Asymmetrical through-fault current needs a time-resolved source and material model. This note screens whether a steady linear field calculation covers that duty; it does not calculate a qualified fault temperature or support load.
Define the waveform from the fault study
A common waveform descriptor is i(t) = iac(t) + Ioffset exp(−t/τ). The actual alternating component, offset and decay constant must come from the fault study; the descriptor alone supplies no design duty.
Retain the winding current relationships and lead paths. Through-fault ampere-turn balance does not eliminate leakage fields, and a transient imbalance or offset must be represented consistently across the relevant circuits. [1]
Follow the nonlinear magnetic trajectory
| Time-history record | Quantity | Decision supported |
|---|---|---|
| Specified currents | Instantaneous values and relative timing | Correct transient source |
| Shunt response | Local B and H trajectory | Excursion into the nonlinear operating region |
| Bypass field | Tank and neighbouring structural field | Whether saturated collection shifts the burden |
| Dissipation | Loss versus time and location | Accumulated short-duration heat |
Use the material description appropriate to the excursion and initial state. A constant-permeability or infinitely permeable collector cannot establish saturation margin. The published multi-step field method uses such simplifying assumptions in selected source-transfer stages; its ordinary benchmark does not validate a biased fault transient. [2]
An RMS fundamental-frequency solution cannot reproduce a decaying offset. Independent harmonic addition assumes linear response; where permeability depends on the trajectory, use a material formulation that retains the required time history.
Keep transient heat distinct
Evaluate the relevant time-dependent dissipated power and integrate it over the specified interval. Initial temperature, heat capacity, conduction and oil response determine the resulting temperature excursion. A peak magnetic density alone supplies none of those quantities.
If an adiabatic approximation is used for a short interval, justify it against the relevant thermal time scales and state what it omits. Do not use a normal-load steady-state temperature limit as an automatic permission for a fault duration.
Assume a linear calculation gives 1.20 T at 20 kA peak, and material data for this exercise cover only 0–1.80 T in the required direction. Prescribe the teaching current i(t) = 20 sin(2π50t) + 20 exp(−t/0.040) kA, with t in seconds. At 5 ms, i = 37.65 kA and linear scaling gives B = 2.259 T, outside that assumed data range. Reject the linear model as evidence for this duty; 2.259 T is an unsupported extrapolation, not a predicted saturated waveform. Obtain directional transient material data and solve field redistribution before calculating loss energy or support forces.
Report which waveform portions and field regions exceed the model’s supported domain. Fault heating and support acceptance remain separate calculations requiring resolved loss and force histories; the steady shield benchmark does not supply them. [2]
References
- I. A. Hernandez-Robles et al., Analysing and Computing the Impact of Geometric Asymmetric Coils on Transformer Stray Losses, Applied System Innovation 7, 26, 2024. DOI: 10.3390/asi7020026.
- Z. Jurković, B. Jurišić and T. Župan, Multi-Step Approach for Fast Calculation of Magnetic Field in Transformer Tank Shields, Energies 17, 1378, 2024. DOI: 10.3390/en17061378.