WP4: Agile Fault Detection and Accurate Failure Prediction

Objective

Develop fast fault diagnosis and proactive failure prediction algorithms that improve IoE resilience and reliability.

Lead: Catalonia Institute for Energy Research (IREC) · Timing: M7-M42 · Main ESR involvement: ESR10-ESR11

Key Tasks

• Use deep semi-supervised learning and echo state networks to detect subtle node-level faults with low delay.
• Improve rare-fault performance through few-shot learning and robust feature extraction from mixed labelled and unlabelled data.
• Predict cascading failures using denoising autoencoders, multi-graph convolution, and multi-scale temporal modelling.

Programme Links

This work package contributes primarily to RO-4 and is part of the proposal-stage programme structure captured in the SAILING Part B document.