Buyer's guide

Dry-type or oil-immersed distribution transformer: what actually forces the choice

12 August 2026 13 minute read

The question an export buyer or panel builder actually faces is not which technology is better, but two narrower ones in sequence: which one is admissible at this site, and what it costs to own for the next twenty years. Fire load and installed location settle admissibility first. Insulation class, cooling, overload margin, part-load losses and maintenance settle the cost question after.

Where each may actually be installed

The US model electrical code treats both technologies under a single article, but it keys the location rule to kVA, voltage and insulation class rather than issuing a blanket verdict on which medium is preferable. That structure is itself informative: the code does not treat dry as automatically safer or oil as automatically restricted — it scales the requirement to the specific unit.

Temperature class is the lever that buys siting flexibility for dry-type units. Above 112.5 kVA, an indoor dry-type transformer normally needs a dedicated fire-resistant transformer room — unless it uses Class 155 or Class 180/H insulation, in which case roughly 1.8 m of horizontal clearance and 3.7 m of vertical clearance from combustible material is an accepted substitute for that room.

Oil-insulated units installed indoors generally require a vault, with fire-resistance construction for the walls, roof and floor; a one-hour reduction in that rating is commonly permitted where automatic sprinklers are fitted. Above 35 kV, the vault requirement applies to dry-type transformers as well — so voltage, not insulating medium, drives that particular threshold.

The obligation that attaches only to the oil-filled case sits outside the electrical code altogether. A facility holding enough oil-filled transformer capacity must maintain a written Spill Prevention, Control and Countermeasure plan under US federal environmental rule, because transformers are explicitly named as a covered class of oil-filled equipment. A dry-type installation carries no equivalent obligation — not because of an exemption clause, but because there is no oil inventory for the rule to apply to.

Working space clearance around either technology is a separate requirement again, scaled to system voltage under federal workplace safety rule, and it applies regardless of which transformer type is chosen.

Insulation system and temperature class

What a temperature class buys in clearance

A temperature-class designation — 105/A up through 220/R and higher — fixes a maximum permitted winding temperature and an associated rise over a stated ambient. A higher class lets the same physical winding run hotter without exceeding its design limit, and that headroom is exactly the mechanism behind the clearance exception described above. Class 155 and Class 180/H insulation earn reduced spacing because their design margin makes the reduced-clearance case acceptable, not because the code treats dry-type as inherently lower-risk.

Two independent institutional design standards, from different universities, specify the identical figure for Class H insulation: a 220 °C total temperature limit with a 115 °C rise over a 40 °C ambient. Two separate procurement documents landing on the same number is genuine confirmation that this is a standard specification item, not an unusual or premium option.

One comparability warning belongs here rather than in the losses section below: the regulator applies a different reference temperature to load-loss measurement by technology — 55 °C for liquid-immersed units, 75 °C for dry-type. Published efficiency figures for the two technologies are therefore not directly comparable without adjusting for that difference first.

Cooling designation: what the letters on a nameplate mean

The international cooling-designation system is a four-letter code that states, in order, the internal cooling medium, how it circulates, the external cooling medium, and how that circulates. It describes a cooling arrangement, not a transformer grade — a distinction worth making explicit because buyers sometimes read the code as a quality tier.

A liquid-filled unit's unforced base rating and its fan-assisted rating are two nameplate figures on one core-and-coil assembly, not two different transformers. Dry-type units follow the same logic using air on both sides of the code: adding forced air creates a second, higher rating exactly as fan-assist does for the oil-filled case, but no liquid medium exists at either rating.

For scope-checking purposes: IEC 60076-11:2018 covers dry-type power transformers, including auto-transformers, up to 72.5 kV, and it excludes single-phase units under 5 kVA and polyphase units under 15 kVA — worth confirming before an enquiry assumes a rating sits inside the standard.

Short-term overload and short-circuit withstand

Short-circuit withstand — a winding and support structure's ability to survive the mechanical and thermal stress of an external fault without damage — is governed by its own dedicated standard, applying to winding and core construction generally rather than to one cooling medium specifically. The impedance figure matters beyond the transformer itself: it is one of the factors that sets the fault level a breaker on the low side has to clear, so a technology choice has downstream consequences for protective device selection.

Overload guidance is published as two separate standards rather than one: one covers loading of oil-immersed power transformers, the other covers loading of dry-type power transformers. Both compute permissible loading above nameplate rating against ageing of the insulation system, using the same underlying methodology, but the two documents are not interchangeable, because the two insulation systems age differently. A parallel IEEE guide exists specifically for dry-type loading, confirming the split is a standards-level distinction, not a marketing one.

No numeric overload multiplier is asserted here — no citable source gives one consistent figure for either technology. What can be stated qualitatively: oil's thermal mass and liquid cooling give an oil-immersed unit materially more short-term overload headroom than an air-cooled dry-type unit, and the fact that the standards diverge by technology rather than sharing one table is itself evidence for that difference.

Part-load losses and the efficiency test basis

US regulation recognises three separate distribution-transformer categories for efficiency purposes — low-voltage dry-type under 600 V, medium-voltage dry-type from 600 V to 35 kV, and liquid-immersed — each with its own efficiency table, because loss behaviour and thermal reference conditions differ between them.

Certification testing is not run at full nameplate load. Liquid-immersed and medium-voltage dry-type units are tested at 50% of rated load; low-voltage dry-type units are tested at 35%. Both figures reflect typical in-service average loading more closely than 100% would.

No-load loss is constant once the unit is energised, regardless of what current is drawn. Load loss — winding loss plus stray loss — rises with the square of the load current. Because one loss component is fixed and the other is load-squared, the relative efficiency ranking between the two technologies can shift depending on the load factor a buyer actually expects to run, which is exactly why the worked example below tests a specific load factor rather than comparing nameplate ratings alone.

Worked example: total cost of losses over twenty years

The figures below are constructed for this example only. Assume two 800 kVA transformers at the same rating, purchased for the same duty:

• Unit A, oil-immersed: no-load loss 1,050 W; load loss at rated output 8,200 W.
• Unit B, dry-type: no-load loss 2,150 W; load loss at rated output 9,600 W.
• Average load factor: 0.55, and load loss scales with the square of load factor.
• Energy price: US $0.14/kWh. Service life: 20 years at 8,760 h/year.

Annual loss energy, Unit A:

1,050 W × 8,760 h = 9,198 kWh (no-load)
8,200 W × 0.55² × 8,760 h = 21,729 kWh (load)
Total ≈ 30,927 kWh/yr → $4,330/yr → $86,596 over 20 years

Annual loss energy, Unit B:

2,150 W × 8,760 h = 18,834 kWh (no-load)
9,600 W × 0.55² × 8,760 h = 25,439 kWh (load)
Total ≈ 44,273 kWh/yr → $6,198/yr → $123,965 over 20 years

The gap over the stated life is roughly $37,400. That means if Unit B's purchase price is more than about $37,000 lower than Unit A's, the arithmetic still favours Unit A — the cheaper transformer at the loading dock can still be the more expensive one to own. This is an illustration built on chosen figures, not a claim about which technology is generally cheaper to run; the actual answer depends on the specific units, load factor and local energy price in front of the buyer.

Maintenance obligation over the service life

Dissolved gas analysis — laboratory analysis of gases dissolved in insulating oil — is the standard diagnostic method for detecting incipient winding, insulation or core faults in an oil-filled transformer, and it is codified in a dedicated IEEE guide. There is no equivalent sampling regime for a transformer that carries no oil, because there is nothing to sample.

Both loading guides treat insulation ageing as a rate calculation tied to operating temperature, rather than a fixed service-life figure — that much is true for both technologies. What differs is the input variable available to monitor: oil condition on one side, winding temperature history on the other.

This is not a claim that one technology needs more maintenance than the other. It is that the maintenance regimes differ in kind: one has a fluid to sample and interpret, the other does not.

Noise: two different test standards, not one number

An institutional design standard requires dry-type sound levels to be verified against a national trade-association sound standard, tested at least 3 dB below its stated limit, with vibration isolation and, where applicable, seismic restraint specified alongside. Dry-type is not inherently silent — noise is a specified, testable parameter regardless of cooling medium.

The trade-association sound standard for liquid-immersed transformers explicitly excludes dry-type from its own scope and names the separate standard that governs dry-type instead. The two technologies are tested to different standards with different reference conditions, so a single "X dB quieter" comparison is not supportable from citable material.

What is supportable is two separate, kVA-linked schedules. On the oil-immersed side, the liquid-immersed sound standard's published table gives a guaranteed average sound level of 57 dB at 700 kVA on unforced cooling, rising to 65 dB at 5,000 kVA and 68 dB at 10,000 kVA. On the dry-type side, a separate university design standard sets ceilings of 37 dBA at 9 kVA and below, rising in steps to 61 dBA between 751 and 1,000 kVA. These two schedules are reported side by side deliberately, rather than subtracted into one delta figure, because they come from different measurement bases.

The underlying physical mechanism is shared by both technologies rather than specific to either. Core noise arises from magnetostriction — the core steel's dimensional change under the alternating field — and from electromagnetic force concentrated at structural discontinuities, chiefly the air gaps between laminated steel sheets at joints, where a sudden change in permeability sharply raises local force density. A peer-reviewed study of this mechanism found the discontinuity-driven force to be the larger contributor of the two to core vibration. Since both technologies use a laminated steel core, this is a core-construction phenomenon, not something attributable only to the presence or absence of oil.

Core-type and shell-type: the other axis of "types of transformers"

Cooling medium is one axis of classification; core construction is a separate one, and the two are independent — a core-type or shell-type transformer can be built either dry or oil-filled. Buyers who conflate the two axes are asking two different questions at once without realising it.

Core-type

This is the construction generally chosen for high-voltage duty, because it is comparatively easy to insulate and to cool. In a core-type unit, the windings surround the core limbs, with roughly half of each winding — high-voltage and low-voltage — wound onto each leg.

Shell-type

Here the buyer is trading construction cost against ventilation. Shell-type reverses the arrangement: both windings are wound onto a single central leg first, and the laminated core is then assembled around them. Channelling the flux this way needs fewer turns for a given rating, which lowers construction cost, but the enclosed winding is harder to ventilate than a core-type design of comparable rating.

What the enquiry has to state

Nameplate marking rules set a floor for any transformer, either technology: rated kVA, frequency, and primary and secondary voltage. Above that floor, the requirement diverges by medium — a liquid-filled unit above a stated kVA threshold must show the type and quantity of its insulating liquid, while a dry-type unit above its own threshold must show the rated temperature rise of its insulation system instead.

The general-requirements standard for liquid-immersed transformers anticipates that a formal enquiry states more than a headline kVA figure, through a purchaser data sheet covering electrical and mechanical requirements beyond the base standard. The same expectation applies in practice on the dry-type side. It is also worth remembering that the same assembly usually carries the current transformers inside the same panel as a separate sizing decision from the power transformer itself.

A short checklist for the enquiry: rating, voltage class, insulation and temperature class, cooling designation, indoor or outdoor siting, and any local fire or containment requirement the buyer's jurisdiction adds on top of the base standard.

Comparison table

Questions buyers ask

Can a dry-type transformer go inside an occupied building without a vault?

Often yes. The US model electrical code allows reduced clearance from combustibles — roughly 1.8 m horizontally, 3.7 m vertically — instead of a fire-resistant room, once the unit uses Class 155 or Class 180 insulation. Below 112.5 kVA the clearance requirement is smaller still.

Does an oil-immersed transformer always trigger a spill-plan obligation?

Only once the facility's total oil storage crosses the federal rule's threshold; transformers are named as a covered class of oil-filled equipment. A dry-type installation has no equivalent obligation, because there is no oil inventory to contain.

Which technology handles overload better?

Oil-immersed units generally have more short-term overload headroom, reflecting oil's thermal mass and liquid cooling, but no citable source gives one universal multiplier — the governing standards are separate documents by technology, not a shared table.

Is a dry-type transformer quieter at the same rating?

Not automatically, and the two are tested to different standards, so a single "X dB quieter" figure is not supportable. Each technology has its own kVA-linked sound schedule from a different standards basis.

What is the practical difference between core-type and shell-type?

It is independent of cooling medium. Core-type is the usual choice for high-voltage duty, because it is comparatively easy to insulate and cool, and has roughly half of each winding wound onto each leg. Shell-type winds both windings onto a central leg first and assembles the laminated core around them afterwards — cheaper to build and more efficient in its use of flux, but harder to ventilate.