UPDATE

Satelline

Project Data

Start date:

09/01/2023

End date:

03/31/2024

Budget:

£413,000

What is the project about?

Vegetation management is a critical activity for all utilities and is typically their largest operations and maintenance expenditure. Vegetation management costs have been increasing while customers and regulators expect reduced numbers of interruptions. The majority of tree trimming is planned and based on Light Detection and Ranging (LiDAR) surveys. LiDAR scans are typically run every few years, measuring the distance between the electricity conductors and the vegetation (clearances). Areas with lower clearances are prioritised for trimming.

Project Satelline, partnered with Boston Consulting Group (BCG), explored how a satellite and AI solution could be implemented at UK Power Networks to inform trimming plans instead of the LiDAR-based approach. Satelline leverages satellite imagery and AI to measure clearances between conductors and vegetation. It then conducts AI analysis to prioritise areas for trimming based on the clearance and other criticality factors (e.g., historical outages). The solution aims to reduce the number of vegetation-related outages and optimise trimming spend.

The project focused on proving that a satellite and AI solution was capable of delivering significant benefits across the UK Power Network’s network while having the required accuracy to inform trimming plans.

How we’re doing it

The Satellite and AI solution was tested in a pilot, which covered 400km of overhead lines in Rotherfield and Polegate. Clearances were measured for c. 3,500 spans by identifying the closest vegetation to the conductor with satellite imagery. This measured both radial clearance (closest 3D distance from the conductor to the vegetation) and encroachment type (linear vegetation growth type, e.g., side growth, under growth, through growth). The solution’s clearance output was validated in the field by measuring the distance between the conductors and the vegetation for a random selection of spans. The solution’s clearance output and other UK Power Network’s data have been uploaded into the end-to-end solution. This is a SaaS platform, which creates the trim plans and monitors the progress of the contractors, assigned to complete trimming.

In parallel, a business case for implementation across the entire network was developed.

What makes it innovative

The use of satellite imagery and AI analysis is at the cutting-edge of vegetation management. It enables:

More frequent data updates (at least annual) compared to LiDAR as data acquisition is more cost-effective while data analysis can be completed within 2 weeks (vs 6+ months with LiDAR). This data frequency means that faster growing species can be captured in trim plans regularly.
Use of AI growth models that analyse historical imagery to calculate vegetation growth rates to predict when vegetation should be optimally trimmed.
On-demand data as satellite imagery can be purchased up to three times per day, which can feed into spot-checks after storms and also for automated post-trim scans. These scans can validate trimming carried out by contractors.
More frequent risk tree detection through specialised models, which identify unhealthy trees within striking distance of conductors.

The solution comes as an end-to-end product: all analysis is done within the central platform, which will reduce frictions in current operations and make the process simpler.

What we’re learning

The project found that the accuracy of the solution measuring the clearances between the conductor and the nearest vegetation is very high at ~95%. This was validated in the field by measuring the actual distance with a laser-measuring device. These results will improve further as they are fed back into the solution to train the AI.

The high accuracy of the solution enables it to deliver its associated benefits. From peer experience, the number of vegetation-related outages can be reduced by 25%+, which will generate significant IIS incentive payments. Through targeted trim plans, trimming spend is optimised to maximise impact. Overall, we have learnt that a satellite and AI solution is highly accurate in the UK Power Networks’ area and its range of benefits can provide a

significant boost to tree-cutting teams, as the full scale implementation across the network is considered.

Reports & Documents

Satelline final report

Share your ideas

If you have an idea that could reshape or revolutionise the way we work, we want to hear from you.

Submit an idea

Satelline

On this page

Project Data