How to calculate power transformer capacity
Sizing a transformer wrongly costs money either way. Too small and it overheats; too large and you pay for capacity and no-load losses you never use. This guide sets out how to calculate the rating you need, step by step, with a worked example.
Basic formula
Transformer capacity (kVA) = (total load in kW) / power factor
For three-phase systems: kVA = (√3 × V × I) / 1000
Step 1: calculate total connected load
Add up the ratings of everything connected to the transformer:
- Motors and drives
- Lighting systems
- Heating and cooling
- Process equipment
- Other electrical loads
Use the rated input of each item, not its output. For a motor, that means the electrical input power, and for drives it means the input side. Where the data sheet gives output in kW or horsepower, convert it using the efficiency of the machine.
Step 2: apply diversity factor
Not every load runs at once, so apply a diversity factor, typically 0.6 to 0.9:
Maximum demand = total connected load × diversity factor
Diversity is the step that most often goes wrong, in both directions. Applying an aggressive diversity factor to a plant whose loads genuinely do run together leads to chronic overload; ignoring diversity in a large mixed-use installation leads to a transformer running at 30% load for its entire life, paying no-load loss for nothing.
Step 3: consider power factor
Real power (kW) and apparent power (kVA) differ by the power factor:
kVA = kW / power factor
Typical power factor: 0.8 to 0.95
A low power factor is expensive twice over: it inflates the transformer rating you have to buy, and it increases current for the same useful power, which raises load loss. Where the load is dominated by motors or by lightly loaded induction equipment, power factor correction at the load is usually cheaper than buying a larger transformer.
Step 4: add future growth margin
Allow for load growth, typically 20 to 30%:
Required capacity = calculated load × (1 + growth margin)
Step 5: select standard rating
Choose the next standard rating above your calculated figure. Standard ratings follow a preferred number series, so the choice is not continuous.
For large power transformers, standard ratings include:
- 10 MVA, 20 MVA, 31.5 MVA, 40 MVA, 50 MVA
- 63 MVA, 90 MVA, 120 MVA, 150 MVA, 180 MVA
- 240 MVA, 300 MVA, 360 MVA, 420 MVA, 500 MVA and above
Example calculation
For a substation with:
- Total connected load: 80 MVA
- Diversity factor: 0.8
- Power factor: 0.9
- Growth margin: 25%
Calculation: 80 MVA × 0.8 / 0.9 × 1.25 = 88.9 MVA
Recommended: a 90 MVA or 100 MVA transformer
Beyond the arithmetic: what the number does not tell you
Loading limit and expected life
A transformer rated 100 MVA is not intended to run at 100 MVA indefinitely. Continuous loading limits follow the temperature rise the insulation can tolerate, and standard practice is to design for a hot-spot temperature that yields normal insulation life at rated load. Loading above rated load is possible — the standards define emergency loading — but it consumes life faster. If the load profile has daily peaks well above the average, that belongs in the sizing decision, not in an assumption that the transformer will cope.
Ambient temperature and altitude
Nameplate ratings assume a reference ambient temperature and an altitude below 1,000 m. Both affect the answer. A site with 45 °C summer ambient or an altitude of 2,000 m must derate the unit or select a larger frame; the calculation above says nothing about this, and it is a common source of overloaded installations.
Starting and inrush conditions
Large motor starts and transformer energisation both draw temporary currents well above steady-state load. Protection settings must ride through these, and the transformer must tolerate the voltage dip they cause. Where the site includes large direct-on-line motors, check that the calculated rating still delivers an acceptable voltage dip.
Losses and cost of ownership
Two transformers of the same rating can differ substantially in loss. Because no-load loss runs continuously and load loss scales with the square of current, the right comparison is capitalised loss cost rather than equipment price. Our guides to losses and total cost of ownership and to substation transformer selection cover how that comparison is made.
Summary of the method
| Step | Action | Typical factor |
|---|---|---|
| 1 | Total the connected load in kW | — |
| 2 | Apply diversity factor | 0.6 to 0.9 |
| 3 | Divide by power factor | 0.8 to 0.95 |
| 4 | Add growth margin | 20% to 30% |
| 5 | Select the next standard rating | Preferred number series |
| 6 | Check ambient, altitude and overload needs | Site specific |
Contact our engineers with your load list and we will help you confirm the rating.
Where the calculation gets complicated
The five-step method assumes loads are known and reasonably stable. Three situations make it harder, and all three are common enough to be worth naming.
- Motor-heavy loads. Motors draw several times their rated current at start, so a plant dominated by large direct-on-line motors has to be checked against starting voltage dip, not just steady-state demand.
- Non-linear loads. Rectifiers, drives and IT equipment generate harmonic currents that raise winding eddy losses. The transformer may need a harmonic loss allowance or a higher rating than the steady-state calculation suggests.
- Cyclic load. Where load peaks follow a daily or weekly pattern, average load is not the sizing criterion — the peak is, and the thermal time constant of the transformer determines how long a peak can be sustained.
Confirming the result
Once the rating is selected, check three things before ordering: that the chosen rating delivers an acceptable voltage dip on motor starting; that the losses are specified rather than left to the supplier; and that the impedance value suits the protection and switchgear downstream. Each of these is easier to change at the specification stage than after manufacture.