Optimising decision making in response to unexpected outages

NESO FastOut

£MMs

In potential reductions to GB-wide system costs

From hours to seconds

Time to triage requests

We built an AI-powered advice tool to help prioritise outage requests, which can improve operational efficiency and reduced costs.

Background

Deciding when a generation asset should be taken offline for maintenance, or because of unexpected issues such as storm damage, is a complex and time-consuming task. This often results in poor decisions, a backlog of outages, and regulatory penalties. We continually collaborate with the National Energy System Operator (NESO) team to identify AI use cases, so we developed a proof of concept to tackle this challenge.

Solution

We helped rapidly triage and assess outage requests using machine learning to draw associations between past outages and constraint limits. Each potential slot is flagged as low, medium or high cost, giving engineers the information they need to make a decision about whether it can go ahead.

Impact

Requests are now triaged in seconds instead of hours, which could reduce the risk of supply shortages and enable more effective network scenario planning - potentially saving millions of pounds in system costs. Learn more about the project, funded by the Network Innovation Allowance (NIA) mechanism here.

background placeholder
AI-powered

“Supporting the development of NESO’s AI Centre of Excellence, in this case through a targeted outage planning support tool, is critical for maintaining network stability.”

Niko Louvranos, Business Development Director

Faculty

discovery.