UPS Sizing: How to Match Data Center Capacity and Runtime to Your Load
Size a UPS from the real critical load and the required runtime, not from a room breaker rating or a rough guess. Start by defining the IT equipment that must stay online, add power conversion and distribution losses, choose a capacity that supports the required redundancy, then calculate the battery string separately for the required runtime.
Start With the Critical Load
The critical load is the equipment that must continue running during a utility failure. For a data center project this usually includes servers, storage, network switches, security systems, monitoring, and any facility controls needed for safe operation.
- List connected equipment nameplate ratings in kW or kVA.
- Use measured or projected operating load where available.
- Separate loads that can shed during an outage from loads that must stay online.
- Include UPS losses and the input power needed to recharge the battery system.
kW, kVA, and Power Factor
kW is real power and kVA is apparent power. The ratio between them is power factor. A UPS must be large enough for both the real power demand and the apparent power demand of the connected load. Loads with a low power factor increase the kVA requirement even when kW remains unchanged.
A practical specification should state both kW and kVA at the design load, the power factor used for sizing, the input voltage and phase configuration, and the expected load growth allowance.
Choose Capacity With Redundancy
The UPS capacity decision is separate from the redundancy decision. An N configuration uses one path with no spare capacity. N+1 adds one additional module or UPS so a unit can be taken out of service. A 2N configuration provides two independent power paths, each sized for the full load.
For modular UPS systems, consider the module size and how many modules are needed at the design load. Avoid sizing so tightly that a single module failure or normal maintenance event changes the runtime or overload behavior.
Calculate Battery Runtime Separately
Runtime is not a fixed property of the UPS model. It depends on battery capacity, actual load, end-of-discharge voltage, temperature, battery age, and the UPS discharge curve. The same UPS can support much longer runtime for a light load than for a fully loaded rack.
Specify the runtime target at the design load and decide whether the UPS must bridge to a generator or support an orderly shutdown. Fifteen to thirty minutes may be enough for a reliable generator handoff, while facilities without a generator should calculate the time required for the approved shutdown procedure.
Confirm With a Load Test
A UPS sizing calculation is only as reliable as its input data. After installation, a load test should verify capacity, voltage, current, runtime, alarm behavior, and generator handoff or orderly shutdown. Record the test conditions so future battery replacements can be measured against the same basis.
Часто задаваемые вопросы
How do I choose UPS capacity?
Use the critical load in kW and kVA, add planned growth, select the required redundancy topology, then verify that the UPS can support the battery string needed for the required runtime.
Do I size UPS by IT load only?
No. Include power distribution losses, UPS input losses, monitoring and control loads, and any loads that must stay online during an outage.
What happens if I oversize a UPS?
Oversizing may reduce measured efficiency at a light load and increase capital cost. It can still be appropriate when growth or future loads are planned, so document the design load and expansion basis.
Need help converting your load profile into a UPS capacity and runtime recommendation? Send the load list and operating conditions to VERHI for an engineering review.
This guide is based on engineering selection principles and VERHI's published scope. Site-specific values should be confirmed with the VERHI engineering team.
