What Happens When a Transformer Is Loaded Past Its Rating

Why this matters

A transformer that is overloaded does not trip, does not alarm, and does not fail on the day it happens. It quietly spends insulation life, holds a lower secondary voltage than the equipment downstream expects, and fails months or years later in a way that gets written up as a bad transformer. Meanwhile the same percentage overload, at a different duration and a different room temperature, is genuinely harmless. Knowing which of those two you are looking at is the difference between a real finding and a scare.

The rating is a heat budget, not a trip point

Nothing inside a transformer enforces its kVA. The rating says that at a stated ambient, drawing rated current continuously, the windings will settle at a stated rise and the insulation system will live out its design life. Every term in that sentence can move.

For a common dry-type construction the three terms add like this: a 220 degree C insulation system is built from a 40 degree C maximum ambient, a 150 degree C average winding rise, and a 30 degree C allowance for the hottest spot in the winding running above that average. That breakdown, standard for ventilated dry types, is the single most useful thing to hold about transformer loading, because it shows immediately that ambient and load are interchangeable ways of spending the same budget. Ten degrees C of extra room temperature costs exactly what ten degrees C of extra winding rise costs.

Two more relationships fill in the arithmetic:

• Winding rise scales with roughly the square of the current. Load losses dominate at and above about half load and go with the square of current; core loss is essentially constant. So treating rise as proportional to the square of current is a good approximation near rated load and a pessimistic one well below it.
• Insulation aging is roughly exponential in temperature. The commonly cited approximation is that expected insulation life halves for something in the range of every 8 to 10 degrees C of sustained operation above the design point. The exact figure is specific to the insulation system, and the manufacturer's loading data owns it.

And one that does most of the work in the field: a transformer's thermal time constant is long. For a typical ventilated dry type it is on the order of one to three hours, which means a winding responds to the average load over hours, not to the peak in any given minute. The actual value for a specific unit is in the manufacturer's loading data; the illustrative figure used below is two hours.

The one gate

A transformer is overloaded in the way that matters when the sum of its actual inlet ambient plus its rise at the load it is actually carrying plus its hot-spot allowance exceeds the insulation system's design temperature, sustained over a period comparable to its thermal time constant. Every clause in that sentence is load-bearing. Percentage of rated current alone does not decide it, and neither does a peak reading.

Two overloads that resolve in opposite directions

Both cases use the same unit: a ventilated dry type with a 220 degree C insulation system built as 40 ambient plus 150 rise plus 30 hot spot, and an assumed two-hour thermal time constant. All values are illustrative of a common configuration rather than a specification for any product.

Case A: 120 percent of rated current for 20 minutes, in a 25 degree C room.

The unit runs at about 70 percent of rated current for most of the day, so its settled rise is 150 times 0.70 squared, which is 150 x 0.49 = 73.5 degrees C. At the shift-start peak of 120 percent the rise it is heading toward is 150 times 1.20 squared, which is 150 x 1.44 = 216 degrees C. That target is alarming and it is also irrelevant, because the winding only travels part of the way there.

Twenty minutes against a two-hour time constant is one sixth of a time constant, and an exponential approach covers 1 minus e to the power of minus 0.167, which is 15.4 percent of the remaining gap. So the rise reached is 73.5 plus 0.154 times (216 minus 73.5), which is 73.5 + 21.9 = 95.4 degrees C.

Total winding temperature at the hot spot: 25 ambient + 95.4 rise + 30 hot spot = 150.4 degrees C, against a 220 degree C system. That is roughly 70 degrees C of margin remaining at the worst moment of the day. Case A is not an overload in any sense the insulation notices, and telling the owner otherwise on the strength of a 120 percent clamp reading is a false alarm.

Case B: 105 percent of rated current continuously, in a 50 degree C closed room.

The percentage is far gentler and the verdict reverses. Rise at 105 percent is 150 times 1.05 squared, which is 150 x 1.1025 = 165.4 degrees C, and because the condition is continuous the winding actually gets there.

Total at the hot spot: 50 ambient + 165.4 rise + 30 hot spot = 245.4 degrees C, against the same 220 degree C system. That is 25.4 degrees C over, and it holds all day.

Run that through the aging approximation stated above and the two ends of the cited range give the honest spread: at 10 degrees C per halving, 25.4 degrees is about 2.5 halvings, so expected insulation life falls by a factor of roughly five to six; at 8 degrees C per halving it is about 3.2 halvings, nearer a factor of nine. Either way the unit is spending years per year, and it will keep doing so silently.

Notice where the damage came from. Of the 25.4 degrees C of overage, 15.4 degrees is the extra rise the 5 percent overload bought (165.4 against 150 at rated current) and the remaining 10 degrees is simply the room being 10 degrees C above the 40 degree C ambient the rating assumes. The room contributed nearly as much as the electrical overload did, and it is usually the cheaper of the two to fix, which is why the inlet air temperature is a reading and not an afterthought.

Why the protective devices will not tell you

The overcurrent devices around a transformer are not watching its insulation. Under Article 450 of the NEC as adopted in the edition in force in your jurisdiction, transformer overcurrent protection is set at percentages of rated current that deliberately allow substantial overload so that inrush and normal load swings do not cause nuisance operation, and under Article 240 as adopted the devices on the conductors are protecting those conductors. Nothing in that arrangement opens because a winding is at 245 degrees C. Case B will run for years without a single operation of anything.

That is why an overloaded transformer is found by measurement rather than reported by the system. A tech who waits for something to trip will meet this unit at the failure, not before it.

What it looks like in the room, at both ends of the range

At the mild end, a transformer inside its budget gives you nothing to find: a warm enclosure, an ordinary hum, and a secondary that sags by a predictable fraction of the nameplate impedance under load. There is no smell and no discoloration, and the absence of those cues is not evidence of light loading.

At the far end, the cues arrive in a rough order: an audible hum that rises with load, an enclosure noticeably hotter than the room, a varnish or hot-cloth smell that is strongest at the top vents where the exhaust air leaves, then darkening of the coil surfaces and eventually the insulation failure itself, most often turn to turn inside a coil rather than a dramatic external event. Do not touch the enclosure to judge temperature. A dry-type enclosure at full rated load in a warm room can be hot enough to burn skin on contact; take the surface with a non-contact infrared thermometer from a normal standing distance, and take the air temperature at the ventilation inlet with a probe held clear of the enclosure.

What to measure, and where

Clamp the secondary conductors over a loaded period, not at a moment. What the insulation responds to is the average current over hours, so a single peak reading is not the input to the gate above. Where the load cycles, log for a period at least as long as the assumed time constant, and use the average of that window against rated current. A clamp meter with a min/max/average function or a simple logger both work; the peak alone is the number that produces false alarms of exactly the Case A kind.

Take the ambient at the inlet, not in the middle of the room. The number in the gate is the air the transformer is actually breathing. A unit in a closed electrical closet, or one mounted high in a room with stratified air, or one whose inlet is a foot from another heat source, is in a different ambient than the room thermostat reports, and that difference is frequently the entire finding.

Check the ventilation path is the one the unit was designed with. Blocked bottom vents, stacked storage against an enclosure, and a removed or reversed room exhaust fan all show up as ambient. Restoring an obstructed factory ventilation path is maintenance. Adding forced ventilation, cutting new openings, or changing the enclosure is a modification to a listed assembly, and whether it is permitted is settled by the manufacturer and the authority having jurisdiction rather than by a technician on site. Never operate a transformer with its enclosure panels removed to improve cooling: that converts a closed assembly into exposed energized parts, and the shock and arc-flash exposure it creates is governed by 29 CFR 1910.333(a)(1), which permits energized exposure only where the employer can demonstrate that de-energizing introduces additional or increased hazards or is infeasible.

Every current and voltage reading above is taken with the equipment energized and running, which is that same 1910.333(a)(1) case, because load is the condition being measured; establish boundaries and select PPE on the basis in NFPA 70E-2021, 130.5 and 130.7, binding through your employer's electrical safety program or a contract rather than on its own, and use a meter and leads rated CAT III or better at or above the voltage present. Before any work on the unit itself, open and lock the primary disconnect under 29 CFR 1910.333(b)(2) and prove dead with the live-dead-live sequence at NFPA 70E-2021, 120.5, remembering that 29 CFR 1910.147 excludes this electrical exposure at (a)(1)(ii)(C) and that 29 CFR 1926.417 is the construction counterpart for lockout and tagging of circuits.

References

• 29 CFR 1910.333(a)(1) and (b)(2) for energized work and de-energizing; 29 CFR 1926.417 for lockout and tagging of circuits in construction
• NFPA 70E-2021, 120.5, 130.5 and 130.7, applied through an employer electrical safety program or contract
• NEC Articles 240 and 450 as adopted, in the edition in force in your jurisdiction, for conductor and transformer overcurrent protection
• Manufacturer loading data for the specific unit, which owns its thermal time constant, its insulation system and any question about modifying a listed enclosure
• See related: What a Transformer Nameplate Is Telling You; What a Secondary Voltage Depends On