AI Screening Makes Energy Performance Certificate Data Operational in Future District Heating Work – HOFOR A/S

Customer: HOFOR A/S

In brief

How do you find the buildings that require special attention when the temperature in the district heating network is lowered? For HOFOR, 4BC has combined energy performance certificate data, meter data, and knowledge to create an AI model that screened approximately 2,000 buildings and established a data-driven basis for prioritizing further efforts to prepare the building stock for future district heating.

HOFOR is working to lower the temperature in the district heating network. This reduces heat loss in the network and increases the efficiency of the heat pump that produces the district heating. At the same time, it places greater demands on the connected buildings and their heating systems.

In this context, 4BC has investigated how existing energy performance certificate data can be combined with HOFOR's own meter data to identify patterns and prioritize buildings where a closer technical investigation is expected to provide the greatest value.

EnergyCheck has been an essential part of this work. The platform has made the extensive technical data from energy performance certificates structured, searchable, and comparable across many buildings. The result is not an automatic technical diagnosis of individual buildings, but a screening and prioritization tool that helps HOFOR move from large data volumes to a more targeted effort.

From Thousands of Data Points to the Right Buildings

HOFOR's operational data and energy performance certificate data exist at different levels and can rarely be linked one-to-one between meters, BFEs (Building and Dwelling Register units), and buildings. By combining energy expertise, data processing, and technology, data was structured and linked so that energy performance certificates could go from static documentation to an active tool in district heating work.

Results

From Energy Performance Certificates to Prioritized Effort

4BC has developed a risk screening where buildings can be ranked according to their relevance for further investigation. This allows HOFOR to target analyses and technical investigations rather than treating all buildings uniformly.

Known Problems Are Concentrated, and New Candidates Are Found

On test data, the model concentrates a large proportion of the known high return temperature problems among the highest-ranked buildings and meters. Simultaneously, the analysis identified 114 high-risk installations that had not previously appeared on HOFOR's high return temperature list. These represent potential candidates for more targeted technical control.

EnergyCheck Makes the Analysis Applicable and Transparent

The AI model can combine many variables and find patterns across the portfolio. EnergyCheck supplements the model with a more transparent approach, where concrete technical criteria can be directly examined across thousands of energy performance certificate data.

The project thus demonstrates the value of combining advanced data analysis with simple, energy-technically understandable screening rules: AI can help find patterns and surprising cases, while EnergyCheck makes it possible to investigate, explain, and operationalize the results.

Key figures