Tank-shunt coverage changes where leakage-field concentration occurs. Track shunt ends and adjacent metal with a complete loss and temperature map to determine whether the critical region improves or moves. [1] [2]
Register the end against the leakage source
Map the shunt endpoint, nearby winding ends, leads, wall features and supports. Compare several coverage lengths under the same excitation. Track the field on the shunt, the wall beside its end and the uncovered wall beyond it.
The published multi-step shield study accounts for geometry discontinuities and increased flux entering outermost stacks. Its edge result concerns its specified shield geometry, but it supports examining boundary packets separately from the central region. The winding-asymmetry study also shows that an unprotected wall can receive increased field after shielding elsewhere. [1] [2]
| Coverage map | Quantity | Location to preserve |
|---|---|---|
| Inside the shunt span | Internal density and collected flux | Central and end packets |
| At the endpoint | Field direction and local concentration | End geometry and adjacent wall |
| Beyond coverage | Bypass field and induced loss | Uncovered wall and nearby supports |
| Thermal response | Peak and temperature distribution | Hot-spot coordinates for each candidate |
Inspect peaks without relying on a single mesh cell
Check resolution near the end and the representation of sharp corners. A mathematical corner can create a highly mesh-sensitive peak that does not represent the finite as-built radius. Retain both peak behaviour and the spatial extent of the concentrated region, and compare with the actual end geometry.
Use the relevant vector components as well as field magnitude. Fringing changes direction, which can alter the interaction with laminated material and conducting wall regions. A scalar magnitude plot alone can conceal that change.
Follow the heat to its new location
Transfer the calculated losses into a thermal model with the same mounting and oil-access basis. Magnetic-field and temperature maxima need not coincide because heat spreads and cooling varies. Compare the hottest shunt, wall and support regions across all candidates.
If validating a prototype, place observations around the predicted endpoint and possible migration region rather than leaving a sensor fixed at the original hot spot. A cooler original point does not demonstrate a lower maximum if the new maximum was never measured.
Assume total structural loss remains 200 W. The baseline places 20 W in an end region with assumed effective thermal conductance 2 W/K, giving rise 10 K. A revised coverage moves 40 W into that region and reduces losses elsewhere: local rise becomes 20 K. Against an assumed allowable local rise of 15 K, baseline passes and the revision fails despite unchanged total watts. Keep the baseline candidate while resolving end fringing, its finite heated area and actual cooling. These are independent teaching inputs, not paper test results.
Choose coverage from the complete location map under the required cases. The result should show whether the added material lowers the critical condition, displaces it or creates a new limit at the extended end.
References
- 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.
- 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.