The Transformer Passed Impedance Test. The Cooling Failed at 80% Load.
Quote from chief_editor on April 8, 2026, 7:06 pmPower transformer buyers verify impedance and turns ratio at factory acceptance. Thermal performance under sustained load — the test that predicts field failure — is rarely included in FAT scope.
The mining substation transformer in Western Australia — 20 MVA, 132/11 kV, ONAN cooling, from a Hebei manufacturer — had been in service for nine months when the monitoring system started logging temperature rise alarms on the top oil thermometer at loads above 14 MW. The transformer's nameplate rating was 20 MVA. The temperature rise alarm was activating at 70% of nameplate load.
The factory acceptance test had verified turns ratio, impedance voltage, no-load losses, and load losses at rated current. It had not included a temperature rise test — a test that subjects the transformer to rated load for sufficient duration to reach thermal equilibrium and measures whether the temperature rise at rated load is within the IEC or IEEE specification limit. Temperature rise testing is optional in many FAT scopes because it is expensive and time-consuming: it requires the manufacturer to have load-test equipment capable of supplying rated current for up to 8 hours, which not all Chinese transformer facilities have.
The Hebei manufacturer had a factory acceptance test certificate that covered every test in the agreed FAT scope. The temperature rise test was not in the agreed FAT scope because the buyer's procurement engineer had used a template FAT checklist that did not include temperature rise as a standard test for this voltage class. The cooling system was undersized. Nobody had tested it.
A Transformer Nameplate Rating Is a Manufacturer's Claim, Not a Measured Result
Transformer nameplate ratings — the MVA figure, the voltage ratios, the cooling class — are the manufacturer's statement of what the transformer will do under specified conditions. For the nameplate to be verified, the transformer has to be tested under those conditions. Temperature rise testing under rated load for sufficient duration to reach thermal equilibrium is the test that verifies the cooling system's ability to maintain the rated performance. Without that test, the nameplate is an assertion.
In Chinese transformer manufacturing, the thermal design is the area where cost optimization most directly affects performance. The radiator count, the cooling fin area, the oil circulation design, and the insulation thermal class are all design decisions that can be reduced in cost without changing the electrical performance parameters measured in most FAT protocols. A transformer with insufficient cooling will pass impedance testing, turns ratio testing, and loss measurement testing without showing any deficiency — those tests are conducted briefly and do not allow the thermal condition to develop.
The temperature rise at 70% load — not at 100% load — indicated that the Hebei transformer's thermal time constant was such that the top oil was reaching approximately 95°C under sustained load conditions, against an IEC limit of 60°C top oil rise plus ambient. At the site's summer ambient of 38°C, the effective limit was 98°C. The transformer was operating near its thermal limit at 70% load, with essentially no capacity margin for peak demand periods.
The Solution Was External Cooling. The Cost Was Continuous.
The mine's power engineering team solved the immediate operational problem by installing supplementary forced air cooling on the existing radiator bank — a modification that cost $85,000 and required the transformer to be de-energized for three days during installation. The modification brought the thermal performance within acceptable limits at nameplate load, with reduced margin.
The longer-term concern — transformer insulation life, which is exponentially dependent on operating temperature — meant that the operations team had committed to a transformer that would require early replacement relative to a correctly-designed unit. The actuarial cost of shortened insulation life was not calculable precisely, but the engineering team estimated a useful life reduction from the expected 30 years to approximately 18 to 22 years.
A 20 MVA transformer that passes everything except the test you did not run is a transformer that passes everything except the test that matters.
Keywords: China power transformer procurement quality | Chinese transformer manufacturer cooling, transformer FAT thermal test, power transformer China procurement, mining substation transformer China
Words: 632 | Source: Documented transformer thermal failure — mining substation, Western Australia, 2023. Hebei manufacturer FAT documentation, temperature monitoring data, external cooling modification records. | Created: 2025-01-15T10:25:00Z
Power transformer buyers verify impedance and turns ratio at factory acceptance. Thermal performance under sustained load — the test that predicts field failure — is rarely included in FAT scope.
The mining substation transformer in Western Australia — 20 MVA, 132/11 kV, ONAN cooling, from a Hebei manufacturer — had been in service for nine months when the monitoring system started logging temperature rise alarms on the top oil thermometer at loads above 14 MW. The transformer's nameplate rating was 20 MVA. The temperature rise alarm was activating at 70% of nameplate load.
The factory acceptance test had verified turns ratio, impedance voltage, no-load losses, and load losses at rated current. It had not included a temperature rise test — a test that subjects the transformer to rated load for sufficient duration to reach thermal equilibrium and measures whether the temperature rise at rated load is within the IEC or IEEE specification limit. Temperature rise testing is optional in many FAT scopes because it is expensive and time-consuming: it requires the manufacturer to have load-test equipment capable of supplying rated current for up to 8 hours, which not all Chinese transformer facilities have.
The Hebei manufacturer had a factory acceptance test certificate that covered every test in the agreed FAT scope. The temperature rise test was not in the agreed FAT scope because the buyer's procurement engineer had used a template FAT checklist that did not include temperature rise as a standard test for this voltage class. The cooling system was undersized. Nobody had tested it.
A Transformer Nameplate Rating Is a Manufacturer's Claim, Not a Measured Result
Transformer nameplate ratings — the MVA figure, the voltage ratios, the cooling class — are the manufacturer's statement of what the transformer will do under specified conditions. For the nameplate to be verified, the transformer has to be tested under those conditions. Temperature rise testing under rated load for sufficient duration to reach thermal equilibrium is the test that verifies the cooling system's ability to maintain the rated performance. Without that test, the nameplate is an assertion.
In Chinese transformer manufacturing, the thermal design is the area where cost optimization most directly affects performance. The radiator count, the cooling fin area, the oil circulation design, and the insulation thermal class are all design decisions that can be reduced in cost without changing the electrical performance parameters measured in most FAT protocols. A transformer with insufficient cooling will pass impedance testing, turns ratio testing, and loss measurement testing without showing any deficiency — those tests are conducted briefly and do not allow the thermal condition to develop.
The temperature rise at 70% load — not at 100% load — indicated that the Hebei transformer's thermal time constant was such that the top oil was reaching approximately 95°C under sustained load conditions, against an IEC limit of 60°C top oil rise plus ambient. At the site's summer ambient of 38°C, the effective limit was 98°C. The transformer was operating near its thermal limit at 70% load, with essentially no capacity margin for peak demand periods.
The Solution Was External Cooling. The Cost Was Continuous.
The mine's power engineering team solved the immediate operational problem by installing supplementary forced air cooling on the existing radiator bank — a modification that cost $85,000 and required the transformer to be de-energized for three days during installation. The modification brought the thermal performance within acceptable limits at nameplate load, with reduced margin.
The longer-term concern — transformer insulation life, which is exponentially dependent on operating temperature — meant that the operations team had committed to a transformer that would require early replacement relative to a correctly-designed unit. The actuarial cost of shortened insulation life was not calculable precisely, but the engineering team estimated a useful life reduction from the expected 30 years to approximately 18 to 22 years.
A 20 MVA transformer that passes everything except the test you did not run is a transformer that passes everything except the test that matters.
Keywords: China power transformer procurement quality | Chinese transformer manufacturer cooling, transformer FAT thermal test, power transformer China procurement, mining substation transformer China
Words: 632 | Source: Documented transformer thermal failure — mining substation, Western Australia, 2023. Hebei manufacturer FAT documentation, temperature monitoring data, external cooling modification records. | Created: 2025-01-15T10:25:00Z