The measured harmonic current increased from 11 A to 72 A for a specified 3+1 group of four single-phase SVC transformers rated 400/16.7 kV and 60/80/100 MVA. Combining 72 A with a 433 A fundamental gives only a 1.34% increase in total RMS current, but RMS current alone does not determine transformer heating or redesign cost.
Where does the harmonic current travel?
Follow the current from its measurement point through the terminals, winding conductors, leakage field, structural metal, cooling system, and insulation. Each harmonic sees a different effective winding resistance and produces a different leakage-field distribution. The responsible source may be background grid distortion, the SVC, or another nonlinear load; simultaneous voltage and current measurements at defined transformer terminals establish direction and operating condition.
Layer one first: confirm the current-transformer ratio, channel scaling, polarity, bandwidth, sampling method, measurement location, and aggregation method. A spectrum recorded on the 400 kV side cannot be compared directly with winding or 16.7 kV-side current without referring every component through the applicable ratio.
| Reading | Required identification | Next branch |
|---|---|---|
433 A |
Fundamental RMS, peak, or instantaneous amplitude; terminal and operating point | Normalize before calculating losses |
11 A and 72 A
|
Single harmonic, arithmetic sum, or root-sum-square spectrum | Request individual harmonic orders |
| Voltage spectrum | Magnitude and phase by harmonic at the same location and time | Evaluate flux and harmonic power direction |
| Transformer loading | Tap, cooling state, terminal voltage, and MVA during capture | Match the guaranteed design case |
Do the current values share the same reference?
The stated transformer ratings and currents do not identify a common electrical base. If 433 A is single-phase current at 400 kV, its apparent power is 400 kV × 433 A = 173.2 MVA. If it is line current for a three-phase bank at 400 kV, the result is √3 × 400 kV × 433 A = 300.0 MVA. Neither result directly matches the stated 60/80/100 MVA sequence.
This mismatch does not invalidate the current measurement. It means the design review must identify whether the amperes are primary, secondary, referred, internal-winding, per-unit, peak, or measured under an operating point different from the nameplate rating. The supplier cannot produce an auditable loss calculation until every harmonic uses the same RMS base and transformer side.
- Export the spectrum as harmonic order, frequency, RMS amperes, phase angle, measurement terminal, and timestamp or operating interval.
- Record the fundamental voltage and current, SVC operating state, tap position, and cooling state for the same interval.
- Apply documented CT ratios and refer the complete spectrum to one selected winding.
- Reconcile the resulting fundamental current against the applicable transformer rating and connection.
Is the RMS increase the deciding metric?
If 433 A is fundamental RMS and each quoted harmonic value is the root-sum-square RMS of components orthogonal to it, the totals are:
| Case | Calculation | Total RMS |
|---|---|---|
| Original | √(433² + 11²) |
433.14 A |
| Measured | √(433² + 72²) |
438.95 A |
| Change | (438.95 / 433.14 − 1) × 100% |
1.34% |
That calculation describes conductor current magnitude, not frequency-dependent loss. Loss in a winding can be represented by P = Σ[I(h)² × R_ac(h)]. Skin effect, proximity effect, and leakage-field eddy currents make R_ac(h) rise with frequency and depend on conductor dimensions, transposition, winding geometry, and nearby structural parts.
Even before applying frequency weighting, the squared contribution associated with 72 A is (72/11)² = 42.84 times the contribution associated with 11 A, if both values represent the same harmonic order and measurement basis. If either number is an arithmetic sum or represents different spectral distributions, this comparison is invalid; calculate each harmonic separately.
Which physical loss path can force a redesign?
| Loss path | Primary driver | Possible design response | Required proof |
|---|---|---|---|
| DC winding loss | Total RMS current and winding resistance | More conductor area or revised cooling | Winding loss at guaranteed temperature |
| Winding eddy loss | Harmonic order, conductor size, and leakage field | Smaller strands, conductor rearrangement, or transposition changes | Loss by winding and harmonic |
| Structural stray loss | Leakage flux entering clamps, tank, leads, and shields | Geometry or magnetic shielding changes | Hot-spot and stray-loss calculation |
| Core loss | Applied voltage waveform, frequency, flux density, and possible DC bias | Lower flux density, different steel, or thinner laminations | Voltage spectrum and calculated core flux |
| Thermal duty | Total losses, cooling state, ambient condition, and duty cycle | Additional cooling or greater active-part dimensions | Top-oil, winding, lead, and structural hot-spot results |
Load-current harmonics do not automatically create proportional core-flux harmonics. Core flux follows applied voltage approximately as Φ(h) = V(h)/(2πhfN). A claim for substantially more core steel therefore needs the voltage-harmonic spectrum, selected flux density, and calculated core-loss or hot-spot change. More core area can also increase mean winding turn length, so a magnetic redesign may add copper as a secondary effect.
A distribution-transformer K-factor is not an adequate substitute for a loss-by-harmonic study of these custom high-power single-phase units. The decision input is the individual spectrum combined with the proposed winding and structural geometry.
What must the revised quotation disclose?
Separate technical redesign from commercial escalation. Copper and steel commodity adjustments may change price independently of harmonic performance, so the quotation needs distinct material-price and design-change lines.
- Issue one agreed harmonic design spectrum with normal, maximum, and credible contingency operating cases. State whether components are continuous or tied to a defined duty interval.
- Request original-versus-revised guaranteed losses, including DC winding, winding eddy, structural stray, and core loss.
- Request the calculated temperature effect by winding, lead, core, clamp, shield, and tank region rather than one total-loss value.
- Ask which dimensions, conductor arrangement, steel grade, lamination thickness, shielding, and cooling provisions changed, without requiring disclosure of proprietary manufacturing detail.
- Require separate cost deltas for active material, cooling equipment, manufacturing complexity, transport effects, and commodity-price adjustment.
- Compare a design using the measured spectrum with any alternative based only on
438.95 A. Reject the RMS-only option if it omits frequency-dependent winding and stray losses.
How is the resolving design verified?
Verification must preserve the same data path used for design. Confirm instrument scaling and bandwidth, reproduce the agreed operating state, and capture simultaneous terminal voltage and current spectra. Compare each measured harmonic with its design-envelope value rather than comparing only total RMS current.
| Check | Pass condition |
|---|---|
| Spectrum | Every measured component remains within the agreed design spectrum for its operating case |
| Loading | Fundamental current, voltage, tap, and cooling state match the test or calculation basis |
| Losses | Guaranteed and measured losses use the same temperature reference and component definitions |
| Thermal performance | Specified winding and structural hot-spot limits are met under the agreed harmonic duty |
| Commercial reconciliation | Technical modifications and commodity adjustments remain separately traceable |
FAQ
What happens if I size the transformer from 438.95 A alone?
The calculation captures total RMS current but misses the increase in R_ac(h) and structural stray loss at higher harmonic orders. Use the individual RMS harmonic spectrum in the loss model.
What happens if 72 A is an arithmetic sum of harmonics?
Do not combine it with 433 A by root-sum-square. Export the RMS value for every harmonic order and calculate total RMS as √ΣI(h)².
What happens if the 433 A and 72 A values come from different transformer sides?
The comparison has no common base. Apply the documented transformation and CT ratios, refer both readings to one winding, and then reconcile the result with the applicable rating.
What happens if the supplier attributes the increase to core steel?
Request the applied voltage spectrum, selected flux density, calculated core-loss change, and identified core hot spot. Current harmonics primarily require winding and stray-loss analysis unless voltage distortion or another magnetic condition changes core flux.
How do I verify the revised transformer design?
Repeat simultaneous terminal voltage and current spectral measurements at the agreed tap, loading, SVC state, and cooling condition. Complete verification by confirming every harmonic and calculated thermal result remains inside the approved design envelope.