Resource

Oil-Immersed vs Dry-Type Transformer: How to Choose

The choice is usually decided before any electrical parameter is discussed: if the transformer sits inside or below an occupied building, fire regulation pushes you to cast resin dry-type. If it sits outdoors or in a purpose-built substation, oil-immersed is cheaper per kVA, quieter and easier to source at high ratings. Where both are permitted, the decision turns on lifetime loss cost, noise limit and the enclosure you can actually ventilate.

Compared against IEC 60076-1/2/3/5/10/11 · IEC 62271-202 (compact substations) · EN 50464-1 · ANSI/IEEE C57.12.00 / C57.12.90 · IEEE C57.110 (K-factor) · IEEE 519 / IEC 61000-2-4 (harmonic limits on the building side) · ISO 9001 / 14001 / 45001 · UL 1561 listing — confirm per market · dB(A) figures below — confirm per type test report

Side-by-side comparison

Criterion Oil-immersed (S11–S22 / SH15) Cast resin dry-type (SCB / SCBH15)
Insulating medium Mineral oil IEC 60296; synthetic ester IEC 61099 or natural ester IEC 62770 on request Epoxy resin encapsulated windings, air cooled
Fire behaviour Flammable liquid inventory; requires containment, separation and — indoors — a fire-rated vault IEC 60076-11 class F1: self-extinguishing, low smoke, low toxicity. No flammable liquid
Typical location Outdoor, utility substation, dedicated transformer room Indoor, below grade, rooftop plant deck, occupied buildings
Rating range 15 – 31500 kVA 30 – 5000 kVA
Cooling ONAN / ONAF / OFAF / OFWF AN (natural air) / AF (forced air)
Efficiency Strong at all ratings; SH15 amorphous available for lowest no-load loss SCBH15 amorphous available; load loss typically higher than equivalent oil unit — confirm per rating
Noise Generally quieter — oil and tank damp core and winding vibration Windings radiate directly; AF cooling adds several dB over the AN rating
Maintenance Oil sampling and testing; sealed corrugated tank design removes conservator and breather No oil testing, no leakage risk; visual and thermal inspection, dust control
Overload capability Higher thermal inertia; forgiving of short-term overload Lower thermal inertia; derating required at high ambient or altitude
Environmental exposure Outdoor-ready; anti-corrosion to ISO 12944 C4/C5 IP20/IP23 indoor; IP31/IP54 or NEMA 3R for outdoor — with derating
Capital cost Lower per kVA, increasingly so above 2000 kVA Higher per kVA
Governing standards IEC 60076-1/2/3/5/7 · EN 50464-1 · ANSI/IEEE C57.12 IEC 60076-11 · EN 50588-1 · IEEE/ANSI C57.12.01 and C57.12.51 · UL 1561 — confirm per market

Fire safety: usually the deciding factor

Cast resin dry-type transformers contain no oil, so there is no flammable liquid inventory, no oil containment bund and — in most jurisdictions — no requirement for a separate fire-rated transformer vault. Units meeting IEC 60076-11 fire behaviour class F1 are self-extinguishing with low smoke and low toxicity, and halogen-free construction is available on request. That combination is what makes dry-type the default for indoor electrical rooms, basement and below-grade plant rooms, and rooftop plant decks.

Where a vault would otherwise be needed, the dry-type premium frequently cancels itself out. A fire-rated room with bunding, drainage, ventilation and oil-condition monitoring costs far more than the difference between two transformer prices — and it consumes floor area you may not have. Run that comparison before concluding that oil-immersed is the cheaper option.

Efficiency and lifetime cost: match the core to the load profile

Lifetime cost is dominated by losses, and which loss dominates depends on how hard the transformer actually works.

  • Energised 24/7 at partial load — no-load loss dominates. An amorphous core, which cuts no-load loss by roughly 40–50% — confirm per type test report, pays back fastest here. This is the classic lightly loaded rural feeder or a data hall running 30–40%.
  • Loaded near rating for most of its life — load loss dominates, and the amorphous premium buys much less. A high-tier silicon-steel design such as S20 or S22 is usually the better value.

The honest caveat on amorphous: it is not a free win. Amorphous cores are typically a few dB noisier than the equivalent silicon-steel unit, which matters immediately next to a ward, classroom or bedroom. In those locations, a silicon-steel unit plus anti-vibration mounting and acoustic treatment is often the easier route to a noise limit — and considerably more convincing to a client than “amorphous is best”.

Noise: read the declared figure correctly

Noise is the number one complaint source on building transformers, and it is also the number most often misread. The value declared from a type test per IEC 60076-10 is a sound power level, LWA — not a sound pressure level, and not what you will measure in the room.

The conversion is straightforward: in a free field, sound pressure at 1 m is approximately LWA − 11 dB, and it drops a further 6 dB every time the distance doubles. A real plant room with hard reflective walls will run higher than that estimate, because reflection adds energy rather than removing it.

Two further points that catch people out:

  • AF cooling is a hidden noise source. A dry-type unit rated AN can be very quiet, then jump several dB the moment the forced-air fans start. A noise-sensitive room must be specified against the peak cooling condition, not the natural-cooled figure.
  • Structure-borne sound beats airborne sound. Core and winding vibration travels into the building fabric and re-radiates elsewhere. Anti-vibration mounting is usually the single most cost-effective acoustic measure available, before any enclosure treatment.

Where each type wins

Application Recommended Why
Utility distribution feeder Oil-immersed (S20 / S22) Outdoor duty, cost per kVA, sealed tank removes maintenance
Lightly loaded rural feeder Oil-immersed amorphous (SH15) No-load loss dominates; fastest payback
Hospital, school, high-rise Cast resin dry-type (SCB13) Fire class F1, no vault, indoor placement
Data centre Cast resin dry-type, K-rated if measured THD justifies it Fire safety plus harmonic tolerance
Metro, tunnel, marine Cast resin dry-type No flammable liquid in confined or below-grade space
Renewables step-up, outdoor plant Oil-immersed (S20 / S22) Higher ratings, outdoor-ready, cost
Furnace and rectifier duty Special oil-immersed, designed to duty cycle Impedance and short-circuit withstand per IEC 60076-5
Rapid-deployment EPC YBW compact substation (either medium) Factory-tested single unit per IEC 62271-202

Selection guide — decide in this order

  1. Fix the location and fire position first. Inside or below an occupied building almost always means dry-type. This single decision removes most of the remaining options.
  2. Confirm the governing standard. IEC 60076 or ANSI/IEEE C57 changes insulation levels, temperature rise limits and terminal arrangement. Getting it wrong is the most common cause of site-acceptance failure.
  3. Rate against the real load profile. Partial load favours an amorphous core; near-full-load duty favours high-tier silicon steel.
  4. Measure harmonics before specifying a K-factor. Modern IGBT-rectifier UPS front ends often sit below 5% THD and need no K-rating. IEEE C57.110 defines K-factor; IEEE 519 and IEC 61000-2-4 are the limits that apply on the building side.
  5. Decide cooling method and enclosure together. IP54 severely restricts heat dissipation: it forces forced cooling or derating, and forced cooling pushes noise back up. These are one decision, not two.
  6. Correct for altitude and ambient. IEC 60076-2 requires temperature-rise correction above 1000 m. A site at 2500 m changes the rating, not just the paperwork.
  7. Specify noise as a pressure level at a stated distance. Ask for the guaranteed figure, not the LWA, and ask for it at the peak cooling condition.

Compliance and standards

Scope Standard Relevant to
General requirements IEC 60076-1 Both
Temperature rise, altitude correction IEC 60076-2 Both
Insulation, dielectric tests IEC 60076-3 Both
Short-circuit withstand IEC 60076-5 Both
Sound level determination IEC 60076-10 / -10-1 Noise-sensitive sites
Dry-type requirements, fire class F1 / E2 / C2 IEC 60076-11 Dry-type
Dry-type energy performance EN 50588-1 Dry-type, EU
Oil-immersed distribution ≤ 2500 kVA / 36 kV EN 50464-1 Oil, EU
North American dry-type IEEE/ANSI C57.12.01 and C57.12.51 · UL 1561 — confirm per market Dry-type, North America
North American general ANSI/IEEE C57.12.00 / .90 Both, North America
K-factor definition IEEE C57.110 Harmonic-rich loads
Harmonic limits, building side IEEE 519 / IEC 61000-2-4 Both
Energy efficiency GB 20052 · EU Ecodesign 2019/1783 (PEI) · DOE 10 CFR 430 Market-dependent — confirm scope per destination

Frequently asked questions

Which is cheaper, oil-immersed or dry-type?

Oil-immersed is consistently cheaper per kVA at the same rating, often by a wide margin above 2000 kVA. The dry-type premium buys fire safety and removes oil maintenance, containment and — in many jurisdictions — the fire-rated vault. On sites where those would otherwise be required, the saving frequently cancels the unit premium entirely.

Is a dry-type transformer safer than an oil-immersed one?

Dry-type removes the flammable liquid inventory, which is the property that matters in occupied buildings. Cast resin windings meeting IEC 60076-11 class F1 are self-extinguishing with low smoke and low toxicity. Oil-immersed units are safe outdoors and in purpose-built substations, but most codes restrict significant oil inventories inside occupied buildings.

Which transformer is quieter?

At the same rating, oil-immersed is generally quieter, because the oil and tank damp core and winding vibration. On a dry-type unit the windings radiate directly into the air, and the AF forced-air rating can add several dB over the quiet AN rating — so specify against the peak cooling condition, not the natural-cooled figure.

How do I convert the declared sound power level to an expected sound pressure?

The type-test value per IEC 60076-10 is a sound power level LWA. In a free field, sound pressure at 1 m is approximately LWA − 11 dB, falling a further 6 dB each time the distance doubles. A real plant room with hard reflective walls runs higher. Ask for the guaranteed sound pressure at a stated distance rather than the LWA alone.

Compare the SCB13 dry-type series →