During the laser powder bed fusion (L-PBF) process in metal additive manufacturing, the formation of columnar dendritic microstructures induces an anisotropic mechanical behavior which is key in understanding the development of stresses and defects such as hot cracking (solidification cracking). As the grain structure directly results from heat transfer and fluid flow in the melt pool, the objective of this study is to concurrently address fluid dynamics, grain structure formation and stress build-up in L-PBF process simulation, focusing on investigating defect formation through thermal-metallurgical-mechanical simulations. This study presents a coupled thermo-fluid-solid numerical model and a grain growth simulation framework for L-PBF, incorporating a two-step partitioned approach [1] to separately resolve fluid flow, grain structure evolution and solid mechanics at each time increment. Thermo-hydrodynamics in melt pool is simulated within a level-set based finite element framework. Grain structure evolution is modeled using a Cellular Automaton (CA) method, which simulates growth within the mushy zone. Concurrently, stress evolution is studied using a crystalline elasto-viscoplastic (CEVP) model, considering slip systems of individual grains to understand stress distribution based on crystallographic orientation. Already applied in the simpler context of multiple laser scan lines on a substrate [2], this approach is applied to L-PBF, at the scale of several adjacent tracks, to investigate the mechanisms driving stress generation, with a particular focus on structural effects. This involves analyzing non-uniform deformations within individual grains, intra-granular texture evolution, and the inherent intergranular stresses that develop during solidification. The insights obtained from these analyses are expected to contribute to the establishment of a refined criterion for predicting hot cracking. References [1] Zhang Shaojie, Guillemot Gildas, Gandin Charles-André, Bellet Michel, A partitioned two-step solution algorithm for concurrent fluid flow and stress-strain numerical simulation in solidification processes, Computer Methods in Applied Mechanics and Engineering 356 (2019) 294-324, doi.org/10.1016/j.cma.2019.07.006 [2] Li Zixuan, Bellet Michel, Gandin Charles-André, Upadhyay Manas, Zhang Yancheng, Metallurgically-driven thermomechanical analysis of multiple side-to-side laser melting on a 316L substrate, submitted to Additive Manufacturing, Februar