Surface-based lattice structures, particularly those derived from triply periodic minimal surfaces (TPMS), offer a promising combination of high stiffness-to-weight ratio, geometric isotropy, and self-supporting characteristics. These properties make them ideal candidates for laser powder bed fusion (LPBF). However, the non-uniform wall thickness inherent to TPMS geometries poses challenges in both simulation and manufacturing, especially for graded structures. This work introduces a simulation strategy that enables efficient modeling of TPMS lattices with spatially varying thicknesses. Shell elements are enhanced using mapped local thickness fields, which are generated from offset surface geometries by employing a standard and efficient nearest-neighbor search algorithm. This approach approximates the actual geometry of complex TPMS structures while significantly reducing computational cost compared to full solid models. Comparative studies demonstrate improved fidelity of shell-based simulations, particularly for low to moderate volume fractions. Simulation results cover several TPMS types (Gyroid, Diamond, Neovius) as well as plate-based unit cells (FCC, BCC), evaluated across various relative densities and thickness gradation strategies. The influence of thickness gradation on mechanical performance is analyzed and benchmarked against both uniform shell models and solid element simulations. Additionally, periodic and symmetric boundary conditions (PBC, SBC) are employed in RVE analysis to streamline the extraction of effective material properties. The proposed methodology lays the foundation for simulating large-scale, graded TPMS sandwich structures. It supports rapid design iteration and integration into structural components. This approach is highly relevant for additive manufacturing applications in lightweight engineering, where tailored material distribution plays a critical role.