What this UPS runtime calculator estimates
UPS sizing has two different dimensions that are often mixed together.
The first is power capacity: how many kW or kVA the UPS can support at one moment. The second is energy capacity: how long the battery system can keep supplying that load.
Adding battery cabinets can increase runtime without increasing the UPS inverter's power rating. Conversely, installing a larger UPS power block does not automatically create more battery autonomy.
The calculator therefore starts with the critical load and the installed battery energy as separate inputs.
The basic energy equation is simple — real battery behavior is not
At the most basic level, energy is power multiplied by time.
Turn the equation around and an ideal battery with 100 kWh of usable output supporting a constant 500 kW load would last:
A real UPS battery plant is more complicated. Nominal battery energy is not necessarily the amount of AC energy eventually delivered to the IT load.
Battery age, state of charge, allowable discharge depth, discharge rate, temperature, DC voltage behavior and UPS conversion losses all affect the result.
This calculator makes those derating assumptions visible
Rather than hiding a generic runtime factor inside the formula, the tool exposes three battery-side adjustments.
How much of nominal stored energy the planning model allows to be discharged.
Reduces modeled capacity for deterioration or an end-of-life design condition.
Allows a project-specific correction for non-ideal discharge behavior.
These percentages are project inputs, not universal battery constants. The defaults in this calculator are illustrative and should be replaced with manufacturer and project data before using the result for engineering decisions.
The complete concept-stage formula
First calculate the total nominal battery energy:
Then apply the battery-side factors:
The model then applies UPS battery-mode efficiency:
Finally:
A 500 kW worked example
Take the calculator's starting assumptions:
The two battery banks contain 160 kWh of nominal energy.
After the three battery-side factors:
After 95% battery-mode UPS efficiency:
At 500 kW:
That is a planning result, not a prediction that a real battery model will deliver exactly 11.2 minutes.
The inverse calculation answers a second useful question
Many projects know the required runtime before they know the battery size.
The calculator therefore also solves for the nominal battery energy required to meet the target runtime:
With the default assumptions, supporting 500 kW for 10 minutes requires approximately 143.3 kWh of nominal battery energy.
The modeled 160 kWh installation therefore has approximately 16.7 kWh of nominal-energy margin against that target.
Do not confuse battery kWh with UPS kW
A battery plant can contain enough energy for the required runtime and still sit behind an undersized UPS.
For example, a battery system might theoretically contain enough energy to support 1 MW for ten minutes. That does not help if the UPS inverter, static bypass or distribution path can support only 750 kW.
The Data Center UPS Cost guide treats protected power and battery runtime as separate design dimensions for exactly this reason.
Adding more batteries usually increases runtime, not UPS output power
Eaton's UPS sizing guidance separates capacity and runtime in the same way. UPS capacity determines how much connected load can be supported, while larger battery capacity or additional external battery packs increase the time that load can remain supported.
This distinction matters when comparing proposals.
Two systems can both be described as “1 MW UPS systems” while one has five minther has fifteen. Their inverter power may be similar while the battery plants are materially different.
Load has a direct effect on runtime — but real curves are not perfectly linear
In a pure energy-balance model, halving the load doubles runtime.
Actual battery behavior can depart from that relationship, particularly at high discharge rates. Voltage sag, chemistry, temperature and battery design affect how much useful energy remains available before the UPS reaches its DC cutoff condition.
Schneider's runtime guidance illustrates why manufacturer curves matter: runtime changes substantially with load, and the company recommends checking the runtime graph for the specific UPS.
The calculator's discharge factor exists to make this limitation explicit, but one percentage cannot reproduce an entire discharge curve.
Manufacturer runtime curves should supersede this calculator
Once a specific UPS and battery configuration has been selected, vendor runtime data is more useful than a generic energy calculation.
Eaton publishes runtime graphs for UPS models with different external battery configurations. Schneider likewise publishes load-versus- runtime information for specific products.
Those curves incorporate characteristics that a simple kWh model cannot reconstruct accurately.
Battery condition changes the result over the life of the system
A new battery plant and an aging battery plant should not be assumed to provide identical autonomy.
Schneider notes that runtime estimation is influenced by battery age, connected load and battery configuration. Its calibration guidance also explains that the runtime displayed by a UPS remains an estimate and changes with actual battery condition.
This is why runtime planning often needs an end-of-life or state-of- health assumption rather than using beginning-of-life nominal capacity indefinitely.
Usable capacity is a design decision, not simply the battery label
A battery may be sold with a nominal energy value, but a reliability model does not necessarily assume every nominal watt-hour is available to the critical load.
The battery management system, UPS cutoff voltage, chemistry, warranty conditions and project operating strategy can all constrain the usable portion.
That is why the calculator exposes usable capacity instead of silently treating nominal energy as fully deliverable energy.
Battery-mode UPS efficiency also consumes part of the energy budget
Energy stored on the DC side has to pass through power electronics before reaching the protected AC load.
If 100 kWh of usable DC energy passes through a 95% efficient path, the simplified delivered AC energy is 95 kWh.
For short-runtime systems, even a small percentage difference can move the expected autonomy by seconds or minutes depending on the scale of the battery plant.
Final calculations should use the battery-mode performance of the selected UPS rather than its best advertised efficiency in another operating mode.
Nominal kWh can be estimated from battery voltage and amp-hours
Battery documentation is not always presented directly in kWh. A simplified nominal-energy conversion is:
For example, a nominal 480 V battery string rated at 200 Ah contains:
That still does not mean 96 kWh will be delivered as usable AC energy. The same discharge, age, cutoff and conversion factors still apply.
Parallel strings need to be modeled in the correct failure state
A battery plant may contain multiple parallel strings or cabinets. Simply multiplying all installed strings together can overstate available runtime if the resilience design requires the system to meet its target with one string unavailable.
That is why the calculator asks for available battery banks rather than total physical banks.
If four banks are installed but the design requirement must be met with one bank isolated for maintenance, use three in that scenario.
Then repeat the calculation for normal operation if the normal-state runtime is also useful.
Battery runtime and generator start time solve different problems
In a conventional data center architecture, the UPS battery often bridges the period between utility loss and stable standby generation.
Battery autonomy therefore needs to cover more than the engine's nominal start sequence. The complete transfer path can include failure detection, generator start, stabilization, paralleling or transfer, and enough operational margin for abnormal starts.
Our backup generator cost analysis covers the longer-duration standby-power layer separately.
The correct battery runtime is an architecture decision: it depends on what event the batteries must bridge and what happens if the next source does not become available on the first attempt.
Five minutes versus fifteen minutes can materially change the battery plant
Hold every derating assumption constant. Required nominal battery energy rises approximately in proportion to target runtime inside this planning model.
These values show ideal load energy only. Actual nominal battery requirements will be larger after project-specific capacity, condition, discharge and conversion factors are applied.
AI loads make the protected-load assumption more important
A large AI rack does not necessarily draw a perfectly flat electrical load. Our GPU rack power analysis describes modern rack-scale systems where both design power and workload behavior matter to the upstream electrical system.
A runtime model therefore needs to say whether the protected kW is a design maximum, expected sustained battery-mode load or another agreed operating envelope.
Using theoretical maximum rack power across every rack can create an oversized battery calculation. Using a low average can create the opposite risk.
Temperature can materially affect available battery performance
Battery performance and life are temperature-sensitive. A battery room operating outside the conditions used for the manufacturer's runtime data can produce a different result from the published curve.
The effect also depends on chemistry. That is another reason this calculator does not automatically assign a universal VRLA or lithium-ion runtime factor.
Use the selected battery manufacturer's temperature and discharge corrections when those values become available.
VRLA and lithium-ion should not be reduced to one generic runtime multiplier
The existing UPS cost analysis compares the lifecycle implications of VRLA and lithium-ion batteries.
For runtime sizing, chemistry is only one part of the battery specification. Cell design, discharge rate, voltage limits, state of charge, temperature and battery management strategy all affect usable performance.
I would therefore avoid a calculator that claims, for example, that “lithium automatically provides X% more runtime.” The correct comparison is between actual battery systems and their published discharge characteristics.
Runtime at beginning of life is not enough for a resilient design
Imagine a battery system that provides exactly ten minutes when new and the business requirement is also ten minutes.
The architecture has no modeled room for aging, temperature differences, manufacturing tolerance or an unavailable bank.
A more useful question is whether the system still meets the target under the design condition that matters: beginning of life, end of life, maintenance state or another defined operating scenario.
The calculator's condition factor and available-bank count make those assumptions visible instead of burying them.
What this calculator deliberately does not calculate
This tool is intentionally a concept-stage energy model. It does not attempt to reproduce:
- manufacturer-specific discharge curves;
- cell-level voltage sag;
- Peukert-style or chemistry-specific discharge behavior;
- temperature correction tables;
- battery management system limits;
- DC cable losses;
- minimum UPS DC bus voltage;
- individual string imbalance;
- short-circuit or protection studies;
- UPS overload capability;
- generator transfer sequencing;
- battery recharge time after an event.
Those questions require equipment-specific battery and UPS data.
How I would use the calculator during concept-stage design
- Define the critical load that genuinely needs uninterrupted power.
- Define the operating or failure scenario: normal battery plant, one bank unavailable or another required condition.
- Enter the nominal battery energy available in that scenario.
- Replace the default usable-capacity assumption with the project's actual design basis.
- Model battery state of health or end-of-life capacity explicitly.
- Use manufacturer discharge data to establish an appropriate rate correction rather than accepting the illustrative default.
- Use the selected UPS battery-mode efficiency.
- Compare the resulting autonomy with the target runtime.
- Repeat the model at different loads and failure states.
- Replace this energy-balance estimate with manufacturer runtime curves before equipment selection.
Research basis and methodology
Eaton's UPS sizing guidance separates UPS capacity from battery runtime and explains that runtime can be increased through larger battery capacity or additional external battery packs.
Eaton also publishes model-specific battery runtime graphs across different UPS and battery-pack configurations.
Schneider Electric's UPS sizing and runtime guidance similarly shows that runtime changes with connected load and advises checking the runtime graph for the specific UPS.
Schneider's runtime-estimation documentation also identifies battery age, battery type, number of battery packs and applied load as inputs that affect estimated runtime.
The energy equation in this calculator is standard dimensional arithmetic: kWh divided by kW produces hours. The additional factors are exposed planning assumptions rather than claimed universal battery constants.
Methodology: use the calculator for early scenario comparison only. Final runtime should be verified using the selected UPS manufacturer's discharge curves, actual battery configuration, design temperature, state-of-health requirement and required failure condition.