UPS sizing: calculate kW, kVA and a practical load schedule

UPS sizing starts with the load’s real power in kW and apparent power in kVA. Check both against the proposed UPS ratings, then assess growth, load changes, redundancy, runtime, and site limits as separate decisions.

A load can fit below a UPS’s kW limit and still exceed its kVA limit. Another proposal may carry the load but lack the battery time you need. One headline capacity cannot answer both questions.

Comparing quotes? Start with a clear load sheet. This UPS sizing guide shows how to test both power limits and record each reserve. You’ll have inputs a supplier can review, rather than a model chosen from one rounded number.

Key takeaways

• Use the actual load power factor when converting kW to kVA.
• Check the UPS’s kW and kVA ratings separately.
• Keep growth, target loading, and redundancy visible so reserve is not counted twice.
• Request battery runtime at the stated load and configuration.
• Treat the calculation as a candidate check, followed by model and site review.

UPS sizing starts with a defined load schedule

List the circuits that must stay powered. A factory control supply, network rack, and motor circuit are different loads. Decide which belong in the UPS scope before adding their ratings.

For UPS sizing, record the operating state behind each measurement. Idle demand may differ from peak work, startup, or a planned shutdown. Note which loads run together and which events can overlap.

A power-supply label can provide useful information, but it is not proof of the equipment’s normal draw. If you use a nameplate limit because measurements are unavailable, mark it as such. Do not present that estimate as a measured load.

Build a sheet you can review

Copy the table below and add one row per device or circuit. Use kW and kVA consistently; divide watts or VA by 1,000 when entering them in those columns.

The power columns contain row totals, so don’t multiply them by quantity again when adding the sheet. Keep source notes beside the figures. If a value is unknown, leave it unknown and identify who will measure or confirm it.

UPS sizing benefits from a measured aggregate kW and kVA at the relevant operating state. Adding individual VA values can serve as a conservative planning estimate, but it isn’t an exact calculation for every mixed-load network. Do not add different power factors or average them without a valid load basis.

Start with the load sheet. Use the industrial UPS selection guide for the wider site and topology decisions, then bring these power inputs into the configuration discussion.

For UPS sizing, use one clear measurement boundary. A reading at the UPS output represents the load side. A reading upstream can include UPS losses and charging. Label the location, operating mode, and meter record so those readings are not mixed.

Keep the evidence with the number

For each important value, retain the meter record, device document, or stated assumption. Note when it was taken and what equipment was active. A total without that context is hard to update when the site changes.

This is a working UPS sizing worksheet, not an automatic product selector. It exposes the inputs and gaps. The proposed UPS still needs checks against its own limits and the site’s operating conditions.

UPS sizing formula: check both kW and kVA

For a load with a known true power factor, the relationships are:

kW = kVA × load PF

kVA = kW ÷ load PF

Power factor is a ratio, not an efficiency percentage to subtract from the load. Use the value for the load and operating state being assessed. If you don’t know it, don’t silently insert 0.8, 0.9, or 1.0.

Separate three power-factor figures

UPS sizing can involve three different numbers. Keeping them separate avoids a common conversion error.

For example, Banatton’s public three-phase UPS rating table lists BT9100H-33 at 100kVA/90kW. The ratio is 0.9. That does not establish the power factor of whatever equipment you connect.

Other products have different rated ratios. Fuji Electric’s UPS6000DX-T3 specifications include a 100kVA/100kW unit. Check the actual model rather than assuming all 100kVA products have the same kW limit. Neither product reference establishes suitability for your load.

Illustrative scenario: one limit passes and another fails

Assume an 80kW load and a candidate rated 100kVA/90kW. At a load PF of 0.8, apparent power is 80/0.8 = 100kVA. At a PF of 0.7, it is about 114.29kVA.

The second case exceeds the kVA rating even though 80kW is below 90kW. The first case reaches the apparent-power limit before any allowance for growth or other conditions. These are assumed loads, not customer results or an approved equipment choice.

The practical UPS sizing test is simple: does the design load fit both ratings under the stated conditions? If either answer is no, that candidate does not pass the power check. Passing both is only the next step, not final approval.

Use a visible two-limit check

Make the UPS sizing decision repeatable. Keep two separate pass/fail fields in the comparison:

For the narrower purchase decision, see the 100kVA UPS configuration guide. Keep fixed-model details separate from this general calculation method.

Add growth and test the failure state

UPS sizing needs an explicit reserve plan. Growth, a target loading level, and spare units answer different questions. Do not hide all three inside an unexplained safety factor.

Eaton’s sizing guide separates capacity, runtime, growth, and redundancy. It recommends at least 15% growth allowance and uses a five-year planning example. That is attributed manufacturer guidance, not a rule that every project needs exactly the same reserve.

Schneider Electric’s sizing FAQ recommends loading around 80% in its general guidance. Treat that as a stated loading objective, not the same thing as adding 15% growth. The project needs to decide which assumptions apply.

Illustrative scenario: make each allowance visible

Assume a current load of 60kW with PF 0.9. Its apparent power is about 66.67kVA. Now assume a documented 20% growth plan and unchanged PF. Future load becomes 72kW and 80kVA.

If the project separately requires future loading to stay at or below 80%, divide those future loads by 0.8. The mathematical thresholds become 90kW and 100kVA. Check both when reviewing a candidate.

This UPS sizing example uses two distinct project assumptions. It is not a default recipe. If the future load sheet already includes the planned additions, do not add the same growth again.

The UPS sizing arithmetic changes with the reserve definition. Adding 20% to load uses a factor of 1.2. Limiting load to 80% of rating uses 1/0.8 = 1.25. Write which calculation you mean before comparing the result.

Check what remains after a unit is unavailable

In an N+1 arrangement, N units are required for the load and one additional unit provides redundancy. The installed total and the capacity remaining after losing a unit are different figures.

The UPS sizing review must check the defined failure or maintenance state. Confirm that the remaining equipment meets both power limits, along with the required runtime and operating conditions. Also review shared bypass, battery, control, and distribution elements.

Do not assume that a model supports parallel operation just because the nameplate arithmetic works. Request the approved configuration, accessories, limits, and behavior from the supplier. A redundant count does not remove every common point of failure.

Record why the reserve exists

Keep a separate line for growth, another for any chosen loading objective, and another for redundancy. State whether each applies to normal operation, a failure state, or both. This makes later changes to the UPS sizing basis easier to audit.

A larger proposal is not automatically better. Compare its operating load, efficiency curve, installation scope, and service needs against the same requirements. Choose reserve for a reason you can explain.

Check phase, startup and operating limits

For a single-phase load, apparent power can be calculated as:

kVA = V_rms × I_rms ÷ 1,000

For a balanced three-phase load:

kVA = √3 × V_LL × I_L ÷ 1,000

V_LL is line-to-line voltage and I_L is line current. State the voltage and phase assumptions beside the calculation. An unbalanced or complex mixed load needs more than a casual use of the balanced formula.

Keep supply-side and load-side checks separate

An output load-current calculation is not the UPS input rating. Input conditions can include conversion losses and battery charging as well as the protected load. Use the selected model’s documents for the supply review.

For UPS sizing, confirm the required input and output arrangement before using a formula. The single-phase vs three-phase UPS guide explains that distinction. The wider building supply does not determine every individual load’s requirement.

Test load events against curves

Record startup magnitude, duration, repeated starts, and large load steps. Request the candidate’s overload curve for the relevant mode and conditions. A continuous kVA rating alone doesn’t show how it will handle those events.

The UPS sizing sheet should also identify the expected temperature, altitude, ventilation, and any manufacturer derating. Mark what is measured and what still needs a site survey. Do not replace those limits with an industry label such as hospital or factory.

Ask the responsible designer to check fault response and protection coordination. This guide does not select cables, breakers, grounding arrangements, or an installation procedure. Those decisions require the actual system and applicable requirements.

Specify runtime without mixing power and energy

UPS sizing for power asks whether the unit can carry the load. Runtime asks how long a defined battery configuration can support it. Keep those questions separate in the worksheet and quotation.

State the required event: orderly shutdown, source changeover, or continued operation for a defined interval. Give the supplier the load profile during that event. A shutdown sequence may not draw the same power throughout.

Request the matched runtime curve

Ask for runtime at the stated load, using the proposed batteries and mode. Record battery type, configuration, temperature assumptions, discharge limit, and aging allowance. Check what the curve includes and what the supplier excludes.

Multiplying load power by time gives a load-side energy figure. It does not by itself specify a real battery bank. Simple Wh/W or Ah formulas cannot guarantee runtime without the relevant loss, discharge, and operating data.

For UPS sizing enquiries, do not treat a 1000VA label or a 500W load as enough information to promise minutes. The battery configuration and load conditions are still missing. Request evidence rather than filling that gap with a generic time range.

Turn the worksheet into a configuration enquiry

Put the design state, growth plan, power limits, runtime requirement, and site constraints on one requirements sheet. Ask each supplier to quote against that same basis. Note any exceptions next to the proposed model.

Illustrative scenario: two installed units, one required spare

Assume two compatible units, each rated 100kVA/90kW. Their summed nameplate ratings are 200kVA/180kW. If the requirement is to carry the load with either unit unavailable, the remaining unit provides only its own applicable ratings.

The protected load must therefore fit within 100kVA and 90kW in that assumed state, before considering other limits. This is a hypothetical UPS sizing check. It does not verify a manufacturer’s parallel approval, transfer behavior, or battery arrangement.

The next action is to request the approved system diagram and failure-state description. Confirm which components are shared and what happens during service. Arithmetic helps ask the right question; it does not certify the architecture.

Send a quote-ready package

Include these inputs with your UPS sizing worksheet:

1. Protected circuits, measured kW/kVA, operating states, and source records.
2. Startup events, load steps, and any unknown power-factor inputs.
3. Planned additions, reserve assumptions, and the required failure state.
4. Input/output supply, environment, access, and space constraints.
5. Runtime, shutdown or transfer sequence, and battery assumptions.
6. Monitoring, maintenance, commissioning, and service requirements.

Request the exact model/version, both power ratings, overload data, derating limits, approved parallel options, and the runtime curve. Ask suppliers to identify missing inputs and deviations. Do not treat a catalogue listing as a signed design.

Compare the same scope. Send the worksheet through Banatton’s contact page and request a configuration enquiry response with supporting documents. Separate the UPS, batteries, bypass, monitoring, installation responsibilities, and service terms in the quotation.

Keep the decision traceable

Retain the load sheet version used for the quote. If equipment is added or the required outage interval changes, update the relevant inputs and ask whether the proposed configuration still fits. Avoid changing one total while leaving the reserve and battery assumptions untouched.

Round up to a real product only after checking both ratings. In UPS sizing, a larger kVA number alone does not prove that the required kW rating, overload duty, or site limits are met. Record the model and version beside the result.

A useful UPS sizing result is a candidate range plus a clear list of checks. It should help the buyer and supplier resolve the same questions. It should not hide uncertainty behind a precise-looking model recommendation.

Frequently asked questions and final checks

How do I calculate the UPS size I need?

Define the protected load and establish its kW and kVA for the relevant operating state. Check both against candidate ratings, then review growth, load events, failure state, runtime, and site limits. Keep missing data visible.

Should I use load PF or the UPS output rating ratio?

Use the actual load PF when converting that load’s kW to kVA. The UPS’s rated kW/kVA ratio describes its output limits; it does not tell you the connected load’s PF. Input PF belongs to a separate supply-side review.

Is 20% headroom always enough?

No universal percentage settles UPS sizing. Define growth, target loading, startup behavior, and redundancy separately. A load multiplied by 1.2 is not the same calculation as selecting capacity so that load occupies 80% of it.

How much capacity remains available under N+1?

Check the equipment left in service in the required state. Do not use the sum of installed nameplates as the remaining capacity. Verify the approved configuration, both power limits, shared elements, and battery duty.

How long will a 1000VA UPS or a 500W load run?

Those figures do not define a battery runtime. Request the curve for the specific unit, battery configuration, load, mode, and conditions. Power capacity and stored energy are separate inputs.

What does a supplier need to check my worksheet?

Provide the load evidence, phase/supply details, growth and redundancy assumptions, runtime event, site constraints, and service scope. Ask for model documents and an explanation of any unresolved requirement.

Finish UPS sizing with two checked power limits and a documented basis for reserve, runtime, and site review. Keep the calculations visible so another person can test the same assumptions.

Ready to compare a configuration? Send your UPS sizing worksheet and required operating states through Banatton’s contact page. Request a matched scope and supporting documents before choosing equipment.