Secondary-side nonlinear loads can overheat a transformer even when a primary-side survey reports no problem. Harmonic current usually transfers through the transformer according to its ratio, but winding connections, grounding, source stiffness, instrument setup, and the measured quantity can make the primary result look much cleaner. Diagnose this installation with simultaneous current, voltage, loading, temperature, and dissolved-gas measurements on both sides; no universal transfer percentage applies.
Symptom Interpretation
The installation uses a 2500 kVA silicone-fluid transformer with a 24.94/14.4 kV primary, a 480/277 V secondary, and a stated Y-Y connection. Reported loading is 70% to 80%, oil temperature is 80°C at an 18°C ambient, and dissolved-gas analysis found excessive H2, CH4, and C2H6. The transformer was replaced three times over four years; the earlier units were remanufactured, while the latest unit was new and came from a different manufacturer.
The measured oil-to-ambient temperature difference is 62°C. Do not directly compare that number with the stated 55/65°C transformer rise rating until the nameplate or test report identifies whether the rating applies to average winding rise, hottest-spot rise, top-oil rise, or another defined temperature. Oil temperature and winding temperature are not interchangeable.
If the 70% to 80% figure represents balanced three-phase kVA, rated secondary current is approximately 3007 A:
I = 2500 kVA / (sqrt(3) × 0.480 kV) = 3007 A
The corresponding load range is about 2105 to 2406 A. At 24.94 kV line-to-line, rated primary current is approximately 57.9 A, with a 70% to 80% range of about 40.5 to 46.3 A. Confirm how the reported loading was calculated because fundamental-only meters can understate total RMS current when distortion is high.
Harmonic Transfer Mechanism
A nonlinear load draws current at the fundamental frequency plus integer multiples called harmonics. Those secondary harmonic ampere-turns require balancing primary ampere-turns, so harmonic current generally appears on the primary after transformation by the turns ratio. Transformer leakage impedance and system impedance modify each harmonic, while connection and grounding determine whether a sequence component has a complete path.
For a Y-Y transformer, triplen harmonics such as the third, ninth, and fifteenth are zero-sequence components. Their transfer depends on neutral connections, grounding, and any tertiary winding. Read the nameplate and drawings to establish those paths. A delta tertiary, if installed, can circulate zero-sequence current internally; primary phase loss or primary voltage unbalance can also produce severe circulating current and heating. Its presence and kVA rating must be verified rather than inferred from the main winding designation.
Primary voltage distortion is not a substitute for primary current distortion. A stiff utility source can hold voltage close to sinusoidal while supplying distorted current. Harmonic current may also become diluted at a monitoring point that includes other loads. Even current total harmonic distortion can differ across the transformer because THD is a ratio to the local fundamental current, not a fixed percentage that the transformer must reproduce.
Harmonics add RMS current and increase winding eddy-current and structural stray losses. Higher-frequency components can therefore produce disproportionate heating without exceeding the nameplate kVA calculated from an incomplete measurement.
Coordinated Measurement Procedure
- Record the complete nameplate data, cooling class, temperature-rise definitions, tap position, Y-Y neutral and grounding arrangement, and any tertiary winding. Check actual conductor connections against the one-line diagram.
- Install suitable power-quality instruments on the primary and secondary for the same operating interval. Capture all three phase currents and voltages, neutral current where accessible, total RMS, fundamental RMS, harmonic spectrum, kW, kVAR, kVA, displacement power factor, true power factor, and voltage unbalance.
- Trend transformer oil temperature, ambient temperature, load, and cooling operation against time. Inspect airflow and heat-transfer surfaces; verify fans or pumps only if the installed cooling system includes them.
- Validate CT ratios, polarity, phase association, bandwidth, burden, and saturation margin. Confirm that each analyzer records current harmonics rather than only voltage disturbances or averaged RMS values.
- Compare matching harmonic orders on both sides after converting current by the transformer ratio. Investigate omitted zero-sequence components separately because their path depends on the verified grounding and winding arrangement.
- Repeat dissolved-gas sampling under controlled conditions and trend individual concentrations and generation rates. The reported analysis used
ASTM D3612; apply interpretation criteria appropriate to silicone fluid rather than substituting mineral-oil limits.
Root-Cause Decisions and Corrections
| Finding | Mechanism | Action |
|---|---|---|
| High secondary current distortion with corresponding primary orders | Nonlinear load current is transferring through the transformer | Identify the producing loads, quantify each contribution, and evaluate load reduction, redistribution, or a properly engineered harmonic-mitigation method. |
| High secondary triplen current but weak primary triplen current | The Y-Y neutral, grounding, or tertiary path is changing zero-sequence transfer | Measure neutral current and verify the actual winding and grounding connections before interpreting the primary survey. |
| Low distortion but high RMS current or low true power factor | Loading, reactive current, or phase imbalance is driving losses | Correct the measured cause. Model harmonic resonance before adding power-factor capacitors. |
| Normal electrical loading with excessive temperature | Cooling restriction, temperature indication error, internal loss, or transformer design condition remains | Verify sensors and cooling performance, then compare measured losses and temperature rise with manufacturer test data. |
| Gas generation continues as temperature or load changes | Active thermal or electrical degradation remains inside the unit | Use silicone-fluid DGA trends with electrical and thermal data to determine whether overheating, insulation breakdown, or partial discharge is active. |
Recurrence across remanufactured units and a new unit from another manufacturer puts common installation factors high on the decision list: load waveform, voltage unbalance, grounding, cooling environment, protection operation, and measurement quality. It does not by itself identify harmonics or exclude a transformer design problem.
Numbered Verification Checks
- Check 1: Repeat simultaneous primary and secondary recording during the same production cycle. Expect matching non-zero-sequence harmonic orders after ratio conversion, within differences caused by system impedance and measurement accuracy.
- Check 2: Recalculate loading from total RMS phase current and measured three-phase kVA. Expect each phase and the transformer total to remain within the applicable nameplate rating.
- Check 3: Compare oil temperature minus ambient against load over time. Expect temperature to stabilize for a stable load and cooling condition; compare the correct measured temperature with the matching manufacturer rise definition.
- Check 4: Trend H2, CH4, and C2H6 after corrective work. Expect gas generation rates to decline or stabilize; a continuing rise calls for internal fault assessment.
Recurring Diagnostic Pitfalls
A “no problem” primary report has little diagnostic value without channel lists, instrument configuration, measurement location, interval, load state, and harmonic spectra. Accepting voltage-only monitoring as proof of clean current is wrong practice.
Do not treat the 80°C oil indication as a winding temperature, or compare it blindly with a 55/65°C rise marking. Do not interpret silicone-fluid DGA using mineral-oil criteria. Avoid adding capacitors solely to raise power factor until a frequency-domain study checks for resonance with the transformer and system impedance.
Average readings can hide production-cycle peaks, phase loss, voltage unbalance, cooling failures, and intermittent harmonic loading. Correlate electrical, thermal, and gas data on a common time base.
Frequently Asked Questions
Why does the primary show no harmonic problem when the transformer overheats?
The survey may have measured voltage rather than current, monitored upstream where other loads diluted current THD, or omitted the relevant operating period. Measure simultaneous primary and secondary current spectra under the same load.
Why does a Y-Y transformer show different harmonic levels on each side?
System impedance changes each harmonic, while triplen transfer depends on the neutral, grounding, and any tertiary winding. THD percentages can also differ because each side uses its local fundamental current as the denominator.
How do I verify that corrective work stopped the overheating?
Repeat the same load-cycle measurements and expect total RMS current and kVA to remain within nameplate limits, the correct temperature rise to stabilize within its manufacturer criterion, and the H2, CH4, and C2H6 generation rates to decline or stabilize.