Total owning cost of a transformer: how to evaluate losses in a bid

In short: Total owning cost is the bid price plus an A factor in dollars per watt of guaranteed no-load loss plus a B factor in dollars per watt of guaranteed load loss at rated load. At the 14.53 cents per kWh EIA reported as average commercial revenue per kWh for July 2026, a ten-year evaluation at 8 percent gives A = $8.54/W and, at 0.45 per-unit load, B = $1.73/W — with a worked 500 kVA two-bid example, the break-even A factor that decides the award, and the loss budgets behind the DOE floor.

The cheapest transformer is rarely the least expensive to own. Two units meeting the same DOE efficiency floor can differ by thousands of dollars over their service life: below, the bid priced $3,400 higher wins by $1,208, and the break-even A factor that decides it is $6.30/W. The loss and rating data come off the same nameplate the transformer sizing calculator works from.

What does total owning cost actually add up?

A transformer stays energized for decades — EPA’s ENERGY STAR buying guide puts the average distribution transformer lifespan at 32 years — so even a fraction of a percent of throughput turned into heat accumulates into a recurring energy bill that never appears on the purchase order. The method putting price and losses on one axis is stated directly in that guide:

TOC = P + (A × NL) + (B × LL at 85 °C)

Two factors are needed instead of one efficiency figure because the components behave in opposite ways. No-load losses — core losses — run continuously whenever the unit is energized, 8,760 hours a year, at full load or none. Load losses are the I²R and stray losses in the windings and vary with the square of load current, so a unit at half load produces about a quarter of its rated load loss. A lightly loaded transformer is governed by core loss, a hard-worked one by winding loss.

How do you calculate the A factor and the B factor?

Both factors are computed, not assumed. The ENERGY STAR guide defines A and B but not how to build them, and IEEE C57.120-2017 holds the full treatment; the two formulas below are the energy-only simplification, stated here so every input is visible. Start with the present worth of an annual cost stream over the evaluation period:

PWF = [1 − (1 + i)^−n] ÷ i

The A factor capitalizes one watt running every hour of every year:

A = 8760 × EC × PWF ÷ 1000

The B factor carries the same economics but only materializes when current flows. It multiplies the load loss at rated load, so the scaling for expected loading belongs in the factor, not in the loss figure — B is the A factor times the square of the per-unit evaluation load, applied once:

B = A × L²

EPA’s anchor is that an A factor at or above $5/W corresponds to an average retail price at or above 9.7 cents per kWh, with about 22 percent of US utilities at or above that level — implying a present-worth factor near 5.9, shorter than the 6.710 here. Mind the inputs: that threshold rests on 2017 EIA data and the rates below on July 2026, and EIA publishes average revenue per kWh as a proxy rather than collecting rates, so for a ten-year levelization substitute a twelve-month figure for the supply actually contracted.

Two caveats. This is the energy-only form; IEEE C57.120-2017 also carries demand and escalation terms, so where demand charges are a large share of the bill the true A factor exceeds $8.54/W and the computed value is a floor. And reference temperatures differ by document: the guide’s TOC equation states load loss at 85 °C, while DOE’s appendix A to subpart K certifies liquid-immersed units at 20 °C no-load and 55 °C load loss. Fix one basis in the enquiry and correct every bid to it. The example below stays on the DOE basis, the one the efficiency floor is written on.

All rows use an 8 percent discount rate; the present-worth factors are 6.710, 8.559 and 9.818. Pair the result with a rating from the transformer sizing calculator.

A worked example: two bids on a 500 kVA pad-mounted transformer

A plant is fed by a three-phase liquid-immersed pad-mounted transformer at 500 kVA, averaging 0.45 per-unit load, buying energy at 14.53 ¢/kWh. Establish the minimally compliant design first — that is the baseline every bid is measured against.

1. Find the code floor. 10 CFR 431.196 requires 99.35 percent for a 500 kVA three-phase liquid-immersed unit, measured at 50 percent of nameplate load.
2. Convert efficiency to watts. Output at half load is 500 × 1000 × 0.5 = 250,000 W, so total loss = 250 000 × (1 ÷ 0.9935 − 1) = 250 000 × 0.006543 = 1,636 W.
3. Split the loss. The ENERGY STAR guide states that a minimum-DOE-compliant design carries LL = 4 × NL. At half load the load-loss contribution is 4 × NL × 0.5² = NL, so total loss = 2 × NL, giving NL = 818 W and LL = 3,272 W, both on the DOE certification basis of 20 °C and 55 °C. The two components contribute equally at the reference load.
4. Capitalize the low bid. Bid A is quoted at $42,000 with guaranteed losses of 820 W and 3,270 W. Capitalized no-load loss = 8.54 × 820 = $7,003; capitalized load loss = 1.73 × 3,270 = $5,657.
5. Total it. TOC = 42 000 + 7 003 + 5 657 = $54,660.
6. Capitalize the dearer bid. Bid B is quoted at $45,400 with a low-core-loss design at 250 W and 3,420 W. Capitalized no-load loss = 8.54 × 250 = $2,135; capitalized load loss = 1.73 × 3,420 = $5,917; TOC = 45 400 + 2 135 + 5 917 = $53,452.
7. Award it. Bid B costs $3,400 more and owns for $1,208 less.

The same answer arrives from the energy side, and that is the check worth running. Annual loss energy = 8760 × (NL + LL × L²) ÷ 1000, so Bid A burns 8760 × (820 + 3 270 × 0.2025) ÷ 1000 = 12,984 kWh a year against Bid B’s 8,257 kWh. That 4,727 kWh gap is worth 4 727 × 0.1453 = $687 a year, present worth 687 × 6.710 = $4,610; less the $3,400 premium, $1,210 remains — the $1,208 gap to within rounding, and a payback of 3 400 ÷ 687 = 4.9 years on a 32-year asset.

The same two bids at a lower energy price

Nothing about the transformers changes. Move the energy price to the 9.77 ¢/kWh industrial figure and the factors fall to A = $5.74/W and B = $1.16/W.

At the industrial rate the low bid wins by $302. Both awards are defensible, and the difference lies entirely in the buyer’s own inputs — which is why they belong in the enquiry.

Which bid wins, and at what A factor

One number settles it. Set the two total owning costs equal and solve for the A factor that exactly repays the premium:

A at break-even = ΔPrice ÷ (ΔNL − ΔLL × L²)

Above $6.30/W the low-core-loss bid wins; below it the low bid does. The $8.54/W and $5.74/W cases sit either side — one line for an award memo, instead of a table of scenarios.

What does 10 CFR 431.196 require, and what changes in 2029?

The DOE minimum is a floor, and a minimally compliant design is optimized to just meet it at the reference load — 50 percent for liquid-immersed units. Realized efficiency falls away either side of that point, which is why a unit optimized for its actual load factor wins.

Loss columns come from the in-force efficiency at 50 percent load, taking the no-load figure as half the total and the load loss as four times that. Read them as what a minimum-compliant design looks like, not as separate legal limits: the standard fixes only total loss at the reference load, and a compliant unit may shift the split — which is the whole point here.

Two features are worth reading twice. The 2029 increase is steep at and below 300 kVA, 0.15 to 0.17 points, and nearly flat above it, 0.02 to 0.03 points. And the ordering inverts: from April 23, 2029 a 300 kVA unit must reach 99.42 percent while a 500 kVA unit needs only 99.38 percent. That is deliberate. DOE splits three-phase liquid-immersed units into equipment class 2A, “ranging from 15 to less than 500 kVA”, and class 2B at 500 kVA and above, setting them at different efficiency levels so units “more likely to have high currents” keep “additional flexibility in meeting efficiency standards”. The boundary sits at exactly 500 kVA — and the order inverts again at the top, 2500 kVA at 99.55 percent, 3750 at 99.54, 5000 at 99.53. Budget loss per rating, not per family. The rules themselves are covered in transformer efficiency standards and the 2029 update.

Where Canada differs

Canada regulates the same equipment through its own instrument. Canada’s Energy Efficiency Regulations cover a dry-type transformer defined as single-phase 15 to 833 kVA or three-phase 15 to 7500 kVA, 60 Hz, with a high-voltage winding of 35 kV or less — then exclude thirteen types, among them auto, drive, grounding, rectifier, sealed and welding transformers, so check the exclusions before assuming a unit is in scope. The federal requirement is NRCan’s own energy-performance table, tested to appendix A of 10 CFR subpart K rather than a Canadian test method. The CSA minimum-efficiency standards for these classes are C802.1, C802.2 and C802.3.

Two divergences follow. The reference load is not common across classes: 10 CFR 431.196 fixes low-voltage dry-type efficiency at 35 percent of nameplate load and liquid-immersed at 50 percent, so the two tables describe different operating points. And the timelines are not aligned: the published Canadian dry-type requirement is the level for units built on or after January 1, 2016, while the amended US levels bind from April 23, 2029. A design qualified only to the Canadian table will not meet the US standard for units manufactured on or after that date. The rule binds date of manufacture, not of sale, so earlier units stay lawfully saleable — but it is still a real constraint on one drawing serving both markets. Only the energy price and the floor change across the border; the arithmetic does not.

What to specify

• Guaranteed no-load and load losses in watts, not an efficiency percentage, tested per IEEE C57.12.90-2021 and rated to IEEE C57.12.00 for liquid units or IEEE C57.12.01 and NEMA ST 20 for dry-type. Without watts there is nothing to capitalize.
• The A factor and the B factor, published in the enquiry. They come from the buyer’s energy price, discount rate and evaluation period, and telling bidders how loss is scored lets them optimize rather than guess the target.
• The per-unit evaluation load, from a load study or interval meter data rather than the nameplate. The B factor is meaningless without it, and it moves the award.
• The governing efficiency floor and the build date — 10 CFR 431.196 as it stands, or the amended levels for units built on or after April 23, 2029; in Canada, the Energy Efficiency Regulations table.
• A loss-penalty clause tying dollars per watt above the guarantee, so the awarded figures are backed by the test report.
• One stated temperature basis. DOE certifies liquid-immersed units at 20 °C no-load and 55 °C load loss; the ENERGY STAR TOC equation states 85 °C load loss. Name the basis and the waveform correction in the enquiry, because uncorrected figures are not comparable between bidders.

Common mistakes

• Comparing efficiency percentages across classes. A 112.5 kVA dry-type unit at 98.74 percent is not worse than a liquid unit at 99.11 percent — the first is measured at 35 percent load, the second at 50. Compare watts, or the right column of 10 CFR 431.196.
• Using the average load factor as the B-factor load. Load loss varies with the square of current, and the mean of a square is always at least the square of the mean, so an average understates load loss on any varying profile. Use RMS-equivalent loading; IEEE C57.120-2017 covers the loss-factor treatment.
• Treating the DOE floor as the design target. A minimum-compliant 500 kVA unit still burns 1,636 W at half load and 12,984 kWh a year at 0.45 per-unit load. EPA’s criteria cut losses 11 to 29 percent while remaining compliant, at a first-cost premium EPA puts at no more than about 10 percent.
• Evaluating over the wrong horizon. A ten-year present-worth factor of 6.710 against a 32-year asset understates the A factor by about 40 percent versus the 11.435 of a full-life evaluation. State the horizon and why.

Where Entogo fits

Loss evaluation only rewards a supplier that can hold guaranteed loss figures across a production run and prove them on a test report. Entogo builds transformers and substation equipment in its own vertically integrated factory, designed and built to IEEE C57.12.00, C57.12.34 and C57.12.90, UL 1561/1562, CSA C88 and CSA C802.2, and DOE 10 CFR 431; UL (cULus)/CSA certifiable on request. Guaranteed no-load and load losses are quoted per unit, and the factory test report is issued with the equipment. Where a high A factor justifies it, low-core-loss constructions are available — a wound-core distribution transformer in its oil-filled pad-mounted variant, or an amorphous-core dry-type transformer where the application is dry-type. Both carry lower no-load loss than an equivalent stacked grain-oriented silicon-steel core, which is what pays back on a continuously energized utility network.

Because the same factory builds the transformer and the substation and power-distribution assembly around it, the guaranteed losses and the ventilation and loading assumptions behind them stay consistent — NEC 450.9 ventilation and the heat a lower-loss unit never rejects are one calculation. Size the rating in the transformer sizing calculator, check the protection side in sizing transformer overcurrent protection, and request a quote with a loading profile attached.

Loss evaluation is not a judgement call dressed up in arithmetic. Publish the A factor, the B factor and the per-unit load, require guaranteed watts on one stated temperature basis, compute the break-even A factor once, and the award follows from numbers every bidder can see. Get the rating right first — kVA selection and dry-type versus liquid-filled precede loss evaluation — then let the two factors decide which compliant bid is the cheap one.