A transformer's kVA rating is a measure of apparent power, the combination of the real power a load actually consumes and the reactive power the load draws to build magnetic fields in motors, transformers and other inductive equipment. It tells you the maximum current the transformer's windings can carry continuously at rated voltage without exceeding their thermal limit, not the maximum real power (kW) a connected load can draw. Two installations with the same kVA rating can have very different usable real power capacity if their load's power factor is different. Sizing a transformer correctly means starting from the connected load's kVA demand, applying a realistic diversity factor, and leaving headroom for growth, not simply matching a transformer's kVA number to a building's sanctioned load in kW.
Every electrician and plant engineer has run into the same confusion at some point: a transformer nameplate reads, say, 500 kVA, but the load calculation on the drawing is in kW. These are not the same unit, and treating them as interchangeable is one of the more common sizing mistakes seen on site.
Apparent power (kVA) is the vector sum of real power (kW, the power that actually does work, heats a resistor, turns a motor shaft, lights a lamp) and reactive power (kVAR, the power that inductive and capacitive loads draw and return without doing useful work, but which still has to flow through the transformer windings and cabling). The relationship is:kVA = kW / power factor
A transformer has to be sized on kVA because its windings and core see the full current associated with the apparent power, regardless of how much of that current is doing useful work. A transformer does not know or care whether the current flowing through it is real or reactive, it only cares about total current and the heat that current generates in the winding resistance. That is why transformers, generators and UPS systems are rated in kVA rather than kW, while end loads like motors and heaters are more commonly discussed in kW.
This distinction matters directly for load capacity planning. A facility with a poor power factor, common in installations with a lot of induction motors, welding equipment or older fluorescent lighting, draws more kVA for the same amount of useful kW than a facility with a well corrected power factor close to unity. Two buildings with an identical 200 kW real load can require noticeably different transformer kVA ratings if one runs at 0.95 power factor and the other at 0.75 power factor.
For a three phase transformer, the standard relationship between kVA, voltage and current is:
kVA = (root 3 x Line Voltage in kV x Line Current in Amps) / 1000
Rearranged, this gives the maximum continuous current the transformer can supply at its rated secondary voltage:
Full Load Current (A) = (kVA x 1000) / (root 3 x Line Voltage)
This is the number that actually matters for load capacity decisions on the ground. When a design engineer or electrician asks "how much load can this transformer carry," what they are really asking is what current the secondary winding can deliver continuously without exceeding its temperature rise limit, at the transformer's rated voltage. The kVA figure is simply a convenient way of expressing that current limit independent of voltage, which is useful because the same transformer might be tapped to slightly different voltages in service.
It is worth being precise here because this formula gets misapplied often. The current a transformer can deliver is fixed by its kVA rating and its secondary voltage. It is not fixed by the load's kW demand alone. A transformer sized purely against kW figures, without accounting for the connected load's actual power factor, will run its windings hotter than the nameplate current suggests, because the real current flowing is higher than the "kW equivalent" current would imply.
There are three separate but related questions that come up when sizing or evaluating a transformer against a real building or plant load, and it helps to keep them distinct.
Connected load is the sum of the nameplate ratings of every piece of equipment that could be switched on. Actual demand is what the installation draws at any given time, which is almost always lower than the connected load because not everything runs simultaneously, and because motors and other equipment rarely run continuously at their full nameplate rating. A diversity factor (sometimes called a demand factor) is applied to connected load to arrive at a realistic maximum demand figure, and it is this demand figure, converted to kVA using the expected power factor, that should drive transformer sizing, not the raw connected load total. Sizing purely against connected load routinely produces an oversized, underutilised transformer.
As covered above, the same kW demand at a lower power factor requires a higher kVA transformer rating. This is precisely why many Indian electricity boards impose a power factor penalty on industrial and commercial consumers with poor power factor, and why power factor correction capacitor banks are such a common addition to plants with heavy motor loads. Correcting power factor at the load reduces the kVA (and therefore the current) the transformer has to supply for the same amount of real work, which can materially increase the usable real power capacity of an existing transformer without any change to the transformer itself.
A transformer's kVA rating on the nameplate is typically stated for a specific reference ambient temperature and a specific cooling method (ONAN, ONAF, OFAF, or dry type equivalents such as AN/AF for cast resin transformers). Running the transformer in a hotter ambient than its rated reference, in a poorly ventilated room, or at altitude, reduces its safe continuous loading below the nameplate figure. This is one of the most overlooked factors in Indian installations, where transformer rooms are frequently under-ventilated and ambient temperatures during peak summer months in cities like Hyderabad can sit well above the 40 degree Celsius reference many nameplates assume. A transformer that is technically correctly sized on paper can still run hot and age prematurely if the room it sits in does not allow it to reject heat the way its cooling class assumes.
Take a small industrial unit with the following connected load:
Step 1, apply diversity to get maximum demand in kW:
Step 2, convert each block to kVA using its own power factor, since a blended power factor across dissimilar loads is only an approximation:
Step 3, add headroom for future load growth and for the fact that diversity and power factor assumptions are estimates, not guarantees. A commonly used margin in Indian industrial practice is 15 to 25 percent above the calculated demand, depending on how firm the future expansion plans are.301.3 kVA x 1.20 = approximately 361.5 kVA
Step 4, round up to the next standard commercially available rating. Transformers are manufactured in standard steps (commonly 100, 160, 200, 250, 315, 400, 500, 630, 750, 1000 kVA and upward for distribution transformers), so the practical choice here would be a 400 kVA transformer rather than trying to source a custom 361.5 kVA unit.
Notice what this example demonstrates: the final transformer selection (400 kVA) is meaningfully larger than the raw connected load in kW (350 kW) might naively suggest, once diversity, power factor and margin are properly worked through, and it would be a mistake to read a 350 kW connected load as "needing roughly a 350 kVA transformer," which is a very common shortcut error.
A few patterns come up repeatedly in site audits and consultations:
Beyond the kVA figure itself, a few other nameplate details matter for understanding actual usable load capacity:
Cross referencing the nameplate against actual site conditions, not just the headline kVA number, is what separates a transformer that performs reliably for its full service life from one that runs hot, loses insulation life faster than expected, and needs premature replacement.
Being the top distributor of switchgear in Hyderabad we know transformer kVA rating decisions rarely sit in isolation. They connect directly to cable sizing on both the primary and secondary side, to switchgear and protection device ratings, to earthing design for the transformer's neutral and body, and to power factor correction planning if the facility's load profile changes over time. Getting the kVA sizing calculation right at the design stage, with realistic diversity and power factor assumptions rather than rule of thumb shortcuts, avoids a cascade of downstream problems where cabling, switchgear or protection settings end up mismatched to a transformer that was itself sized incorrectly.
For a deeper look at how transformer sizing interacts with the rest of the distribution design, including switchgear selection and protection coordination, see our switchgear selection guide on enarayan.com for a related technical breakdown from the same engineering perspective.
A transformer's kVA rating is a capacity ceiling, not a target operating point, and understanding this distinction matters for both efficiency and equipment life. Every transformer has two categories of loss. No-load loss (also called core loss or iron loss) occurs continuously whenever the transformer is energised, regardless of how much load it is carrying, because it comes from the magnetising current cycling the core. Load loss (copper loss) occurs in the windings and varies with the square of the load current, meaning it rises sharply as loading approaches the nameplate rating.
This relationship has a practical consequence for sizing decisions. A transformer running at a very low percentage of its rated kVA, say 15 or 20 percent, is still paying the full no-load loss continuously, which as a percentage of the useful energy delivered is comparatively wasteful. A transformer running consistently near its full rated kVA runs hotter and experiences faster insulation ageing, since transformer insulation life is a function of operating temperature over time, broadly following an exponential relationship where every meaningful rise in average operating temperature shortens expected insulation life. The commonly cited sweet spot for continuous loading on many distribution transformers sits somewhere in the 50 to 80 percent range of nameplate kVA, which is part of the reasoning behind the standard practice of sizing to calculated demand plus a reasonable growth margin, rather than either matching demand exactly or oversizing dramatically for headroom that may never be used.
Not every transformer on an Indian industrial or commercial site is the same construction, and the kVA rating interacts slightly differently depending on the type in use.
Oil-filled distribution transformers, the most common type for outdoor substations and larger installations, use mineral oil or increasingly ester-based fluids both as an insulating medium and as the primary heat transfer mechanism, carrying heat from the windings and core to the tank surface and radiators. Their kVA rating is tightly linked to the oil cooling method noted earlier (ONAN, ONAF, OFAF), and physical radiator condition, oil level and oil quality all directly affect how close to nameplate the transformer can actually be loaded in practice.
Dry-type (cast resin) transformers, increasingly common indoors, in basements, and in fire-sensitive locations such as hospitals, data centres and high-rise buildings where oil-filled units raise fire safety concerns, use air as the cooling medium, either natural (AN) or forced with fans (AF). Their kVA rating is more sensitive to enclosure ventilation and room airflow than an oil-filled unit's, since there is no oil reservoir buffering short-term thermal transients, and a poorly ventilated dry-type transformer room can meaningfully undercut the nameplate rating in real operating conditions.
Auto-transformers, used in specific applications such as motor starting or voltage step changes where full electrical isolation between primary and secondary is not required, have a different relationship between their physical size and their apparent kVA rating compared to a two-winding transformer of the same rating, generally making them more compact and economical for the specific use cases where they are appropriate, though not a substitute for a standard isolating transformer where isolation is a safety requirement.
On larger sites, load growth is sometimes met by adding a second transformer in parallel with an existing one rather than replacing the existing unit outright. This is a legitimate approach to expanding kVA capacity, but it comes with conditions that are easy to overlook if the focus stays purely on the combined kVA total. Transformers operated in parallel need matching vector groups, closely matched percentage impedance values, and matched voltage ratios, otherwise circulating currents between the two units can develop even under no external load, wasting capacity and generating unwanted heating. Where these conditions are met, the combined kVA capacity of two paralleled transformers is a genuinely useful and common way to grow capacity incrementally rather than replacing a single unit with a much larger one ahead of confirmed demand. Where they are not met, attempting to parallel mismatched transformers can create more problems than the capacity increase solves, and this is a decision that should always involve a qualified electrical engineer reviewing the specific nameplate data of both units before proceeding.
A transformer's ability to deliver its nameplate kVA reliably over its service life depends heavily on maintenance practice, not just the original design and sizing calculation.
Treating transformer sizing as a one-time calculation at design stage, rather than a figure that should be periodically revisited against actual measured load, is one of the quieter but more consequential gaps in ongoing electrical maintenance practice on many sites.
1. Is a higher kVA rating always better for a transformer?
No. Oversizing has real costs, higher no-load losses running continuously, a larger footprint, and a higher purchase price, without a proportional benefit if the actual demand never approaches the rating. The goal is a rating that comfortably covers calculated maximum demand plus a sensible growth margin, not the largest available unit.
2. Can I estimate transformer size just from the building's sanctioned load in kW?
Not reliably. Sanctioned load figures from the utility are often a contractual ceiling rather than an actual measured demand, and they do not tell you the power factor of the load. A proper kVA sizing calculation, working through connected load, diversity factor and power factor by load category, gives a far more accurate answer.
3. Does power factor correction change how big a transformer I need?
Yes, meaningfully. Correcting a poor power factor at the load reduces the kVA (and current) drawn from the transformer for the same real power output, which can free up genuine spare capacity on an existing transformer, sometimes avoiding an upgrade that would otherwise appear necessary from raw kW growth alone.
4. Why do transformers come in standard kVA steps rather than any custom value?
Standard ratings (100, 160, 200, 250, 315, 400, 500, 630 kVA and so on) reflect manufacturing standardisation across the industry, referenced in Indian and international transformer standards, which keeps lead times, spares availability and cost predictable. A calculated requirement is always rounded up to the next standard step rather than custom manufactured for a specific figure.