated Learning, Edge AI, and Virtualization allow for the execution of algorithms on distributed local equipment.

These devices are generally based on microcontrollers with limited computing and memory capacity – Resource Constrained Devices (RCD) – which must ensure safe and efficient operation in critical applications.

CHALLENGE

The FIRMAR project aims to research firmware development environments and architectures that allow for the cyber-secure incorporation of advanced algorithms for electrical networks into Resource Constrained Edge Devices (RCD).

INFOGRAPHIC

EXPECTED results

Design of firmware architectures and execution environments for complex algorithms for RCD devices, with new functionalities required by electrical networks, and analysis of the transfer of models deployed in the cloud to Edge-RCD devices.

Deployment of advanced algorithms in RCD devices: modeling strategies applicable to different RCD families, orchestration methodology to implement green algorithms, and customized cybersecurity solutions, including post-quantum security.

Proposals for low-capacity but quantum-secure System-On-Chip (SoC) architectures and implementation design on CMOS technology.

Use cases:
– Arteche Intelligent High Voltage Transformer Management (GIAT): monitoring of instrument transformers combining direct measurement with mathematical models and advanced algorithms.

– Ingeteam hybrid platform for critical systems control with advanced analytics capability: use of RCD hardware to integrate critical real-time functionalities into Edge nodes in Substations / Transformation Centers.

– Advanced Home Area Network Controller (HANC) by Tecnalia: providing flexibility in demand management and deferrable generation, using algorithmic models and predictive control.