Quasi-Static Method-of-Moments Modeling of Near-Field Magnetic Coupling in EMI Filter Hardware with Experimental Validation on Multiple PCB Configurations

Electromagnetic interference (EMI) filters are critical to achieving conducted-emission compliance in modern power electronic converters. In compact implementations, however, parasitic near-field magnetic coupling between common-mode (CM) and differential-mode (DM) chokes and adjacent capacitor structures can perturb the intended filter response, promote CM-DM mode conversion, and degrade insertion-loss performance. Although these interactions can be captured with high-fidelity electromagnetic solvers, such as full-wave methods or volumetrically discretized partial element equivalent circuit (PEEC) models, those approaches are often computationally prohibitive for rapid iteration across multiple hardware layouts.This thesis develops a magnetoquasistatic, Method-of-Moments (MoM)-inspired modeling framework for extracting mutual coupling and predicting near-field magnetic field distributions in an EMI filter test printed circuit board (PCB). To reduce computational complexity, the formulation avoids volumetric discretization and instead represents conductive paths as segmented filamentary elements, thereby retaining the dominant inductive interactions governing near-field behavior while significantly reducing memory and runtime requirements. Ferrite cores are incorporated through configuration-specific effective-permeability calibration at 1 MHz, and the common-mode choke is modeled as two explicitly coupled windings on a shared core in order to preserve both CM and DM excitation components. Measured phasor winding currents are used as model inputs, while capacitor-loop currents are obtained through a coupled impedance formulation that captures magnetic interaction between the choke and capacitor network.
The framework is evaluated using four physical PCB configurations in which the chokes are repositioned and reoriented relative to a fixed capacitor network. The resulting coupling matrices and H-field distributions provide quantitative insight into the extent to which component placement and orientation influence coupling strength and spatial field localization. The proposed model therefore serves as a computationally efficient intermediate layer between circuit-level EMI filter synthesis and layout-resolved electromagnetic analysis, while also establishing a foundation for future extensions that incorporate more complete capacitive coupling and PCB power/return-plane effects.