CEA Advances Tools for Energy Cost Reduction and Comfort

Within the EVELIXIA project, CEA is actively developing tools aimed at reducing energy bills and improving building occupant comfort, with a primary focus on Service 1, while also contributing to additional services across the project.

Significant progress has already been achieved. CEA has processed data from multiple pilot sites, including the Austrian pilot via the EVELIXIA platform, as well as data from the French and Finnish pilots, provided through CSV files by ENTECH and Turku University of Applied Sciences.

At the Austrian pilot site, CEA successfully identified the thermal capacitive and resistive parameters of the building model for townhouse #4. This step is essential for the operation of IS1 “Indoor Air Quality” as well as IS10b “HVAC Management System”. In parallel, a machine‑learning model representing occupant behaviour, specifically window‑opening patterns, was trained using real data from the same building. This behavioural model is a key component for both IS1 and IS10b.

Progress has also been made on battery‑related services. For the French pilot site, CEA identified the parameters of the model representing the 40‑foot container hosting a battery energy storage system involved in aFFR grid services. This activity supports the development of IS2‑1 “Battery Cooling System Monitoring”.

In addition, CEA developed a data‑driven battery ageing prognosis model using a machine‑learning approach. The tool was tested using two data resolutions—10 seconds and 1 second—to assess accuracy in estimating discharged capacity. The results showed an error of around 2% at a 10‑second interval, leading CEA and the University of Turku to adopt a 5‑second data resolution to remain below a 1% error threshold. This achievement marks an important step forward in the development of IS2‑3 “Battery Ageing Prognosis”.

CEA has also analysed datasets provided by the Romanian, Danish, Greek, and Spanish pilot sites, identifying several issues such as inconsistent time intervals, data gaps, and missing inputs required for the tools to function properly. These findings have been reported to support data quality improvements across the project.

The next phase of work will focus on delivering an initial integrated demonstration of EVELIXIA’s innovative solutions. This will begin with Service 1 “Daily Optimisation for Building”, including IS10b for the ENTECH building, with office and workshop areas considered separately. In parallel, Service 3 “Energy Assets Maintenance” will be demonstrated through IS2‑1 and IS2‑2 at the French and Finnish pilot sites.

In the coming weeks, CEA also plans to finalise the collection of pending data from the Romanian, Danish, Greek, and Spanish pilot sites, further supporting the large‑scale validation of EVELIXIA services.