Transformer Loss Evaluation in EPC Bids: Capitalization, Test Basis and Award Risk

A transformer bid can look cheaper because its purchase price is lower, while the lifetime cost tells a different story. For an EPC team comparing technically compliant offers, transformer loss evaluation is a way to make the trade-off visible. It only works, however, when the guaranteed losses, test basis, evaluation factors and measurement uncertainty are defined before bid opening.

What loss evaluation is meant to compare

Transformer losses are energy dissipated during operation. No-load loss is associated mainly with the energized magnetic core and is present whenever the unit is energized, even at light load. Load loss is principally associated with winding and stray losses and varies with current; the exact evaluation basis follows the applicable test method and contract. They affect different operating hours, so using one blended number can conceal a material difference between two offers.

For procurement, the comparison should usually distinguish:

• Guaranteed no-load loss at the specified voltage, frequency and tap condition.
• Guaranteed load loss at the stated reference temperature and rated current or agreed test basis.
• Auxiliary consumption, where it is separately included in the project’s energy model.
• Purchase price and the contractual treatment of loss deviations.

The evaluation factors are commercial inputs, not universal IEC constants. The owner or EPC should establish their units, currency, tax basis, escalation assumptions, expected loading profile and evaluation horizon. The bid form should say whether the factors represent a capitalized value per kilowatt or another agreed method. Do not let bidders silently apply different assumptions.

Set a common calculation basis before tender

A transparent comparison can be expressed as:

Evaluated cost = bid price + A × guaranteed no-load loss + B × guaranteed load loss + other agreed adjustments

Here A and B are project-defined capitalization factors, expressed so the terms resolve to the same currency as the bid. The equation is a procurement model, not a standard-prescribed tariff. If the project instead uses an hourly energy model, calculate the no-load component using energized hours and the load component using an agreed load curve or loss factor. Document the model and keep it identical across offers.

Freeze the technical definition, not just the headline number

Two guarantees are comparable only if their boundaries match. Confirm rated power, winding configuration, voltage ratio, frequency, cooling arrangement, tap range and the specific loss definitions. Confirm whether the quoted figures are guaranteed maxima or typical design values. If parallel units, special duty cycles, harmonic currents or frequent reverse power are expected, make those service conditions explicit; they can influence design and should not be introduced after award as an informal assumption.

The same principle applies to tolerances. A standard may establish test and tolerance rules, while the purchase specification and contract determine which guaranteed values and remedies apply. State the governing edition and order of precedence. For the document workflow behind a controlled technical freeze, see transformer datasheet approval for EPC projects. The contract should also align the guarantee schedule with the FAT witness plan.

How to handle measurement uncertainty

Loss results are measured, not exact mathematical constants. Instrument transformers, voltage and current measurement, phase-angle accuracy, wiring, burden, temperature correction and test-system configuration contribute to uncertainty. IEC 60076-19-1:2023 describes procedures for evaluating uncertainty in no-load and load-loss measurements during routine tests, including known-error corrections. That is why the EPC should ask for the calculation method and traceability rather than comparing rounded report values alone.

Do not create an acceptance threshold by adding an arbitrary percentage to a reported value. The contract must define whether a guarantee is assessed against a measured result, a corrected result, an uncertainty statement, or a specific contractual tolerance. The test report should preserve raw readings, correction factors, instrument identification and the calculation trail needed to reproduce the result.

Use the award comparison and FAT as one chain

The bid evaluation is only useful if its assumptions survive into manufacturing and factory acceptance. Carry the accepted guarantees into the purchase order, approved datasheet, routine-test procedure and final report. Before the witness date, agree who supplies test equipment, which taps and connections are used, how stabilization is judged, and which documents the purchaser receives. If a deviation occurs, record it against the guarantee and contract clause before deciding acceptance; do not treat a rerun or corrected value as self-explanatory.

For a suitable product context, review the 2500 kVA oil-immersed transformer and confirm the specific project’s rating and loss guarantees with the manufacturer. Product-page examples do not replace a project datasheet.

Worked example: how project factors can change the award ranking

Consider a deliberately simplified, hypothetical two-offer comparison. Offer Alpha is 12,000 currency units cheaper, but its guaranteed no-load loss is 0.4 kW higher and its load loss is 1.5 kW higher than Offer Beta. With illustrative project factors of 2,000 currency units per kW for no-load loss and 500 per kW for load loss, the evaluated additions are 800 + 750 = 1,550 for Alpha and zero for Beta. Alpha remains lower by 10,450 after the loss adjustment.

If the owner’s approved factors instead assign 25,000 per kW to no-load loss and 2,000 per kW to load loss, Alpha’s additional evaluated loss cost becomes 10,000 + 3,000 = 13,000. Its 12,000 purchase-price advantage is then more than offset, so Beta ranks lower by 1,000 under this simplified model. The values are illustrative only; they are not market benchmarks, equipment guarantees, or a forecast for a Zisheng Electric project.

This sensitivity is the point of the exercise. The EPC should publish the factors and units before tender, not select them after seeing the bids. If factors represent a whole-life present value, document the evaluation period, discount and escalation assumptions. If an hourly energy model is used instead, separate continuously energized no-load energy from load-dependent energy. For a simplified load curve, the load-loss energy term is proportional to the sum of each interval’s duration multiplied by the square of its per-unit current, subject to the agreed loss definition and test basis. Do not substitute one average load percentage for a varying duty profile unless the owner’s method explicitly permits it.

Turn the loss schedule into a contract and test record

Before award, reconcile the commercial workbook with the technical offer line by line. Each guaranteed value should point to a transformer identification, rating, winding and tap basis, cooling designation, reference temperature, voltage and frequency condition, and governing standard edition. Identify excluded auxiliary loads separately. Where bidders use a different calculation model, request a normalized schedule instead of trying to compare narrative claims.

At purchase-order stage, carry the accepted numbers and evaluation assumptions into the guarantee schedule. At FAT planning, agree the routine-test sequence, measurement configuration, correction and uncertainty reporting, witness hold points, and the document format for raw and corrected results. Record any retest as a separate traceable event with a reason and the relevant contract clause. A rerun does not erase the first result; both records help the purchaser understand the disposition.

• Keep bid evaluation factors distinct from technical acceptance tolerances.
• Require the supplier to identify missing or ambiguous reference conditions before design freeze.
• Make any commercial deviation remedy explicit in the contract rather than inferring it from a test standard.
• Close the FAT dossier only after the report, guarantee schedule and approved datasheet agree.

EPC bid review checklist

• Issue one common loss-evaluation schedule and one set of capitalization factors.
• Define guaranteed values, reference conditions, taps, correction basis and governing standard edition.
• Require each bidder to separate no-load, load and auxiliary losses.
• State whether service profile or annual operating hours are used for economic evaluation.
• Carry guarantees and remedies into the purchase order and approved FAT procedure.
• Require a traceable result with measurement uncertainty and known-error corrections stated.

Transformer loss evaluation is most defensible when engineering and commercial teams agree on the same boundaries before they rank bids. Send Zisheng Electric the datasheet, load profile, evaluation factors and draft guarantee schedule for a project-specific review; the engineering team can identify missing assumptions before they become award or FAT disputes.