EMI Reduction and ICs Optimal Arrangement Inside High-Speed Networking Equipment Using Particle Swarm Optimization

In this paper, a numerical approach to the modeling of electromagnetic interference (EMI) from the emissions of integrated circuits (ICs) and printed circuit boards (PCBs), inside rectangular metallic enclosures of communication devices, is presented. The ICs are modeled as small magnetic dipoles, and their interaction with the enclosures are formulated by means of dyadic Green's functions. Given the probability distribution describing the magnetic dipole sources, a Monte Carlo simulation is employed to derive the electric current density on the metallic walls as well as to perform a worse-case analysis. The EMI in a network switch and a network router are investigated by means of sources-device models developed for each one of the devices. These models are based on data obtained from magnetic probe measurements in various types of network equipment. Finally, the particle swarm optimization algorithm is utilized to indicate the optimal arrangement of ICs that results in significant reduction of the electric current density induced on the metallic walls.