To achieve high-frequency inductance and impedance enhancement for effective electromagnetic interference (EMI) mitigation in power electronics, this paper presents an electromagnetic analysis and experimental study of laminated Mn-Zn toroidal ferrite cores. The electromagnetic field is analyzed using a 2D analytical solution based on a simplified Cartesian approximation. Although neglecting curvature, this approach enables efficient eigenfunction expansion and is rigorously validated against cylindrical finite difference (FDM) and 3D finite element (FEM) benchmarks. The results demonstrate that lamination effectively interrupts eddy current loops; notably, a four-layer structure increases the resonant frequency by approximately 2.8 times compared to a monolithic core. Experimental measurements confirm that this design significantly mitigates the skin effect and extends the stable frequency bandwidth. This study establishes a validated, computationally efficient methodology for optimizing core geometries to prevent impedance degradation.

Knowledge distillation achieves model compression by training a lightweight student network to mimic the output distribution of a larger teacher network. However, when the teacher becomes overconfident, its sharply peaked logits break the scale symmetry of supervision and induce high-variance gradients, leading to unstable optimization. Meanwhile, research that focuses only on final-logit alignment often fails to utilize intermediate semantic structure effectively. This causes weak discrimination of student representations, especially under class imbalance. To address these issues, we propose Scale Calibration and Pressure-Driven Knowledge Distillation (SPKD): a one-stage framework comprising two lightweight, complementary mechanisms. First, a dynamic scale calibration module normalizes the teacher’s logits to a consistent magnitude, reducing gradient variance. Secondly, an adaptive pressure-driven mechanism refines student learning by preventing feature collapse and promoting intra-class compactness and inter-class separability. Extensive experiments on CIFAR-100 and ImageNet demonstrate that SPKD achieves superior performance to distillation baselines across various teacher–student combinations. For example, SPKD achieves a score of 74.84% on CIFAR-100 for the homogeneous architecture VGG13-VGG8. Additional evidence from logit norm and gradient variance statistics, as well as representation analyses, proves the fact that SPKD stabilizes optimization while learning more discriminative and well-structured features.