United KingdomEnergy and utilities2023 – ongoingSSE Energy Solutions
Machine learning optimization of combined heat and power networks
SSE Energy Solutions cut manual forecasting time 12x and saved £350K a year on combined heat and power networks
Problem
SSE Energy Solutions, part of SSE plc, ran machine learning optimization models for its combined heat and power networks on virtual machines. Each run needed heavy manual input from data scientists, and limited compute made scaling costly.
Approach
SSE Digital Services moved the workflow to Azure Databricks. Pricing data loads were automated, and optimization runs were scheduled and run in parallel on clusters. Results went to Power BI dashboards, and the outputs feed optimization for four CHP-powered networks across the UK, with weekly jobs at each site.
Results
Takeaway
Automating the optimization runs freed data scientists from a weekly manual forecasting task and tied the models to a yearly cost saving.
Categories
- Size
- Large enterprise
- Industry
- Energy and utilities
- Business function
- Production and quality
- Innovation type
- Process innovation
- AI technology
- OptimizationPredictive analytics and ML
- Adoption
- Off-the-shelf tool or platform
- Evidence
- Vendor case study
Sources
Source summaries
Databricks describes how SSE Energy Solutions moved its CHP network optimization models from virtual machines to Azure Databricks. It reports a 12x cut in manual forecasting time, £350K in yearly savings and forecasting jobs that now take two hours or less. The story does not say how the savings were calculated.
Full details are in the original sources.