Indoor climate baseline
The indoor climate baseline lets myCoreAI factor a building's historical indoor climate into how aggressively it optimizes, rather than relying only on configured temperature limits. It is one of the mechanisms gated by AI Freedom. The intent: give myCoreAI more room to save energy where the building has historically tolerated wider comfort bounds, without ever operating outside conditions it has actually run in.
How the baseline is calculated
myCoreAI builds the baseline from reference data — historical indoor-temperature data from periods when myCoreAI was not in control. From that data it computes reference limits: a per-sensor minimum and maximum that approximate the building's historical comfort bounds.
Applies only to indoor temperature sensors. CO₂, supply temperatures, and energy meters are not affected.
Outliers are filtered before the limits are computed; a single cold or warm hour does not skew the result.
Requires sufficient seasonal coverage. If the reference period lacks enough winter and summer data, the baseline cannot be produced and myCoreAI falls back to the configured limits.
How the setting behaves
The baseline is controlled together with AI Freedom and is effectively binary:
0
Off — only configured limits are used.
> 0
On — any value above zero applies the full baseline.
When active, myCoreAI compares two candidate sets of limits per sensor:
The configured limits set in the platform.
The reference limits derived from history.
For the lower bound, myCoreAI picks whichever set has the lower average during office hours — the one that gives more room to push temperatures down. For the upper bound, it picks whichever has the higher average. If the reference set is not the outermost one, myCoreAI shifts the reference limits so their average matches the configured average; the shape of the reference curve is preserved, but the level aligns with what was configured.
The baseline only ever widens the comfort bounds. It never tightens limits beyond what was configured.
Office hours
myCoreAI separates office hours from the rest of the week, because tolerated comfort bounds often differ between workday and nights/weekends. Office hours are fixed for all buildings at 07:00–18:00, Monday to Friday. Each reference limit therefore has one value for office hours and another for non-office hours.
In a building with a consistent climate, the two values are nearly identical. In a building with strong day/night or weekday/weekend patterns, they diverge more.
Safety floor for historically warm buildings
If a building has historically operated warm but is configured with relatively cold lower limits, applying the reference minimum verbatim could push the building colder than it has ever run. A safety rule prevents this:
If the average reference minimum is above 20.7 °C but the configured minimum is below 20.7 °C, myCoreAI adjusts the chosen limits so their average lands at 20.7 °C.
Worked example
An office building has been configured with a lower indoor temperature limit of 20.0 °C, but two heating seasons of reference data show the building has never actually operated below 21.4 °C during office hours. With AI Freedom set above zero, myCoreAI selects the reference minimum as the outermost lower bound, then applies the 20.7 °C safety floor: the chosen lower limit is shifted so its office-hours average is 20.7 °C instead of 20.0 °C. The upper bound is left at the configured value, since the reference maximum sits inside it. Result: myCoreAI gains a realistic lower bound to work against during night setback and morning ramp-up, without ever driving the building below temperatures it has historically held.
What this means in practice
Configured limits remain in control. The baseline only widens the comfort bounds.
Buildings with rich historical data get more flexibility, because more evidence exists for what the building has tolerated.
Buildings with thin or unrepresentative history fall back to configured limits and behave conservatively.
Recently onboarded buildings see little or no baseline effect until enough off-period data accumulates across seasons.
Last updated
Was this helpful?