Sizing a Distribution Transformer: A Practical Checklist

Add up every load on a project’s schedule and the total can easily suggest a 1000 kVA transformer. Yet the actual peak demand of the same installation may fit comfortably within 630 kVA. Oversizing increases capital cost and no-load losses for the life of the unit; undersizing risks overheating, accelerated insulation ageing and no room for growth.

This article answers the question: how do you move from a list of loads to a correctly sized, standard-rated distribution transformer — and what else must be decided beyond the kVA figure?

Understanding transformer sizing

Connected load vs maximum demand

• Connected load is the sum of the nameplate ratings of all equipment that could be connected.
• Maximum demand is the highest load actually drawn over a defined averaging period (often 15 or 30 minutes).

Not all loads run at full rating, and not all loads run at the same time. Lighting, HVAC, motors, sockets and process loads peak at different times and at different fractions of their rating. Two terms describe this:

• Demand factor = maximum demand of a load or group ÷ its connected load (normally ≤ 1).
• Diversity factor = sum of the individual maximum demands of load groups ÷ maximum demand of the whole system (normally ≥ 1). Its reciprocal is often called the coincidence (or simultaneity) factor (≤ 1).

These terms are used inconsistently across countries and documents — “a diversity factor of 0.8” usually means a coincidence factor. Always state which definition is used.

Sources for demand values, in order of reliability: measured data from similar or existing installations; utility or national design guidance; engineering judgement based on the load type. Each source has limitations, and the assumptions should be recorded.

Diversity and future expansion

After maximum demand is estimated, add a future-expansion margin reflecting planned additions (new building phases, additional production lines, EV charging, electrification of heating). Typical allowances are often in the range of 10–25%, but the right figure is a project decision, not a rule. Where expansion is uncertain, it may be more economical to plan space for a second transformer rather than oversize the first.

kW, kVA and power factor

Transformers deliver apparent power (S, kVA), because their heating depends on current, regardless of how much of that current produces useful work.

• Active power (P, kW) — useful power.
• Reactive power (Q, kvar) — power that circulates for magnetic and capacitive fields.
• Power factor (pf) = P / S.

Therefore: S = P / pf (and S² = P² + Q²).

At a lower power factor, the same kW demand needs more kVA. If power factor correction is planned, size on the realistic corrected value and consider the condition when correction is out of service. Harmonic-producing loads (drives, rectifiers, IT equipment) can also add heating not captured by a simple kVA figure and may require a derated or specially designed transformer; this needs manufacturer confirmation.

Voltage ratio and system configuration

The specification must define:

• Primary voltage — the MV supply (for example 11 kV, 20 kV or 33 kV), with the equipment insulation level.
• Secondary voltage — typically the no-load LV voltage (for example 400 V or 415 V between phases, depending on local practice); confirm whether the stated value is no-load or on-load.
• Tapping range — usually an off-circuit tap changer on the HV winding, for example ±2 × 2.5% (illustrative; confirm with the utility and design).
• Vector group — describes winding connections and phase displacement. Dyn11, widely used for distribution, means: HV winding in delta (D), LV winding in star (y) with neutral brought out (n), and a 30° phase displacement between HV and LV (clock-hour 11). The delta primary helps contain triplen harmonic currents and isolates zero-sequence behaviour between the two sides; the accessible neutral supports single-phase loads and the LV earthing system.

Transformers intended for parallel operation must have compatible vector groups, voltage ratios and closely matched impedances.

Oil-immersed vs cast-resin dry-type

Neither type is universally better. Local fire regulations, building codes, utility requirements and the installation location often drive the decision.

Cooling, ambient temperature and installation conditions

Cooling is designated by codes such as ONAN (oil natural, air natural) and ONAF (oil natural, air forced) for liquid-immersed units, and AN and AF for dry-type units. Forced cooling can increase output but adds fans, controls and maintenance.

Ratings are defined for standard ambient temperature conditions in IEC 60076 (limits for maximum, monthly-average and yearly-average temperature) and altitude up to 1000 m. Hotter climates, higher altitudes, or poorly ventilated rooms and enclosures reduce the effective capacity unless the design accounts for them. State the actual site conditions in the specification, including enclosure type, so the manufacturer can design or derate accordingly.

Losses, efficiency and lifecycle considerations

Transformer losses have two components:

• No-load (core) losses — present whenever the transformer is energised, regardless of load.
• Load (copper) losses — vary with the square of the loading: at a load fraction k, load losses ≈ k² × rated load losses.

For a transformer that is energised continuously but lightly loaded, no-load losses dominate the lifetime energy cost. Capitalising losses — assigning a monetary value per kW of no-load and load loss over the expected life — allows offers with different purchase prices and loss levels to be compared fairly.

Minimum efficiency or maximum loss levels may be mandatory in some markets — for example, the EU Ecodesign regulation for transformers. Requirements depend on the destination country and should be confirmed for each project.

Impedance and short-circuit implications

The short-circuit impedance (%Z or uk) is the percentage of rated voltage needed to drive rated current through a short-circuited secondary. It has two opposing effects:

• Lower %Z — better voltage regulation, but higher fault current on the LV side, requiring higher-rated LV switchgear and protective devices.
• Higher %Z — lower fault current, but greater voltage drop under load and during motor starting.

A simplified first estimate of the three-phase fault current at the LV terminals, assuming an infinite MV source and ignoring cable and other impedances:

Isc ≈ In / (%Z / 100), where In = S / (√3 × U).

This is a screening estimate only. A proper calculation (for example to IEC 60909) includes source impedance, voltage factors and motor contribution. Minimum impedance values and tolerances are addressed in the IEC 60076 series and must be confirmed with the manufacturer for each design.

Worked example: from maximum demand to a standard rating

Scenario (illustrative): a commercial building supplied at MV, with a 400 V three-phase LV system at 50 Hz.

Assumptions (stated explicitly)

1. Connected load from the equipment schedule: 760 kW.
2. A load analysis has already applied demand factors to each load group. The sum of the individual group maximum demands is 500 kW.
3. Coincidence factor between groups (because the groups do not peak simultaneously): 0.8 (equivalent to a diversity factor of 1.25).
4. Design power factor at peak demand: 0.9 (after any planned correction).
5. Future-expansion allowance: 20%.
6. Standard ambient conditions, altitude below 1000 m, linear loads (no significant harmonic derating), continuous ONAN or AN operation.
7. Standard ratings considered: 400, 500, 630, 800 and 1000 kVA.

Step 1 — Maximum demand in kW

P = 500 kW × 0.8 = 400 kW

Step 2 — Convert to apparent power

S = P / pf = 400 kW / 0.9 = 444.4 kVA

Step 3 — Add future expansion

S_future = 444.4 kVA × 1.20 = 533.3 kVA

Step 4 — Select a standard rating

• Without expansion, 500 kVA would be the next standard size above 444.4 kVA, giving a day-one loading of 444.4 / 500 = 88.9% — limited headroom for estimation error or growth.
• With the 20% expansion allowance, 533.3 kVA exceeds 500 kVA, so the next standard rating, 630 kVA, is selected.

↔ Scroll to see the full diagram

Step 5 — Check the resulting loading

• Day one: 444.4 / 630 = 70.5%
• After expansion: 533.3 / 630 = 84.7%

Both are below 100%, leaving some margin for estimation uncertainty. At 70.5% loading, load losses are about 0.705² ≈ 0.50 × rated load losses.

Step 6 — Contrast with connected-load sizing

If the transformer had been sized on connected load: 760 kW / 0.9 = 844.4 kVA, pointing to a 1000 kVA unit — more expensive to buy, with higher no-load losses for its entire life, and a day-one loading of only about 44%.

Step 7 — Indicative LV fault level

For 630 kVA at 400 V: In = 630 000 / (√3 × 400) ≈ 909 A.

• If %Z = 4%: Isc ≈ 909 / 0.04 ≈ 22.7 kA
• If %Z = 6%: Isc ≈ 909 / 0.06 ≈ 15.2 kA

The impedance choice changes the prospective LV fault current by roughly a third, directly affecting the breaking capacity required for the LV main breaker and downstream devices.

Limitations of this example

• The demand and coincidence values are assumptions; real projects should use measured data or recognised design guidance.
• Harmonics, unbalanced loading, cyclic loads and motor starting are not considered.
• Ambient derating, altitude and enclosure effects are not applied.
• The fault-level figures are a simplified estimate assuming an infinite MV source and ignoring cable impedance. Source impedance, motor contribution, voltage factors and other network conditions all affect the calculated fault duty, so the actual value may be higher or lower; a proper short-circuit study (for example to IEC 60909) is required, and this estimate is not a substitute for one.
• The %Z values are illustrative; actual values depend on the design and applicable standard.

Practical selection guidance — checklist

1. Build a load schedule and estimate maximum demand, not connected load.
2. State demand and coincidence/diversity factors and their sources.
3. Define the design power factor, with and without correction.
4. Add a documented future-expansion allowance — or plan for a second transformer.
5. Convert to kVA and select the next standard rating; check day-one and future loading.
6. Specify primary and secondary voltages, tapping range and vector group.
7. Choose oil-immersed or dry-type based on location, fire regulations and environment.
8. State ambient temperature, altitude, installation (indoor, outdoor, enclosure) and cooling.
9. Define loss requirements or capitalisation values, and any mandatory efficiency rules in the destination market.
10. Specify the impedance and coordinate it with LV switchgear fault ratings.
11. Confirm accessories: temperature monitoring, protection devices, cable boxes or bushings, and earthing provisions.

Common sizing mistakes

• Sizing on connected load alone, leading to oversized units with high lifetime no-load losses.
• Confusing kW and kVA, or ignoring power factor.
• Using “diversity factor” without defining it, so it is applied in the wrong direction.
• Adding safety margins at every stage (demand, diversity, expansion and “round-up”), compounding into heavy oversizing.
• Ignoring harmonic loads that require derating or a specially designed transformer.
• Not stating site ambient conditions, altitude or enclosure type.
• Selecting impedance without checking the LV fault level and switchgear ratings.
• Specifying the LV voltage without clarifying no-load vs on-load.
• Comparing offers on purchase price only, without considering losses.

Related AdepaTrade solution

AdepaTrade sources oil-immersed and cast-resin dry-type distribution transformers from Türkiye-based and international manufacturers, specified to your project’s ratings, losses and site conditions. Transformers