Hospital UPS runtime: how to calculate, test and maintain it

Hospital UPS runtime is the time a particular system can supply a defined load, under specified battery and temperature conditions, before it reaches a safe end point. It is not a fixed number on a product sheet. For an operating theatre, intensive care unit or other critical service, the useful figure is the runtime that can be demonstrated under a realistic operating scenario.

A defensible estimate starts with the measured load profile, the UPS efficiency and the battery manufacturer’s discharge data. It also accounts for ageing, temperature and the behaviour of the alternative supply. Verification then requires comparable measurements and a planned test that protects clinical activity. The aim is to establish a usable window for the intended transition, not to drain the batteries simply to record the longest possible time.

What must the UPS keep running, and for how long?

The hospital first needs to define the load and the event it is preparing for. Bridging the start-up and stabilisation of a generator differs from keeping a clinical procedure running to a safe conclusion, supporting a controlled shutdown or maintaining services during a longer loss of supply. Each case leads to a different load, target duration and acceptance criterion.

NHS HTM 06-01 refers to three hours of battery autonomy for certain Group 2 medical locations; this may be reduced to one hour where the safety supply becomes available within 15 seconds. That is UK guidance, not a substitute for the rules applicable to a project elsewhere. Its practical lesson is that the target duration belongs to a documented continuity strategy agreed with clinical and engineering teams.

From connected equipment to the actual load profile

The sum of equipment nameplate ratings rarely equals the continuous demand. Runtime calculations should use measurements taken during representative procedures, recording both the steady load and short peaks. The team should distinguish equipment that must remain energised throughout an interruption from loads that can be shed in a controlled way or are needed only during transfer.

Multiplying power by time gives a useful starting estimate of energy demand. A proper battery selection must then account for UPS losses, minimum battery voltage, discharge rate and the manufacturer’s limits. A battery does not deliver the same usable capacity at every current: a higher discharge current can shorten runtime non-linearly. Manufacturer discharge curves or tables are therefore more reliable than dividing a nominal ampere-hour rating by the expected current.

Five variables that change available UPS runtime

1. Load profile. Average power matters, but so do peaks, the order in which equipment connects and the loads that remain active during a contingency.
2. Battery condition. Remaining capacity, internal resistance, variation between blocks and connection quality influence the whole bank.
3. Temperature. Low temperatures can reduce available capacity, while high temperatures accelerate ageing. Record the temperature during every test.
4. Age and operating history. Calendar time, cycles, deep discharges, float charging and environmental conditions can make otherwise similar banks behave differently.
5. Transfer to the alternative supply. Generator start-up is only one part of the sequence. Stabilisation, load acceptance and failed-transfer scenarios also consume battery time.

Estimate runtime without suggesting false precision

An initial calculation compares the energy required by the critical load with the usable energy in the battery bank. The answer must retain its conditions: output power, power factor, temperature, end voltage, assumed battery age and system redundancy. A claim of “60 minutes” without those conditions implies a level of certainty the installation may not provide.

The design should also address a credible worst case rather than an impossible combination of faults. That might be a mains failure during a high-demand procedure, delayed generator start-up or one unavailable redundant branch. Engineering and clinical teams must decide which events the electrical design covers and which require an additional operational measure.

UPS selection and integration

IEC 62040-1 and IEC 62040-3 provide references for UPS safety, performance and test methods. The project must also address applicable installation rules, redundancy, bypass arrangements, protection, ventilation and generator compatibility.

ETKHO’s ETK UPS range offers uninterruptible power systems for hospital loads. Selection should start with the measured profile, not nominal power alone. Relevant questions include overload capability, efficiency at the normal operating point, maintenance arrangements and the interfaces available for alarm integration.

An oversized UPS may spend much of its life operating at a less efficient load point. A system sized without a justified margin may struggle with future additions or transients. The reserve must be coordinated with the battery bank: increasing UPS power rating does not, by itself, increase runtime.

A battery bank is only as dependable as its weakest links

In a series string, a deteriorated block can limit the performance of the entire bank. Total voltage alone does not reveal the full picture. Differences between units, resistance or conductance, temperature, connections and protective devices all warrant attention. Readings should be interpreted against manufacturer guidance, a documented baseline and comparable modules in the same bank.

ETKHO lists AGM batteries in several capacities, including a 12 V, 40 Ah model. The illustration uses that rating as a visual reference. String configuration, block count, voltage and protection must be selected for the actual UPS and required duration. Mixing technologies, ages or capacities calls for an explicit manufacturer assessment.

Runtime testing: scope, safety and end point

A useful test begins with an approved procedure. It specifies the load, target duration, end voltage, expected alarms, abort conditions, responsibilities, recovery method and clinical contingency plan. Where possible, testing uses a controlled load or an agreed maintenance window. A bank that has not recharged must not be left exposed to a second mains failure without a suitable contingency.

1. Confirm that the mains supply, bypass and alternative source are available, and investigate any active alarms.
2. Record the load, temperature, string voltages and initial condition of supervised modules.
3. Start the discharge against a common time reference and check UPS and protection behaviour.
4. Log voltage, current, temperature and alarms at defined intervals without changing the acceptance rules during the test.
5. Stop at the target duration or pre-defined limit, before a damaging depth of discharge.
6. Restore supply, confirm recharging and document abnormal modules, connections or alarms.

Thermography can reveal high-resistance terminals and differences between comparable components. Conductance or internal-resistance measurements can reveal trends, but neither measures the energy delivered over the full required interval on its own. Acceptance should combine suitable methods and retain traceable records of instruments and conditions.

Maintenance based on trends as well as schedules

Design life is a planning reference, not proof of remaining capacity. A hospital maintenance programme should include visual inspection, cleaning, temperature control, connection checks, alarm review, module trends and functional tests proportionate to risk. Location, ventilation and the precautions relevant to the battery technology must be considered alongside electrical measurements.

Investigation thresholds should be set before the UPS reports an end-of-life condition. Increasing variation between blocks, shorter measured runtime, persistent heat or more frequent alarms may justify earlier action. After battery replacement, a new baseline makes future comparisons meaningful.

Redundancy and maintainability: runtime cannot stand alone

A sufficient battery duration has limited value if the system cannot be maintained safely. The architecture should define what happens when the UPS, bypass, battery string or output distribution board is taken out of service. In a redundant arrangement, losing one module must not overload the others or reduce runtime below the agreed target. “N+1” is useful only when the redundant component and load conditions are specified.

The maintenance bypass belongs in operating procedures and staff training. An incorrect switching sequence can remove protection even when every item of equipment is functional. Drawings, interlocks, labels and procedures must match the installed system. Alarm communication also needs testing: a warning shown on the local UPS screen but not delivered to the person responsible is of limited operational use.

Recovery after a discharge is part of the design case. While the batteries recharge, the bank has less than its full autonomy and the UPS may draw additional power. The project should consider source capacity, charging strategy, ventilation and the risk of a second interruption. If continuity must be preserved throughout recovery, independent strings, staged charging or temporary procedures may be required.

Test generator transfer as a complete sequence

The UPS bridges the interval between loss of the normal supply and availability of a stable alternative source. The sequence includes fault detection, generator start command, acceleration, voltage and frequency stabilisation, breaker closure and progressive load acceptance. A delay or rejected load uses up battery time. An integrated test can reveal oscillation, overload or repeated transfers that a UPS-only test would miss.

The plan should also consider a failed first start, a generator out of service and the return of mains power during transfer. Not every scenario requires a full battery discharge, but each needs documented logic and proportionate verification. Runtime then becomes a known operating window rather than an assumption first tested by an incident.

What an acceptance report should record

The report should make it possible to reconstruct the test months later. Identify the UPS, strings and blocks, relevant configuration, connected load, ambient conditions, date and initial state. A time-based record should show the start, alarms, transfers, end voltage, stop and restoration, including any manual intervention.

Do not hide deviations behind a simple pass or fail. If one block shows unusual temperature or conductance, the load changes during the test or the generator takes longer than planned, record the observation and assign an action, owner and deadline. That makes testing useful for maintenance decisions.

Common mistakes

• Confusing the UPS power rating with battery runtime.
• Adding equipment nameplate ratings without measuring the load profile.
• Using ampere-hours without consulting the discharge curve.
• Ignoring temperature, ageing or variation between blocks.
• Testing without an end point, contingency or recovery period.
• Assuming a generator removes the need for adequate UPS runtime.
• Keeping a report without the load, temperature, instruments and alarm log.

A runtime figure the hospital can defend

A defensible runtime protects an identified load during an agreed event, under recorded conditions and to a safe end point. That traceability gives engineering, maintenance and clinical teams a common basis for decisions, and makes later tests comparable.

Frequently asked questions