CASE39 · Data centers · Cooling / PUE
Real-time PUE diagnosis and cooling optimization across data centers
A multi-center performance-improvement proposal integrates facility, IT and cooling electricity with operating data. It connects PUE to cooling-system causes and recurring optimization actions.
01 · CASE ANALYSIS
Baseline & operating challenge
The reported 2016 totals for three centers were 176,885 MWh/year of facility electricity and 97,429 MWh/year of IT electricity, with an average PUE of 1.82. The target table instead uses a 2017 estimated reference PUE of 1.77; these bases must remain distinct.
The three centers had 594 installed cooling units and a reported load-based requirement of 132 active units at average demand. The difference indicates staging headroom, not a recommendation to remove 462 units or ignore redundancy.
02 · CASE ANALYSIS
Data quality & performance diagnosis
Stored data intervals differed between centers: one minute, 15 minutes and one hour. Some cooling-unit, pump-power and outlet-temperature measurements were missing. Additional metering, calibration and time alignment were proposed before optimization.
PUE = facility electricity ÷ IT electricity. Cooling electricity relative to IT electricity and delivered cooling relative to cooling-unit electricity provide more specific diagnostic views. The public source uses “real-time diagnosis” without fixing one interval because report pages differ.
| Reference | Value |
|---|---|
| 2016 facility electricity, three centers | 176,885 MWh/year |
| 2016 IT electricity | 97,429 MWh/year |
| 2016 electricity cost | KRW 20,486 million/year |
| 2016 average PUE | 1.82 |
| 2017 estimated target-table reference PUE | 1.77 |
03 · CASE ANALYSIS
Proposed optimization actions
Compare room temperatures and setpoints to identify overcooling; match cooling-unit and chiller staging to IT and cooling loads; optimize free-cooling transitions using outdoor conditions and floor loads.
Check chiller condition, COP, temperature settings and load sharing. Optimize cooling-tower and pump flow, temperatures and active units, and improve airflow distribution around hot spots.
The proposed IPOS platform combines integrated monitoring, efficiency diagnosis, trends and correlations with repeated analysis, corrective action and result review. Operator training supports sustained improvement.
04 · CASE ANALYSIS
Performance targets
The first-stage proposal targets PUE 1.70 over 24 months, with 6.5 GWh/year estimated savings and approximately KRW 753 million/year in cost savings. The longer-term proposal targets PUE 1.65.
A separate electricity-contract review identified approximately KRW 167 million/year of potential cost reduction at one center. This is tariff optimization, distinct from cooling-efficiency savings, and is not included as an achieved result here.
| Stage | PUE | Estimated electricity saving | Estimated cost saving | Reported cooling-energy reduction |
|---|---|---|---|---|
| First stage / 24 months | 1.70 | 6.5 GWh/year | Approximately KRW 753 million/year | 8.81% |
| Longer term | 1.65 | 11.3 GWh/year | Approximately KRW 1,310 million/year | 15.3% |
05 · CASE ANALYSIS
Measurement & verification
Validate facility and IT electricity, outdoor conditions and cooling operation under aligned measurement intervals. Evaluate center/floor PUE together with cooling and equipment efficiency.
PUE and all savings figures in the target table are pre-implementation targets or estimates. They must be verified after implementation with IT-load and weather effects considered; a simple electricity before/after comparison is insufficient.
APPLY TO YOUR SITE
Connect your operating data to an actionable energy strategy.
Discuss the measurement boundary, data quality, operating constraints and verification approach for your facility.
Discuss your project All English cases