How to Size UPS Battery Runtime for a Real Load

“We need a 6 kVA UPS to run for 20 minutes” is a starting point, not a battery specification. The same UPS might supply a 1.5 kW network rack or a 4 kW group of instruments. Those loads will draw very different amounts from its battery.

The kVA rating tells you the UPS output capacity. Runtime depends on the power that remains connected during the outage, the battery arrangement, and how the selected batteries perform at that discharge rate. A label marked 100 Ah does not settle the question either.

The sizing path:

• Measure the protected kW and note when the load changes.
• Convert that load to a battery discharge demand using the UPS's battery-mode data.
• Check the required minutes against the exact battery's discharge table and the UPS's approved configuration; verify the installed system under load.

The first calculation points to the battery data worth checking.

Start with the equipment that must keep running after utility power fails. Measure or obtain its operating kW, and note whether the load changes during the outage. A server room may shed nonessential equipment after several minutes. A production line may have a motor that starts while the supply is on battery. One steady-state number will miss both situations.

Also record kVA or power factor and any starting or peak current. Runtime is mainly an energy and discharge-rate problem; the UPS itself must still carry the load's continuous kW, kVA and short-duration peaks. If those ratings do not fit, a larger battery will not fix the UPS selection. Our industrial and medical UPS sizing guide covers that first check in more detail.

For a variable load, give the supplier a short load schedule: for example, the power required during the first five minutes, what can shut down, and what remains for the rest of the outage. A single runtime figure quoted at one constant load may not represent that sequence.

Choose the target time from the job the UPS must finish: keep equipment alive for an orderly shutdown, bridge a generator handover, or support continued operation until another supply is available. Those are different promises. A site's average outage length is useful background, but it cannot replace the required operating time for the protected load.

“20 minutes” should mean 20 minutes at the protected load, through the point at which power is restored or the equipment has shut down safely. If a generator is involved, count the whole handover: starting, reaching usable voltage and frequency, transferring the supply, and the UPS accepting that input. Engine start time alone is not the required battery time.

Decide whether the target applies only to a new, fully charged battery at a specified room temperature or must still be met later in the battery's service life. Those are different design requirements. If repeat outages are possible, check how much charge the system can recover between events rather than assuming every event begins with a full battery. Our UPS battery charging guide explains why charger capacity and the available recharge window matter.

The simplest first-pass calculation is:

Battery power (kW) ≈ protected AC load (kW) ÷ DC-to-AC efficiency in battery mode

Use the efficiency or battery discharge-current data for the actual UPS model and load level. Include auxiliary DC loads if the manufacturer has not already included them. For a final design, use the maker's runtime tool or discharge-current data: UPS efficiency and battery voltage can change as the battery discharges.

Here is a worked illustration using hypothetical assumptions:

• Measured protected load: 3.0 kW
• Required runtime: 20 minutes
• Assumed UPS battery-mode efficiency: 90%
• Assumed battery arrangement: one series string of sixteen 12 V VRLA blocks, six cells per block

The estimated battery-side power is 3.0 ÷ 0.90 = 3.33 kW. Divided across 16 blocks, that is about 208 W per block, or 34.7 W per cell across 96 cells. That is the discharge demand to check in the battery table.

Hypothetical comparison — not a TAFENG specification or runtime guarantee. The load makes a visible difference even with the same assumed UPS efficiency and battery string:

These figures define the battery demand. The discharge table then tells you whether a given battery can meet that demand for 20 minutes.

Now open the proposed battery model's constant-power discharge table. Find the 20-minute value at the temperature and end-of-discharge voltage allowed by the UPS. The battery must supply at least the required power per cell after the project's aging and temperature allowances are applied. If the maker provides watts per 12 V block instead, compare against 208 W per block. Do not mix a per-cell table with a per-block calculation.

The assumed 90% efficiency and sixteen-block string make the arithmetic visible; they must be replaced with the actual UPS data. In particular, never infer an approved battery count merely by dividing a nominal DC voltage by 12.

Multiplying nominal battery voltage by Ah gives a nominal energy figure. It does not tell you how much energy the bank can deliver in 20 minutes before the UPS reaches its cutoff voltage. An Ah rating is measured under a stated test rate and temperature; a short UPS discharge can be much more demanding.

Series and parallel wiring also matter. In a series string, block voltages add while the Ah rating stays that of one block. An approved second string in parallel increases available capacity, but the UPS and battery makers must permit the arrangement and specify its protection. Do not multiply the Ah by the number of blocks in a single series string.

Battery manufacturers therefore publish discharge data by time, end voltage and temperature. A 20-minute design should use a 20-minute rating or a manufacturer-approved sizing method, not a ten-hour or twenty-hour Ah rating stretched into a 20-minute promise. Check the exact battery model: two batteries with the same Ah marking can have different high-rate performance.

Before quoting, check the battery bank against the UPS's approved DC voltage, charger settings and permitted string count. A battery that looks right on paper can still be the wrong configuration for the UPS. Then check the protection devices, cable size and distance to the cabinet. The battery-cabinet distance guide explains why DC current and cable voltage drop matter, especially near the end of discharge.

For an online UPS that needs extended runtime, this external-battery online UPS model is one option to discuss with TAFENG. The battery arrangement has to be approved for the specific UPS model, load and site. Adding storage does not increase the UPS's output rating, and it does not make runtime unlimited.

Before signing off, test the installed system with the intended protected load and an agreed test procedure. Record the initial battery charge, room temperature, load profile and the point at which the UPS stops supplying it. Manufacturer curves are the basis for selection; a commissioning test checks the installation you actually built.

For an online UPS battery backup system, send the supplier the details that will decide the battery size:

• Measured operating kW and kVA, plus any load changes during the outage
• Peak or starting load and how long it lasts
• Required minutes and the event that ends the battery's job
• Proposed battery model or chemistry, if already chosen
• The actual room temperature at the battery cabinet, not just the comfort setpoint, and the UPS model

With those details, a supplier can begin a defensible calculation. Without them, “6 kVA, 100 Ah, 20 minutes” does not give a supplier enough to quote.