Softwire and NGED team up to pilot AI-powered drone inspections of remote transformers
Drone imagery and computer vision pilot aims to detect transformer deterioration earlier, improve asset health reporting and target maintenance to help prevent failures and power cuts.
Softwire is collaborating with National Grid Electricity Distribution (NGED) (opens in new tab) on a Network Innovation Allowance (opens in new tab) (NIA) funded project to develop and test a proof of concept (opens in new tab) using drone imagery and machine learning to assess the condition of pole-mounted transformers.
The project is exploring how aerial imagery, AI, and computer vision can automate the condition assessment process, identifying corrosion and other visible defects to produce consistent asset condition data. This could help improve asset health reporting, target maintenance more effectively, reducing the cost and effort associated with manual inspections and assessments.
Detecting deteriorating transformers earlier could allow NGED to act before transformers fail, avoiding unplanned power cuts, particularly in rural communities, and limiting safety risks and environmental damage from oil leaks.
Trial locations for the project have been selected across NGED’s network regions, including remote and rural areas, to capture a representative range of asset conditions. Drone inspection specialists, Team UAV (opens in new tab)has captured high resolution images of 500 pole-mounted transformers. Softwire is developing the computer vision models and a prototype web application to present the results.
Working with NGED specialists, Softwire's data science team will label the images and train models to assess asset condition. The team will also consider how factors including weather, light, location and seasonal variation could affect model performance, and ensure the training and test data represent the range of conditions captured during the trial.
The prototype application is expected to map its condition assessments, confidence levels and supporting rationale against the Common Network Asset Indices Methodology, the standard used by electricity networks to assess asset health and risk. This could support future reporting to Ofgem through the Network Asset Risk Metrics framework during Electricity Distribution 3 (ED3), the next period for which Ofgem will set funding and performance requirements for local electricity networks. The model could later be expanded to assess a wider range of conditions and support a predictive approach to maintenance, reducing costs and improving network reliability.
Marcus Gardiner (opens in new tab), Client Director, Energy & Utilities at Softwire (opens in new tab), said:
“Critical infrastructure providers face a major challenge: how to monitor large, geographically dispersed asset bases closely enough to identify deterioration early and lay the foundations for predictive, targeted maintenance. This project brings together drone technology, data science and software engineering to test a smarter, more scalable approach.“We are delighted to be working with NGED and Team UAV on an innovation project with the potential to improve how hard-to-reach assets are inspected and managed. It is a powerful example of how data science and engineering expertise can be applied to complex real-world problems, supporting better operational decisions and the long-term resilience of essential public infrastructure.”
The initial phase is due to be completed by the end of 2026. It will include the trained proof-of-concept model, prototype web application and guidance for future drone image collection.
Favour Eboh (opens in new tab), Innovation Engineer at National Grid Electricity Distribution (opens in new tab), said:
“Pole-mounted transformers play a critical role in delivering electricity to homes and businesses, particularly in rural communities.“This innovation project will explore whether aerial imagery, AI, and computer vision can automate aspects of the condition assessment process, generating more consistent condition data to inform asset health reporting and investment decisions.“Working with Softwire and Team UAV, we will test the technology on an initial set of transformers and assess its potential to support future network asset management.”
Mark DeGiovanni (opens in new tab), Technical Lead, Team UAV (opens in new tab), said:
“Working with NGED and Softwire has been a great opportunity to demonstrate how specialist drone operations and data science can work together. By combining our experience in infrastructure inspections and data capture with Softwire's expertise in AI and software development, we're helping to explore a more efficient approach to assessing asset condition, with the potential to support better maintenance decisions across electricity networks."
Subscribe
to our monthly newsletter for our latest expert content.
