This research introduces the concept of physically-based periodic Representative Volume Elements (RVEs) tailored for metal components fabricated via electron-based and laser-based Powder Bed Fusion (PBF) processes. Microstructural modeling approaches can be classified into two fundamentally different categories [1]. Geometrically-based methods, such as Voronoi tessellation, synthetically generate microstructures to resemble experimental statistics. In contrast, physically-based methods, including Cellular Automata (CA), simulate microstructure formation by tracking solidification and grain growth during manufacturing. While geometrically-based approaches easily support periodicity, they do not account for the physics of microstructure formation. Physically-based models capture these formation mechanisms, but introducing periodicity into these models remains a major challenge. To address this gap, the proposed study adapts CA modeling to generate microstructures exhibiting inherent periodicity suitable for RVE extraction. Periodic RVEs offer two major advantages. First, they enable the use of Fast Fourier Transform (FFT) solvers [2], which leverage periodicity to significantly reduce computational cost in mechanical problems, especially when compared to conventional finite element solvers. Second, they eliminate artificial boundary effects, reducing the required size of the RVE to accurately capture effective material properties. Periodic RVEs have already been developed for composite materials, random polycrystals, and geometrically-based PBF microstructures. However, their extension to physically-based PBF microstructures remains unexplored. The study builds upon an established CA framework for laser-based PBF processing of 316L [3], introducing modifications to the initial baseplate microstructure to enforce periodic boundary conditions. Key research questions addressed include verifying the periodicity of the generated microstructures, validating their accuracy against experimental observations and baseline non-periodic CA simulations, assessing variability due to different initial baseplate conditions, and determining the optimal RVE dimensions. Ultimately, this methodology is expected to reduce the computational cost of mechanical properties prediction under defined PBF process parameters, supporting faster microstructural optimization and residual stress estimation in part-scale simulation models.