2500 kVA Customized Stacked Core with Rectangular Silicon Steel Sheet
Oct 21, 2025
Leave a message
A stacked core is a closed magnetic circuit component formed by layering and fastening insulated silicon steel sheets flat in a controlled direction. For large power transformers in the 2500 kVA class, the rectangular stacked core is the mainstream structure because it combines a low-loss magnetic circuit with the mechanical rigidity needed for heavy windings and continuous grid duty. This article explains how such cores are designed, which grades are available, and how stacking practice affects transformer performance.
What Is a Stacked Transformer Core?
A stacked core is built from thin sheets of electrical steel, each coated with an insulating layer, stacked face to face and clamped into a closed magnetic circuit. The insulating coating divides the core into independent thin layers so that alternating flux cannot form large-scale eddy currents inside the steel; this is the primary mechanism that keeps eddy-current loss low. The sheets are oriented along the magnetic circuit direction, and for grain-oriented grades the rolling direction of the strip is aligned with the flux path to exploit the material's high permeability and low specific core loss.
Rectangular stacked cores are preferred for 2500 kVA power transformers because the rectangular window accommodates concentric cylindrical windings efficiently, the straight limbs simplify clamping and coil mounting, and the core cross-section can be stepped to approximate a circle and minimize the mean length of the winding turns.
Why Choose a Stacked Structure for a 2500 kVA Core?
The stacked design is chosen for three engineering reasons:
Suppressing eddy-current loss: thin laminations (0.20–0.30 mm typical) shorten eddy-current paths; combined with the insulating coating, total iron loss stays within the guaranteed values of the selected grade.
Optimizing magnetic circuit continuity: sheets are stacked along the flux direction, and with staggered (step-lap) joints the air gaps at the corners are reduced to a minimum, lowering magnetizing current and noise.
Mechanical strength: interleaved laminations and through-bolt or band clamping produce a rigid core that withstands short-circuit forces and transport vibration.
Stacking Methods
Staggered stacking (step-lap): the sheet ends are cut in steps and the joints of adjacent layers are offset. Flux crosses the joint region gradually instead of through a single straight gap, which reduces iron loss and audible noise. This is the standard for 2500 kVA distribution and power transformers.
Flat stacking: joints are aligned layer to layer, which is simpler to cut but creates a continuous air gap at the joint line, increasing magnetizing current and loss. Flat stacking is confined to low-cost, low-performance equipment and is not recommended for large power transformers.
Core Material: Grain-Oriented Silicon Steel
The magnetic performance of a stacked core is defined by the silicon steel grade. Non-oriented grades have uniform properties in all directions and are used in motors and small transformers. Grain-oriented (CRGO) grades offer high permeability and low core loss along the rolling direction and are the correct choice for a 2500 kVA transformer core, where every 0.1 W/kg of core loss matters over a 20–30 year service life. Test methods follow IEC 60404 (magnetic properties) and GB/T 2522 (coating and testing of electrical strip).
Available Grades and Typical Magnetic Properties
The table lists typical grades supplied for rectangular stacked cores, with width from 35 mm to 1200 mm. Core loss is measured at P1.7 T / 50 Hz, lamination factor is given as a minimum percentage, and induction is the minimum value at field strength B800 (A/m).
| Type | Grade | Max Core Loss @P1.7T/50Hz (W/kg) | Lamination Factor (%) | Min Induction @B800 (T) |
|---|---|---|---|---|
| Common | 23Q110 | 1.10 | 97.2 | 1.82 |
| Common | 23Q120 | 1.20 | 97.2 | 1.82 |
| Common | 27Q120 | 1.20 | 97.5 | 1.82 |
| Common | 30G120 | 1.20 | 98.0 | 1.82 |
| High induction | B20P085 | 0.85 | 97.0 | 1.86 |
| High induction | B23P085 | 0.85 | 97.2 | 1.88 |
| High induction | B23P090 | 0.90 | 97.2 | 1.88 |
| High induction | B23P095 | 0.95 | 97.2 | 1.88 |
| High induction | B23P100 | 1.00 | 97.2 | 1.88 |
| High induction | B27P095 | 0.95 | 97.5 | 1.89 |
| High induction | B27P100 | 1.00 | 97.5 | 1.89 |
| High induction | B30P120 | 1.20 | 98.0 | 1.89 |
| Domain refined (laser) | B20R070 | 0.70 | 97.0 | 1.86 |
| Domain refined (laser) | B23R085 | 0.85 | 97.2 | 1.88 |
| Domain refined (laser) | B27RK085 | 0.85 | 97.5 | 1.89 |
A wider range of grades is available for selection, including B30P105, B30R100, 27QG120, 27QG100, 23QG100, 23QG095, 23RK090, 23RK085, 23QG085, 23RK080, 23RK075, B23R070, 20QG085, 20RK080, 20RK075, B20R070, B20R065 and the M0H, M1H, M2, M3, M4, M5 series. Each grade is selected to match the transformer's guaranteed loss targets, working flux density and manufacturing method.
Customization for 2500 kVA Class Cores
As a transformer core manufacturer, GNEE New Material supplies rectangular stacked cores to the transformer designer's drawing, including core window dimensions, limb and yoke cross-section steps, number and pitch of cooling ducts, clamping plate design, and painting or varnish finish. Core loss and magnetizing current are guaranteed per the selected grade and are verified on finished cores by no-load testing before delivery.
FAQ
What is the difference between step-lap and flat stacking?
In step-lap stacking the joints of adjacent lamination layers are offset in steps, so flux transfers between layers gradually and the equivalent air gap is much smaller. Flat stacking aligns all joints in the same plane, creating a continuous gap that raises magnetizing current, loss and noise. Step-lap is therefore standard for 2500 kVA transformer cores.
Which silicon steel grade is best for a 2500 kVA transformer core?
Selection depends on the guaranteed no-load loss and the working flux density. Typical choices are 23Q110–30G120 for conventional designs, high-induction grades such as B23P085–B30P120 when a higher design flux density (up to 1.88–1.89 T) is needed, and domain-refined grades such as B20R070 or B23R085 for the lowest loss targets.
Why are grain-oriented grades used instead of non-oriented grades?
Grain-oriented silicon steel has much lower core loss and higher permeability along the rolling direction, which is aligned with the flux path in a transformer core. Non-oriented grades are cheaper but have higher loss and are normally limited to motors and small transformers.
How does lamination thickness affect core loss?
Eddy-current loss rises with the square of lamination thickness, so thinner sheets (0.20–0.23 mm) give lower total loss at 50/60 Hz. The trade-off is a lower stacking factor and higher manufacturing cost, so the thickness is chosen to meet the guaranteed loss at the design flux density.
How is core loss verified on a finished stacked core?
Finished cores are checked by no-load (open-circuit) testing of the assembled transformer, where no-load loss and no-load current are measured at rated voltage and frequency per IEC 60076-1. The measured values are compared against the guaranteed values derived from the grade table.

