DDM-Based Accurate and Efficient RIS Modeling for Over-the-Air Mutual Coupling Analysis and Comparison vs Macromodeling

Artikel i vetenskaplig tidskrift, 2026

Existing modeling approaches for reconfigurable intelligent surfaces (RISs)—including scalar approximations, boundary-condition-based formulations, and conventional full-wave solvers—either lack accuracy in capturing aperiodic interactions or become computationally prohibitive for electrically large, multi-scale structures. To address these limitations at least for a certain class of RISs, which are composed of open-cavity-based meta-atoms, this paper presents a rigorous and computationally efficient hybrid domain decomposition method (H-DDM). The method decomposes each meta-atom into interior and exterior subproblems, enabling fast and accurate modeling of over-the-air mutual coupling effects while preserving microscopic field behavior. The proposed H-DDM is cross-validated against CST full-wave solvers and compared against two widely used approximate methods: macromodeling (MM) and the element-in-array-pattern (EIAP) technique. Using RIS panels of varying sizes which are beamformed to have a non-specular reflection under a normal plane wave incidence, we demonstrate that H-DDM achieves very good agreement with full-wave reference solutions, while MM and EIAP exhibit increasing deviations in sidelobe regions due to their inherent assumptions. The results confirm that H-DDM provides a powerful and reliable tool for analyzing open-cavity based RIS panels and offers a promising pathway toward fast, high-fidelity EM modeling of next-generation beamforming metasurfaces.