UPS runtime calculator
A battery rated for twenty hours does not deliver a twentieth of that in one hour. This accounts for it.
Enter the load, the battery bank and how deeply you are willing to discharge it. The calculator returns autonomy in minutes, the usable energy behind that figure, and the capacity a chosen target runtime would actually require.
The derate is the part most quick calculations skip. Battery capacity is published at a slow discharge rate, and a UPS empties its bank in minutes rather than hours, so the energy available at that rate is materially less than the nameplate arithmetic suggests.
Worked example — 100 kW on two 480 V strings
Two 100 Ah strings, discharged to 80%, with 70% of the published capacity available at that rate.
- Load on the UPS — 100 kW. Real power drawn from the UPS output, not the UPS nameplate rating.
- DC bus voltage — 480 V. Nominal battery string voltage seen by the inverter.
- Capacity per string — 100 Ah. Published amp-hour rating, normally quoted at a 10 or 20 hour discharge rate.
- Parallel strings — 2 strings. Battery strings connected in parallel on the same bus.
- Depth of discharge — 80 %. How much of the rated capacity you are prepared to use before the inverter shuts down.
- High-rate availability — 70 %. Share of the published capacity actually delivered at a short discharge rate. Around 60-75% is typical for a 10-15 minute discharge.
Estimated autonomy: 32.3 minutes — 53.8 kWh usable at 100 kW
| Nameplate bank energy 480 V × 100 Ah × 2 strings | 96.0 kWh |
|---|---|
| Usable energy after derates 80% depth of discharge, 70% high-rate availability | 53.8 kWh |
| Average discharge current At nominal bus voltage; real current rises as the bank sags | 208 A |
| Energy needed for 5 minutes | 8.3 kWh |
| Energy needed for 10 minutes | 16.7 kWh |
| Energy needed for 15 minutes | 25.0 kWh |
| Autonomy at half load Roughly — the rate factor improves at lower discharge rates, so this is conservative | 64.5 minutes |
Formula
Runtime (h) = (V × Ah × strings ÷ 1000) × DoD × high-rate factor ÷ load kW
The two derates are what separate this from the nameplate arithmetic: depth of discharge is a policy choice about battery life, and the high-rate factor accounts for the fact that a fast discharge yields less energy than a slow one. Neither is a manufacturer discharge table, which is what a real design uses.
What autonomy is normally bought for
Autonomy is sized around what has to happen during the discharge, not around a round number of minutes.
| Runtime | Bought to cover | Depends on |
|---|---|---|
| 30 seconds – 1 min | Flywheel or supercapacitor ride-through | A generator that starts and accepts load reliably |
| 5 minutes | Transfer to standby generation | Start sequence, transfer timing and fuel that is actually there |
| 10 – 15 minutes | Generator start plus a margin for a failed first attempt | A second start attempt or a second generator |
| 30+ minutes | Orderly shutdown with no generation on site | Shutdown automation that has been tested, not assumed |
Why the nameplate arithmetic overstates runtime
Amp-hour ratings are published at a slow discharge — usually ten or twenty hours. Pull the same bank down in ten minutes and the chemistry cannot keep up: internal resistance, diffusion limits and voltage sag all reduce the energy you can actually extract. Lead-acid banks commonly deliver around 60 to 75 per cent of their published capacity at typical UPS discharge rates, and the effect is steeper the faster you discharge.
On top of that, batteries age. A string at the end of its service life is generally defined as one that has fallen to 80 per cent of rated capacity, which means a design that was exactly adequate when commissioned is inadequate before the batteries are formally due for replacement.
Temperature quietly sets the rules
Battery rooms are held near 20 to 25 °C because capacity falls as temperature drops and life falls as temperature rises. A common rule of thumb for lead-acid is that every 8 to 10 °C above the design temperature roughly halves service life, while cold reduces available capacity on the day you need it.
That makes battery room cooling part of the power system, not a comfort load. A cooling failure that leaves the room warm does not produce an alarm anybody links to autonomy until the next discharge proves it.
What this does not replace
- Manufacturer discharge tables, which give energy at each specific discharge rate and end voltage rather than a single derate.
- A battery monitoring system reading real impedance per block, which is the only thing that finds a weak cell before it is tested by an outage.
- A load bank test, which is the only evidence that the bank delivers what the calculation claims.
- Lithium-ion behaviour, which is significantly flatter under high-rate discharge and does not follow the same rules of thumb.
Frequently asked questions
How do you calculate UPS battery runtime?
Work out the usable energy in the bank — bus voltage times amp-hours times the number of strings, divided by 1,000 — then reduce it for depth of discharge and for the high discharge rate, and divide by the load in kW. The result is hours; multiply by 60 for minutes.
How long should a data center UPS run on battery?
Most facilities with standby generation buy 5 to 15 minutes: enough for the generator to start, stabilise and accept load, with margin for a failed first attempt. Sites without generation size for an orderly shutdown instead, which is usually 30 minutes or more.
Why is my actual runtime shorter than the calculation?
Almost always because the published amp-hour rating was measured at a much slower discharge than a UPS performs, and because the bank has aged. Battery temperature above the design point, an unbalanced string and a single weak block all reduce it further.
What depth of discharge should I use?
Around 80 per cent is common for lead-acid in this application. Going deeper extracts energy from the part of the curve where voltage is already collapsing and shortens service life significantly, so the extra minute is expensive.
Does doubling the battery double the runtime?
Slightly more than double, in practice. Twice the bank carries the same load at half the discharge rate, and a lower rate means a larger share of the published capacity is actually available — so the high-rate penalty eases as you add capacity.
Do lithium-ion batteries change the arithmetic?
Yes. Lithium chemistries hold voltage far better under high-rate discharge, so the high-rate derate is much smaller, usable depth of discharge is typically greater, and capacity fade with age follows a different curve. The rules of thumb here are written for lead-acid.
Where this calculation stops
This is a sizing estimate, not a battery design. A real design uses the manufacturer discharge curves at the specific end voltage, accounts for ageing and temperature, and is proven by a load bank test rather than by arithmetic.
The model assumes a healthy, balanced bank. A single failing block can end a discharge early regardless of what the rest of the string is capable of.