Envision developed a software-based machine learning tool designed to generate insights into Low Voltage (LV) networks more quickly and cost-effectively than traditional physical monitoring methods. The project marked a significant step toward expanding the flexibility market and supporting the development of a smart grid—one that would enable cleaner, greener energy resources to connect faster and at lower cost.
What is the project about?
Detailed visibility of the network was critical as we planned for a low-carbon, Net Zero future. In simple terms, ‘visibility’ referred to the collection of real-time data on the network’s performance—ranging from power flow patterns to areas of high electricity demand and the times at which that demand occurred.
We explored how our existing data could be leveraged to model demand and enhance the visibility of the Low Voltage (LV) network. By incorporating external data sources alongside internal datasets, we were able to build a more comprehensive and accurate picture of current network activity, ultimately helping us improve the service delivered to customers.
As more renewable energy sources connected to the network and the adoption of electric vehicles and electric heating increased, having accurate, granular data enabled us to plan and prioritise infrastructure investment—ensuring it was deployed at the right time and in the right locations.
How we’re doing it
The Envision project explored new approaches to modelling demand and improving visibility across our Low Voltage (LV) network to support both planning and operations. The first phase focused on developing and trialling a software-based machine learning tool designed to enhance network visibility and generate fresh insights into demand. This approach significantly reduced costs and delivered results more quickly than traditional monitoring methods.
The second phase examined the availability and value of third-party data by engaging with organisations that held relevant datasets. This included data from distributed energy resources (DERs), electric vehicle charge point operators, local heat networks, and community energy schemes. The insights gathered helped us assess whether, and how, these external datasets could be used to further improve visibility and understanding of our network.
What makes it innovative
Envision developed a first-of-its-kind, machine learning-based software tool that produced a predictive model of the Low Voltage (LV) network, delivering deeper insights than previously possible. This innovation demonstrated the significant value of applying artificial intelligence to network-specific challenges, highlighting its potential to transform how electricity networks are monitored, understood and managed.
What we’re learning
Envision generated new insights into our Low Voltage (LV) network by modelling predicted load, enabling our teams to plan more efficiently and effectively for a Net Zero future. The key outcomes of the project are summarised below:
Enhanced Network Visibility: We developed and validated data models for both ground-mounted and pole-mounted transformers, providing estimated load values across all LV substations. The models achieved high prediction accuracy, with 83% overall confidence for ground-mounted transformers and 61% for pole-mounted units.
Improved Planning Tools: We created a tool to help prioritise where hardware-based monitoring devices should be installed. In addition, we developed a user interface to support network planning and Distribution System Operator (DSO) teams in making informed decisions on network reinforcement deferral and flexibility procurement.
Expanded Flexibility Procurement: The predicted load values from our modelling enabled a more targeted approach to flexibility procurement, particularly at unmonitored sites, improving the efficiency and impact of procurement strategies.
We use cookies to give you the best possible experience when using our website. Many of these cookies are essential to the efficient operation of this site.
