“How many minutes will the UPS run?” sounds like a simple question, but the answer changes with load, battery size, chemistry, temperature, age, DC voltage limits and inverter efficiency. Small packaged UPS products often publish runtime curves that are easy to use, while larger three-phase systems require manufacturer-supported battery calculations. The most important step comes before the mathematics: define why autonomy is needed. A data centre may need enough time to bridge generator start and stabilisation. A healthcare system may need to preserve a clinical function until a safe state is reached. A remote telecom site may need hours because restoration takes longer. Once the scenario is clear, the battery can be sized and tested against an explicit end-of-life requirement rather than a vague number of minutes.
Define the autonomy scenario first
Write the event as a timeline. At time zero, the normal source fails. The UPS immediately supports the critical load. What happens next? Does a generator start? Does an operator receive an alarm? Are non-essential loads shed? Is a process moved to a safe state? At what point is controlled shutdown initiated? The required autonomy should cover these actions plus a justified margin.
This approach exposes unrealistic assumptions. A specification may demand 15 minutes because that is a familiar number, even though the generator is required to be available much sooner. Conversely, a remote location may be given ten minutes despite a one-hour technician response. Tying runtime to the operating sequence helps the organisation decide whether more battery, faster recovery, load shedding or a second energy source is the most effective way to reduce risk.
Why simple watt-hour division is unreliable
A rough energy calculation can help build intuition: power multiplied by time gives energy. However, a UPS battery is not an ideal energy bucket. Available capacity depends on discharge rate, chemistry, temperature, age, cell voltage and the UPS DC operating limits. The inverter also has losses. Lead-acid batteries, in particular, do not deliver the same usable ampere-hours at every discharge rate.
For packaged systems, use manufacturer runtime curves at the intended load. For larger systems, request a formal battery-sizing calculation. The supplier should state the number of cells or modules, end voltage, temperature, ageing margin and load assumptions. If proposals use different assumptions, their battery sizes and prices cannot be compared fairly. A transparent calculation is more useful than an unexplained “20-minute battery” label.
Specify runtime at a stated load
Runtime falls as load increases. If a 100 kW UPS is only supporting 40 kW today, a battery may appear to provide generous autonomy. Future growth to 80 kW can reduce that time substantially. Decide whether the project requirement applies to current load, design load or some future scenario. For modular expansions, confirm whether batteries will be expanded at the same time.
Also decide whether all protected loads remain connected throughout the event. Staged shedding can preserve runtime for the most critical functions. A telecom cabinet might disconnect cooling or ancillary loads according to temperature and battery state; an industrial process might shed non-essential controls after a defined period. Model the load profile over time rather than assuming one constant figure if the operating sequence is more complex.
Include ageing and end-of-life performance
New batteries should exceed the minimum required capacity because their performance will decline. If the system must still provide the target autonomy shortly before planned replacement, the initial battery must be sized with an appropriate ageing assumption. Manufacturer design tools commonly include ageing factors or permit end-of-life criteria to be specified.
The operating organisation should know what condition triggers replacement. Calendar age alone is a weak indicator because temperature, cycling and charging history affect degradation. Combine planned replacement windows with monitoring and periodic testing. If a battery string no longer supports the agreed autonomy, the UPS system may still appear normal during utility operation while its resilience case has already failed. Treat reduced runtime as a capacity defect, not merely a future maintenance concern.
Temperature changes the result
Battery calculations normally use a reference temperature. In colder conditions, available discharge capacity can fall; in hotter conditions, short-term capacity may not be the main concern but accelerated ageing can shorten life. The correct design temperature depends on the actual battery environment. A cabinet on an outdoor telecom site requires different assumptions from a conditioned data-centre battery room.
Monitor the battery environment continuously where consequence warrants it. Avoid locating temperature sensors only near room air returns if the hottest cells sit elsewhere. During an outage, check whether cooling remains available. A battery room that rises quickly in temperature because mechanical cooling is not on the resilient supply may experience conditions outside the design case during the very event the batteries are intended to cover.
Recharge time and repeated outages
Autonomy is only half of the energy-storage story. After a discharge, the battery must recharge. If a second outage occurs before full recharge, available runtime may be reduced. The recharge rate also creates an electrical load on the UPS input and therefore on a standby generator if mains has not returned. Large, rapid recharge can compete with other essential loads.
Specify the required recharge performance and coordinate it with source capacity. Many UPS systems allow charger current or input current to be limited, which can protect a generator at the expense of slower recharge. In locations with unstable utility supply or frequent cycling, repeated partial discharges deserve particular attention. Battery chemistry and cycle capability may influence the best technology choice.
Testing runtime safely
A battery self-test is useful for detecting some faults, but it does not necessarily demonstrate full required autonomy. Periodic discharge testing can provide stronger evidence, especially for critical systems, but the method must be designed so the test itself does not expose the load. Options can include load banks, redundant strings, controlled partial discharge or testing during maintenance windows.
Record starting state, load, temperature, battery voltage, cell or block readings, elapsed time and end condition. Compare results with the acceptance criterion rather than simply noting “test passed”. If the battery cannot meet the requirement, investigate whether the cause is individual weak blocks, string imbalance, environmental condition, charger issue or general ageing. Corrective action should restore the defined autonomy, not just clear an alarm.
A robust runtime statement
A good specification might say: “The battery system shall support the defined critical load of X kW for Y minutes at the stated ambient temperature, at the agreed end-of-life condition and end voltage, with the UPS operating in the specified mode.” It can also state required recharge time and whether the autonomy includes a future load allowance. The exact values must come from the project, not from a generic example.
This wording makes proposals comparable and commissioning measurable. It also gives the operations team a clear target for future testing. Runtime then becomes an engineered property of the resilience plan rather than a marketing number displayed on the UPS front panel.
Standards and official guidance may be amended. Confirm the edition and project-specific requirements with a competent professional before design, procurement or maintenance work.