How to Size a UPS for Hospital and Data Centre Critical Loads

Sizing a UPS starts with real connected load and its power factor, not the kVA number on an old single-line diagram, and it has to separate two different questions: how much power the load needs and how long the batteries need to hold it up. Get those two questions confused and you end up with a UPS that's the right size on paper and the wrong size the day the grid actually drops.

Connected Load and Power Factor Come First

Most oversizing and undersizing starts at the same place: someone reads a kVA figure off an old drawing and orders against it instead of measuring what's actually connected today. A UPS rated in kVA has a real kW output that depends on its power factor rating, and modern IT and medical loads often run at a power factor close to unity, while older UPS designs were rated at 0.8. Mixing those two assumptions is how a UPS that looks correctly sized ends up short on usable watts.

The fix is to build the load list from actual equipment schedules, not assumption, and confirm the power factor the UPS is rated at before comparing it against the kW the load draws. A hospital ICU wing and a data hall rack row have very different load profiles even at the same kVA figure.

Add a Growth Allowance Before You Lock the Number

A UPS sized exactly to today's connected load leaves no room for the server refresh, the added imaging suite, or the extra rack row that shows up eighteen months later. Facilities that size to the wire, with nothing spare, are usually the ones paying for a second UPS purchase within a few years instead of a planned capacity upgrade.

Modular UPS platforms solve this without forcing an oversized day-one purchase. Adding power modules as load grows means the frame and battery cabinets are sized for the eventual footprint while installed capacity matches what's connected now, and the growth allowance becomes a module swap rather than a full system replacement.

Runtime Is a Battery Decision, Not a kVA Decision

This is where sizing conversations go wrong most often. The UPS module's kVA rating tells you how much load it can carry, but it says nothing about how long it can carry that load. Runtime comes from the battery bank, and a bigger UPS module with a small battery string will drop the load in minutes just as fast as a small one.

Runtime needs to be specified separately, in minutes, based on how long the facility needs to bridge to generator start or safe shutdown. A hospital with a generator that starts and transfers within 10 seconds needs a different runtime target than a data hall bridging to a set that takes 30 seconds, and both are shorter than a facility with no generator that needs a full graceful shutdown window. Specify runtime at the actual load, size the battery string to deliver it, and treat the kVA rating as a separate number entirely.

End-of-Life Capacity vs Day-One Capacity

A battery string rated for 10 minutes of runtime when it's new isn't rated for 10 minutes for its whole service life. Lead-acid and lithium batteries both lose capacity as they age, and manufacturers publish end-of-life capacity, typically around 80% of rated capacity, as the point where replacement is due.

Sizing against day-one capacity and hoping the batteries hold that number for years is how facilities discover a runtime shortfall during an outage rather than during a scheduled test. The correct approach sizes the battery string so the runtime target is still met at end-of-life capacity, which means day-one runtime comes in comfortably above the design brief number.

Room Temperature Effects in Saudi Conditions

Battery life and rated capacity are both specified at a reference temperature, usually 20-25°C, and that reference has little to do with a plant room in Jeddah or Dammam that isn't perfectly climate controlled. Every degree above that reference shortens battery service life, and VRLA batteries in particular can lose a large fraction of their rated life for a sustained temperature rise of just a few degrees.

This matters twice in a Saudi sizing exercise: once for the derating applied to runtime calculations at actual room temperature, and once for the replacement interval the facility should plan around rather than assume from the datasheet. A UPS room without dedicated cooling, running warm through a Riyadh summer, needs both a runtime allowance and a shorter replacement cycle built into the plan from day one.

N+1 vs 2N Redundancy

Redundancy level changes what "sized correctly" even means. N+1 adds one extra module or UPS beyond what the load needs, protecting against a single module failure while keeping the rest of the system carrying load as normal. 2N duplicates the entire UPS system end to end, including separate battery strings and often separate distribution paths, so a total failure of one system doesn't touch the load at all.

Hospitals typically size critical branches to at least N+1, with life-safety and operating theatre loads often justifying 2N. Data centres split by tier and rack criticality, with 2N common for colocation halls carrying multiple tenants and N+1 accepted for less critical support loads. The redundancy level has to be decided before module count and battery capacity are finalized, not layered on afterward.

VRLA vs Lithium Batteries

Valve-regulated lead-acid batteries remain the lower upfront cost option and the more familiar one for maintenance teams, with a design life of 5-8 years under good conditions and shorter under Saudi ambient heat if the room isn't cooled. Lithium iron phosphate batteries cost more per kWh installed but typically run 2-3 times the service life, take roughly half the floor space and weight for the same runtime, and hold capacity better across a wider temperature range.

The right choice depends on the site more than a blanket preference. A retrofit into an existing plant room with limited floor loading often favors lithium despite the higher upfront number, while a new-build with space and cooling already budgeted can make VRLA the more economical choice over one replacement cycle. The decision should follow the runtime and space constraints of the specific site, not a default.

NLS supplies online double-conversion UPS systems with a choice of VRLA or lithium battery banks, sized against actual load lists and runtime targets for hospitals and healthcare facilities and data centres across the Kingdom, as part of its UPS and power backup supply work.

FAQ

Q: How do I size a UPS for a hospital or data centre? A: Start with actual connected load and the UPS power factor rating, not an old kVA figure, add a growth allowance for future load, and specify runtime in minutes separately from kVA capacity, sized so it still meets the target at the battery's end-of-life capacity rather than its day-one capacity.

Q: Does a bigger UPS mean longer runtime? A: No. The UPS module's kVA rating determines how much load it can carry, not how long. Runtime comes from the battery bank, and a large UPS module paired with a small battery string will drop the load just as fast as a small one.

Q: What's the difference between N+1 and 2N UPS redundancy? A: N+1 adds one spare module beyond what the load needs, protecting against a single module failure. 2N duplicates the entire system, including separate battery strings, so a full failure of one system still leaves the load fully protected by the other.

Q: Should I choose VRLA or lithium batteries for a UPS in Saudi Arabia? A: VRLA costs less upfront and is familiar to most maintenance teams, but loses service life faster in warm, uncooled plant rooms. Lithium costs more per kWh but typically lasts 2-3 times longer, takes less space, and holds capacity better across a wider temperature range, which often suits space-constrained retrofits.