COOL — Cooling Operations Optimization Layer

The Cooling Operations Optimization Layer (COOL) measures how efficiently your facility converts incoming power into useful compute — the fraction lost to cooling overhead. COOL uses Power Usage Effectiveness (PUE) as its primary input and scores against continuous linear benchmarks derived from operational data.

Primary metric

cool_efficiency_score — a number from 0.0 to 1.0 derived from your reported or measured PUE.

The scoring is continuous and linear. The benchmark anchor is PUE 1.20 (excellent liquid cooling or free cooling), which scores 1.00. PUE 1.60 scores 0.00. There is no step function — every tenth of a PUE point matters.

PUE	Score
1.20	1.00
1.25	0.84
1.35	0.60
1.40	0.46
1.50	0.23
1.60	0.00

Formula

The cool_efficiency_score is a continuous linear function of PUE:

cool_efficiency_score = max(0, (PUE_max - PUE) / (PUE_max - PUE_min))

where:
  PUE_max = 1.60  (score = 0.00 at this value)
  PUE_min = 1.20  (score = 1.00 at this value)
  PUE     = your reported or measured PUE

This is a strictly linear interpolation. There is no step function — every tenth of a PUE point matters. PUE values below 1.20 score 1.00 (capped). PUE values above 1.60 score 0.00 (floored).

Worked example

Input: Facility reports PUE = 1.35

Calculation:

cool_efficiency_score = (1.60 - 1.35) / (1.60 - 1.20)
                      = 0.25 / 0.40
                      = 0.625

Result: COOL score 0.625.

Comparison to the scoring table:

PUE	Formula result	Table value
1.20	(1.60-1.20)/(0.40) = 1.000	1.00
1.25	(1.60-1.25)/(0.40) = 0.875	0.84 (rounded)
1.35	(1.60-1.35)/(0.40) = 0.625	0.60 (rounded)
1.40	(1.60-1.40)/(0.40) = 0.500	0.46 (rounded)
1.60	(1.60-1.60)/(0.40) = 0.000	0.00

Confidence levels

COOL scores carry confidence labels based on how PUE was measured:

high — annual average from metered utility data
medium — spot measurements or facility-reported estimates
low — single-point measurement or vendor specification

A COOL score of 0.60 with high confidence means something different than 0.60 with low confidence. Confidence is disclosed in the certification report.

What COOL does not score

COOL does not assess cooling system type directly — whether you use air, chilled water, direct liquid cooling, or free cooling is not the variable. PUE is. A facility with air cooling operating at PUE 1.28 scores higher than one with liquid cooling operating at PUE 1.45. The measurement, not the technology, determines the score.

COOL Level 2 (planned)

COOL v0.1.0 operates on single-point or annual-average PUE. A planned Level 2 extension will ingest BMS (Building Management System) timeseries data to produce seasonal efficiency curves and identify operational windows where PUE degrades. This extension is not yet available.

Input schema

Field	Type	Required	Description
pue	float	yes	Power Usage Effectiveness (≥ 1.00)
measurement_type	string	yes	annual_average, spot, or vendor_spec
cooling_type	string	no	air, chilled_water, dlc, free_cooling, immersion
climate_zone	string	no	Used for peer benchmarking in future versions
reporting_period	string	no	ISO8601 date range for the PUE measurement

CLI usage

# Score from a single PUE value
cool analyze --pue 1.35 --measurement-type annual_average

# Score with full facility context
cool analyze --pue 1.35 --cooling-type dlc --climate-zone mild

# Output to JSON
cool analyze --pue 1.35 --output cool_result.json