GridRevive

Project Description

The GridRevive project aims to develop and demonstrate a governed AI-based risk-to-plan decision-support solution that enables DSOs to plan maintenance, reinvestment and reinforcements for ageing underground MV cables under rising electrification demand. Using existing DSO asset registers, fault history, GIS/context, load and consequence/cost data, the solution will produce transparent, explainable and auditable recommendations for multi-year planning and documentation.

The project will:

• Develop uncertainty-aware fault vulnerability models that are robust to incomplete or uncertain legacy asset records and provide confidence measures and explainable risk drivers.
• Implement a controlled model-operations pipeline for performance monitoring, drift-aware updating, model versioning and audit logs.
• Deliver multi-year, budget- and reliability-constrained planning and what-if scenario tools, supported by a guardrailed LLM-enabled natural-language interface.
• Integrate the prototype into KMD's platform and validate it with Dinel and Vores Elnet through operational shadow testing and benchmarking against current planning baselines.

Expected Green Outcomes:

GridRevive supports the green transition by enabling more targeted renewal and reinforcement of distribution grids. Better identification of vulnerable cable sections can reduce premature replacements, unnecessary excavations, emergency repairs and outage impacts, while supporting cost-effective integration of electrification and renewables, improved security of supply, and longer asset lifetimes.

Project Summary