50% more control resolution unlocks additional savings and comfort potential.
After achieving record savings during the winter 22-23, NARA AI, ECCO2’s industry-leading artificial intelligence engine is getting even better.
A combined team of ECCO2’s data and energy specialists designed a new heating distribution logic improving indoor comfort while unlocking even more savings.
They used the Terabyte-size database collected by ECCO2’s High-Resolution Sensing IoT solution to develop, simulate, fine-tune and backtest the improved algorithm against a large volume of real data.
Once this work was completed, the R&D team implemented software and database changes, enabling a 50% increase in control resolution for NARA AI and its underlying database HyperCube.
The day is now split in 6 control and monitoring periods called “Timeblocks” (up from 4 previously) supporting the deployment of an ECCO2-proprietary Night Setback algorithm. The screenshot below shows the corresponding fine-tuning patterns that NARA AI custom-designs for each heating group, based upon High Resolution Sensing and Machine Learning.
The backtesting predicts additional average savings of 1%, while automatic fine-tuning of the control parameters to each heating group will preserve tenant comfort.
Antoine Eddé, Product Strategist at ECCO2 stated:
“NARA AI V3 is a big deal: it implements AI-powered building-specific Night Setback that challenges conventional wisdom on the matter.
The topic of Night Setback is highly debated among specialists.
Many experts believe that the idea of reducing losses during the night makes sense in theory, but the additional complexity and comfort loss yield questionable real-world results. We at ECCO2 blame this on rigid existing heating control logic and lack of available technician time for tuning. And we asked ourselves: could NARA AI solve this problem?
We started a simulation project using our “data gold mine” – the HyperCube database. This research was successful, demonstrating once again the usefulness of ECCO2’s High Resolution Sensing and the scientific value of the HyperCube database.
As a foundation for a successful implementation, we increased the time resolution of the NARA AI control system by 50%, laying the ground for even better analysis and simulation, consistent with our vision of Swiss precision in heating control.
We are ECCO2, and we look forward to an even better contribution to a World in Balance”.