IN Brief:
- National Grid has secured £355,985 for Project FoSMo through Ofgem’s Strategic Innovation Fund.
- The project will develop shared AI models for electricity network asset monitoring.
- Partners include Keen AI, SP Electricity North West, UK Power Networks, SP Energy Networks, and Scottish Hydro Electric Transmission.
National Grid has secured £355,985 in Alpha-phase funding for Project FoSMo, a shared artificial intelligence model for electricity network asset monitoring.
The project is funded through Ofgem’s Strategic Innovation Fund and will be led by National Grid Electricity Transmission in collaboration with DScience Ltd, trading as Keen AI. Network partners include SP Electricity North West, UK Power Networks, SP Energy Networks, and Scottish Hydro Electric Transmission.
FoSMo will develop and test shared AI models for monitoring transmission and distribution assets. The project is intended to improve fault detection, asset reliability, and operational efficiency by applying common models across network companies, rather than training separate systems within individual organisations.
The funding forms part of a wider National Grid innovation package worth more than £1m across five projects. Other funded workstreams cover flexible demand from large electricity users, probabilistic network planning, marine habitat restoration for offshore grid expansion, and a digital tool connecting renewable developers with community energy groups.
Electricity network operators are handling larger volumes of asset data as inspection methods become more digital. Drone imagery, fixed sensors, thermal surveys, historical maintenance records, and visual inspection datasets are now part of everyday asset management. Manual review remains important, but the scale of data being produced is difficult to process consistently without automated support.
A shared AI model could give network companies access to broader training data and more consistent assessment methods. Transmission lines, substations, pylons, insulators, conductors, and related components vary by geography and age, yet many condition indicators and fault categories are common. A model trained across several networks may be able to identify patterns that would be harder to detect within a single operator’s dataset.
Critical infrastructure AI also brings governance requirements that ordinary analytics projects do not face. Data quality, cybersecurity, model validation, decision accountability, and integration with maintenance processes all determine whether a model can support engineering work. Asset-monitoring AI cannot simply identify a visual anomaly; it has to support decisions that affect maintenance priorities, outage planning, and long-term network resilience.
The project sits alongside wider investment in physical grid capacity. A recent proposal seeking £4.5bn of transmission approval showed how reinforcement and asset upgrades remain essential across England and Wales. Digital monitoring does not replace substations, conductors, pylons, transformers, or switchgear, but it can help operators use maintenance budgets and outage windows more effectively.
FoSMo also follows earlier work across autonomous inspection, dynamic line rating, and digital planning. Those technologies are part of a wider transition from periodic, manually intensive network management toward more continuous condition assessment. The shift becomes more important as electrification increases load, renewable connections expand, and ageing assets remain in service for longer.
The Alpha phase will need to test the model under real operating conditions. AI systems can perform strongly on curated datasets yet struggle with inconsistent imagery, weather variation, unusual assets, incomplete labelling, or rare fault types. Field deployment will therefore need robust validation against engineering judgement and existing inspection practice.
If the project progresses successfully, shared AI could become part of the standard condition-monitoring toolkit for electricity networks. Britain’s power system is being asked to connect more renewables, accommodate larger loads, support flexibility, and maintain reliability while major reinforcement programmes move through planning and delivery. Better asset intelligence gives operators another way to manage that pressure with greater precision.
