Advanced Modeling of Large-Scale Electromagnetic Systems with Application to Reconfigurable Intelligent Surfaces

The rapid evolution of wireless communication technologies, particularly with the advent of sixth-generation (6G) systems, has brought significant attention to optimizing electromagnetic (EM) wave propagation environments. A central focus in this field is the development of Reconfigurable Intelligent Surfaces (RISs), which are poised to play a transformative role in enabling flexible, adaptive, and efficient wireless communication systems. This doctoral thesis aims to consolidate recent advancements in EM modeling, exploring how various modeling techniques in both the time-domain (TD) and frequency-domain (FD) approaches enhance the design and functionality of RISs. RISs offer a pathway to optimize EM wave environments by dynamically reconfiguring the propagation channels to improve communication quality and reliability. To support RIS design, EM simulation provides critical insights into complex interactions such as crosstalk, signal dispersion, reflections, radiation, and conductor and dielectric losses, all of which substantially impact electronic systems. These effects become even more pronounced at higher operating frequencies as parasitic elements begin to influence interconnection behavior, and high-frequency behaviors in passive structures risk degrading system efficiency without proper countermeasures. The literature on numerical techniques for full-wave EM analysis generally divides these methods into differential and integral approaches. Differential methods, such as the Finite-Element Method (FEM) and Finite-Difference (FD) techniques, discretize Maxwell's equations across the entire computational domain. Integral Equation (IE) methods, by contrast, use formulations like the Electric Field Integral Equation (EFIE) and Magnetic Field Integral Equation (MFIE), applying the Volume Equivalence Principle for Maxwell's equations. In IE methods, computations are confined to the material-occupied regions (e.g., conductors and dielectrics), making IE methods a powerful choice for scattering problems frequently encountered in electronic and telecommunication applications. Among the IE methods, the Partial Elements Equivalent Circuit (PEEC) method has recently received particular attention for translating EM problems into circuit-based models compatible with SPICE-like simulators. By leveraging the volume equivalence principle, PEEC formulates a volumetric EFIE, using currents flowing within the scatterer as unknowns and sources for the scattered field. Concerning the time-domain modeling, this research introduces and validates a PEEC-based time-domain approach, addressing the limitations of traditional frequency-domain techniques in scenarios involving nonlinearities and time-varying metasurface configurations. To enhance computational efficiency and accuracy, this novel methodology enables the efficient computation of impulse responses and integrates seamlessly with convolution-based solvers, allowing for improved system-level analysis. In particular, we explore innovative macromodeling techniques, such as piecewise constant and piecewise linear approximations of impulse responses, combined with a convolution-based algorithm. Additionally, a quasi-recursive convolution scheme is introduced, reducing the computational complexity. Furthermore, this work emphasizes the importance of frequency-domain techniques for characterizing RISs. Through a comprehensive validation of an analytical multiport model for dipole-based RISs against full-wave numerical simulations within the PEEC framework, we demonstrate the effectiveness of analytical formulations in capturing the key characteristics of RIS-assisted wireless channels. Through this multi-faceted approach, the thesis aims to contribute valuable knowledge to developing RIS technologies, ultimately advancing the adaptability and efficiency of wireless communication systems in next-generation applications.