Calibrate · Sardinia (existing)
Setting up the twin for a facility that already exists
An existing facility skips Design — Calibrate rebuilds the same ledger from measured geometry, a year of meter readings, and harvest records. From here on, Plan and Operate work the same as for a new build.
Envelope captured
1,200 m²
single-span structure · EVA film
Meter baseline
12 mo
electricity + water, monthly bills
Model fit
±8%
monthly energy, model vs meter
Footprint baseline
1.1
kg CO₂e / kg basil · measured energy
Calibration ledger — what was captured, and how well the model fits
The same physics model used for Agadir's design, run backwards: parameters are tuned until modeled monthly energy matches the meter. Deviations flag what the sensors should watch first.
| Ledger entry | Source | Modeled | Measured | Status |
|---|---|---|---|---|
| Geometry & envelopespan, height, film transmissivity | on-site survey | — | captured | done |
| Annual electricityfans, pumps, supplemental LED | 12 monthly bills | 38.9 MWh | 41.8 MWh | −7% · within band |
| Summer peak monthventilation-dominated | July bill | 5.1 MWh | 5.9 MWh | −14% · sensor priority |
| Water use | meter | 10.4 m³/mo | 11.3 m³/mo | −8% · within band |
| Yield per cycle | harvest records | — | 0.98 kg/m² | baseline set |
| Cycle length by season | harvest records | — | 30–41 days | 1 season · needs a full year |
Why this matters
Live sensors don't just monitor — they close the loop on these numbers.
The ±8–14% gaps above are the design assumptions Agadir's business case also leans on (ventilation loads, cycle lengths in a Mediterranean summer). When the Sardinia feed goes live, every week of sensor data tightens both twins: the existing site calibrates the new build. That is the multi-site argument in one sentence.