PILLAR: SMART NETWORKS
Envision and Foresight

What we will do and why
Application of machine learning to model and estimate utilisation at low voltage substation and transformer level, targeting reduction in the need to install physical sensors and monitoring.
Key deliverables
- Model estimating capacity for all secondary substations.
Target outcomes
- Targeted, proactive reinforcement.
- Ensuring customer LCT connections.
Measures of success
- Reduction in number of capacity constraints across the LV network.
- Customer satisfaction.
Stakeholder Engagement
We employed third party data-science expertise in the development of the model, checking results with real measurements through work with field staff. Work involved collaboration between data scientists, subject matter experts, IT professionals, and business stakeholders. In parallel with Envision, we developed alternate models using different methodologies and compared results. Through collaboration between Distribution Network Operator (DNO) and Distribution System Operator (DSO) data teams we have arrived at a process for use of these predictive models in our operations.
Public Detailed Docs
Information on the original project:
Envision
UKPN machine learning simulation to unlock 70MW of network capacity
Planned start:
01.04.2023
Target completion:
01.04.2024
Key milestones:
- Base model created (Oct 2023)
- Internal apps and dashboards (Dec 2023)
- Enhanced model with smart meter data (Apr 2024)
Status:
Complete
Date of change | Category | Description |
---|---|---|
28.06.2024 | Update |
Project successfully completed in 2024. No other change. |
24.12.2023 | Update |
We added links to additional information and added a summary of our stakeholder engagement as part of Envision and the further development we have been doing to this model under the title “Foresight”. |