Back
Efficient and Effective icon
Efficient and Effective

Storm Resilience

Storm Resilience is developing new software to reduce the number and length of power cuts during a storm. The project could help restore power supplies caused by lightning strikes up to 90% faster, and ensure we have engineers in the right places to reduce the time taken to repair faults caused by a storm.

Project-on-a-Page summary

Project data

Start date: 01/12/2019
End date: 16/12/2021
Budget: £665,000

UK Power Networks’ ‘Storm Resilience’ project developed an advanced tool that combines network data, historic fault data and live weather forecasts to predict the number of faults that could occur in an area of our network. This is particularly helpful in times of stormy weather to ensure enough engineers are on standby. A separate part of the project trialled a lightning tracking software to help restore power supplies caused by lightning strikes up to 90% faster.

This project supported the drive to be an even more reliable network operator through improving resilience during severe events. It was split across two separate initiatives;

  • Work stream 1 – Lightning into PowerON
  • Work stream 2 – Resource Estimation Tool

Work Stream 1:

We trialled a proof of concept where UK Power Networks’ Network Management System (PowerOn) received lightning strike locations in real time via an API developed by an international weather consultancy. These locations were then linked to poles and to the network diagram. After this mapping was established, a dedicated alarm was created in PowerOn to notify control engineers that the faulted circuit was likely struck by lightning. This could reduce the time customers don’t have supply due to lightning strikes.

Work Stream 2:

We gained access to advanced weather forecasting from stations across the UK Power Networks licence areas. We trialled the concept of using predictive analytics to combine historic fault data and weather parameters. This built on and enhanced our existing capabilities to forecast the impact of severe weather. Using innovative technology we were able to develop the link between high frequency sampled weather data and the distribution network. These forecasts are now helping us predict the impact of a storm and quantify the expected level of risk each weather event presents.

Lightning into PowerOn meant that, for the first time ever, real-time lightning data was available our network management system. The system is now identifying when a fault is most likely caused by lightning. Using lightning data to this accuracy had never been trialled before.

In the trial, the system captured 37 faults that were due to a suspected lightning fault. The functionality has now been built into our automatic power restoration system to enable us to restore a power cut due to lightning in less than 3 minutes.

The Resource Estimation Tool combined network data, historic fault data, and advanced weather forecasts to predict the number of faults each region will experience during bad weather. This created a ‘probabilistic fault forecast’, which has never been trialled for a UK electricity network. The tool automatically predicts where and when to allocate resources and staff hours far more accurately than humans can. Machine learning will be used to help improve the tool over time. ​

We trialled Lightning into PowerOn and Storm Resource Forecasting throughout 2021.

The lightning trial captured 37 events. This was sufficient to prove the concept on the live system, but was not sufficient to lead to a long term business change. The business has set a target of 200 faults to be recognised by the system before making the solution business-as-usual. This will allow for a thorough verification to take place that the accuracy is as high as required for full deployment. This will not be part of the Storm Resilience project, but will take place as part of follow-on activity.

The resource forecasting tool proved it was capable of forecasting, planning and responding to severe weather events effectively. During the trial we measured how well it predicted outages for different types of weather across our regions. We found that for precipitation up to 20mm per day, the tool worked very well for all regions. For days with precipitation above 20mm per day, the tool slightly overestimated faults for SPN and slightly underestimated faults for EPN.


About Innovations

Find out more

How projects are funded

Find out more