Singapore hosts over 70 data centres consuming approximately 7% of Singapore's total electricity. A figure that prompted the government's data centre moratorium (2019–2022) and the subsequent Green Data Centre Roadmap requiring new data centres to achieve PUE ≤ 1.3 and WUE ≤ 0.4. IMDA's green data centre standards, the Singapore Green Building Council's Green Mark for Data Centres, and Uptime Institute Tier certifications all require documented environmental and energy performance monitoring. This guide covers the measurement instruments enabling Singapore data centres to meet these requirements.
ASHRAE Thermal Guidelines for Singapore Data Centres
ASHRAE's Thermal Guidelines for Data Processing Environments define recommended and allowable temperature and humidity ranges for server inlet air:
| Class | Temperature (°C) | RH (%) | Typical Singapore Application |
|---|---|---|---|
| A1 (Recommended) | 18–27°C | 5.5°C dew point to 60% RH | Enterprise servers, financial trading |
| A2 (Recommended) | 10–35°C | 8% to 80% non-condensing | Most commercial colocation |
| A3 (Allowable) | 5–40°C | 8% to 85% non-condensing | High-density computing, edge |
Singapore's ambient humidity (70–90% RH) makes the risk of condensation and corrosion from humidity-related failures very real. Particularly in equipment near outdoor air intakes or in poorly sealed hot aisle/cold aisle containment systems. Continuous humidity monitoring with Rotronic's RMS system detects containment failures before they cause corrosion damage.
Power Usage Effectiveness (PUE) Measurement
PUE = Total Facility Power / IT Equipment Power. Measuring PUE accurately requires:
- Total facility power: Measured at the utility service entry using revenue-grade power meters or the Fluke 435-II for temporary measurement
- IT equipment power: Measured at the PDU (Power Distribution Unit) level using Fluke 1730 energy loggers or smart PDU monitoring
- Data averaging: PUE must be averaged over at least 12 months for IMDA reporting (avoiding cherry-picking cooler months)
Singapore's IMDA Green Data Centre Roadmap requires new data centres to target PUE ≤ 1.3. Existing data centres should work toward PUE ≤ 1.5. The Fluke 435-II power quality analyser provides the measurement accuracy needed for verifiable PUE calculations submitted to IMDA.
Recommended Instruments
Temperature and Humidity Monitoring
Rotronic RMS with HL-20 Transmitters: Distributed temperature and humidity sensors at server inlet and exhaust, hotspot monitoring in high-density racks, and cooling system performance tracking. The RMS platform provides real-time dashboards, alarm management, and automated reports for data centre operations teams and IMDA compliance reporting.
Power Quality at Data Centre Incoming Supply
Fluke 435-II: Data centres are particularly sensitive to power quality. Voltage sags can trigger UPS transfer and server reboots; harmonics from UPS systems load distribution transformers. The 435-II identifies power quality issues at the service entry and within the data centre distribution infrastructure, enabling corrective action before they affect IT availability.
UPS and Battery Monitoring
Fluke 500 Series Battery Analyser: Data centre UPS batteries must be tested quarterly to verify available capacity. The Fluke 500 series performs conductance testing on VRLA and lead-acid batteries, identifying weak cells before failure. Battery failures during a mains outage event are a leading cause of data centre downtime. Predictive battery testing is a critical resilience measure.
Airflow Measurement for Cooling Efficiency
Data centre cooling efficiency depends on effective hot aisle/cold aisle containment. Airflow measurement at CRAC/CRAH discharge and at server inlet faces identifies recirculation bypass. The main efficiency loss in unconsolidated data hall layouts. Fluke's 922 Airflow Meter measures air velocity at supply diffusers and server faces, enabling hot spot identification and CRAC unit repositioning optimisation.
Worked Example: Finding a Hidden Hotspot With a Combined Approach
A Singapore colocation operator receives a customer complaint about a specific rack running warmer than its neighbours despite the data hall's overall temperature and humidity trends looking normal on the RMS dashboard. Because the fixed HL-20 sensors are positioned at representative points across the hall rather than at every individual rack, the aggregate data alone does not localise the problem. A facilities engineer follows up with a handheld thermal camera scan along the affected row and an airflow meter check at the nearest CRAC discharge and the rack's own cold aisle floor tile. The combination reveals the actual cause quickly: a floor tile with an oversized perforation two racks upstream is short-circuiting a disproportionate share of the cold air supply before it reaches the complained-about rack, starving it of the cooling the aggregate hall data suggested was available. Swapping the floor tile for a correctly rated grille and re-scanning confirms the fix. This is a useful illustration of why fixed environmental monitoring and portable diagnostic instruments (thermal camera, airflow meter) are complementary, not redundant: the fixed system tells you the hall is healthy on average, and the portable instruments find the localised exception the averages hide.
Battery Testing as a Predictive, Not Reactive, Practice
UPS battery failure during a genuine mains outage is one of the more preventable causes of Singapore data centre downtime, precisely because conductance testing is a mature, well-understood predictive technique that most facilities already have the instruments to perform, yet quarterly testing discipline is where many maintenance programmes actually slip. A conductance reading trending downward over consecutive quarters on a specific battery string, even while every cell still technically passes a simple voltage check, is the early warning sign a purely voltage-based inspection misses; conductance responds to internal plate degradation well before voltage under load does. Building battery conductance data into the same facility dashboard used for temperature, humidity and power monitoring, rather than treating it as a separate, occasionally-remembered maintenance task, is what turns battery testing from a compliance checkbox into a genuine predictive maintenance input.