DataCentersX > Facility Ops > Power Monitoring

Data Center Power Monitoring

Power monitoring is the sensor-and-data-collection discipline that produces electrical telemetry across the facility - from utility entrance through distribution to rack-level branch circuits. The discipline feeds multiple consuming systems: EPMS for electrical infrastructure operations, DCIM for capacity and asset management, EMS for energy portfolio orchestration, and the PUE/efficiency reporting that flows to GRC:Sustainability. Power monitoring is the source layer; the consuming platforms are the integrating layer.

Monitoring tiers

Power monitoring runs at multiple tiers, each with different sensor density, sample rate, and consuming systems.

Tier	What it measures	Primary consumer
Utility entrance	Total facility import; revenue metering; power factor; demand	EMS, EPMS, billing reconciliation
Substation and main switchgear	Voltage, current, power, energy, harmonics at MV/HV level	EPMS, EMS
Distribution panels and PDUs	Per-feeder voltage, current, power; circuit-level loading	EPMS, DCIM
Branch circuit (rPDU)	Per-rack and per-outlet power consumption	DCIM, capacity planning, customer billing
UPS	Battery health, runtime, load, alarms, transfer events	EPMS, predictive maintenance
Generator and onsite generation	Output power, fuel consumption, runtime, exhaust emissions, vibration	EMS, predictive maintenance, emissions reporting

Power quality measurement

Power quality monitoring extends beyond simple voltage and current to capture phenomena that affect equipment operation and longevity. The discipline is governed by IEC 61000-4-30 (the international standard for power quality measurement methods) and IEEE 1159 (voltage quality recommendations).

Phenomenon	What it is	Why it matters
Harmonics	Distortion of the sinusoidal waveform from non-linear loads (UPS, VFDs, switching power supplies)	Heating in transformers and conductors; neutral overloading; equipment derating
Voltage sags and swells	Short-duration voltage deviations from nominal	Equipment trips; UPS battery cycling; ITIC/CBEMA compliance
Transients	Brief high-magnitude voltage events (lightning, switching surges)	Equipment damage; surge protector aging; insulation degradation
Frequency variation	Deviation from nominal 50/60 Hz	Generator synchronization; clock-sensitive equipment; grid stability indicator
Flicker	Rapid voltage variations causing visible lighting fluctuation	Indicator of upstream load issues; rare in modern data centers
Imbalance	Unequal voltage or current across phases	Motor heating; transformer derating; root-cause indicator for asymmetric load
Power factor	Ratio of real to apparent power	Utility billing penalty; capacity utilization; reactive power management

Sensors and meters

Device class	Vendor examples	Where deployed
Revenue-grade meters	Schneider PowerLogic ION, Eaton IQ, GE PQM, Siemens SENTRON	Utility entrance, sub-billing, customer cross-charge
Power quality meters	Schneider PM8000/ION9000, Dranetz, Fluke 1760, Hioki PQ3198	Main switchgear, critical distribution panels, site characterization
Branch circuit monitoring	Veris (Schneider), CCS, Packet Power, Setra	Distribution panels feeding racks; per-circuit metering
Intelligent rPDUs	Vertiv Geist, Server Technology, Raritan, Eaton ePDU, APC by Schneider	Rack-level outlet monitoring
Current transformers	Continental Control Systems, Magnelab, Veris	Retrofit branch circuit monitoring; non-invasive installation
UPS monitoring (built-in)	Vertiv, Eaton, Schneider, ABB, Mitsubishi Electric, Riello	UPS internal telemetry exposed via Modbus, BACnet, SNMP, or proprietary protocols

Protocols and integration

Power monitoring data flows from devices to consuming platforms across multiple protocols. Modbus TCP and Modbus RTU dominate at the device level for cost and ubiquity. BACnet/IP is common where building management integration is required. SNMP is standard for network-attached equipment (rPDUs, intelligent UPS). IEC 61850 is the standard for utility-grade substation automation and is increasingly deployed at large data center MV substations. OPC UA is the emerging standard for cross-platform integration. Most large facilities run protocol gateways that translate between device protocols and the EPMS/DCIM platforms - the protocol heterogeneity is a real operational concern, not just a procurement footnote.

Capacity and capacity planning

Per-circuit and per-rack power monitoring is the primary input to capacity planning. The data center industry's traditional measurement of "stranded capacity" - the difference between facility-rated capacity and actual occupied capacity - depends on accurate branch-circuit monitoring to identify the racks and feeders where additional load can be safely added without exceeding circuit ratings. AI training workloads have made this more difficult because GB200 racks at 130 kW and Rubin reference designs at higher densities consume the entire rated capacity of single circuits, leaving no headroom for traditional opportunistic placement. Capacity planning has shifted from "where can we add another rack" to "do we have a circuit that can support this rack class at all" - a fundamentally different planning question that depends on the same monitoring infrastructure.

PUE telemetry

Power Usage Effectiveness (PUE) is the ratio of total facility power consumption to IT load consumption. Calculating PUE accurately requires comprehensive monitoring at both the numerator (utility entrance) and the denominator (IT load aggregated across all rack circuits, or measured at UPS output). Many published PUE figures are calculated against engineering-design assumptions rather than measured data; rigorous reporting under the Green Grid framework, ISO/IEC 30134, and the EU CSRD requires actual measured PUE with documented metering methodology. WUE (Water Usage Effectiveness) and CUE (Carbon Usage Effectiveness) follow the same measurement principle. The reporting frameworks live in GRC:Sustainability; the measurement infrastructure lives here.

UPS battery monitoring

UPS battery monitoring is its own subdiscipline because battery health is the primary determinant of UPS reliability and battery failure is the most common UPS failure mode. Per-cell or per-string voltage, internal resistance (impedance) testing, temperature, and discharge testing are the standard measurements. VRLA (valve-regulated lead-acid) batteries are reaching end-of-life at many older facilities; lithium-ion replacements are increasingly deployed as the technology matures. Battery monitoring vendors include BTECH, Vertiv Albér, EnerSys, and Eagle Eye Power Solutions. The data flows to EPMS and predictive maintenance platforms; battery analytics is one of the highest-value targeted applications of facility AIOps because battery failure consequences are severe and predictive accuracy is achievable.

Where this fits

Power monitoring is the source-layer discipline; EPMS is the integrating platform that consumes it for electrical operations; DCIM consumes it for asset and capacity management; EMS consumes it for energy portfolio orchestration. The boundary between Power Monitoring and EMS is the boundary between sensor-and-data-collection (here) and energy-portfolio-decision-making (there). UPS, generator, and battery monitoring touch power distribution monitoring and feed predictive maintenance via AIOps. Compliance evidence flows to GRC:Compliance and PUE/efficiency reporting flows to GRC:Sustainability.

Related coverage