Phase field simulation of solidification microstructure for Fe-17 wt%Cr model alloy fabricated by laser powder bed fusion

• A simulation model developed to predict solidification microstructure in LPBF. • Multiscale model validated against single- and double-track LPBF experiments. • The FVPF model captures heat transfer, melt pool dynamics, and microstructure. • Effects of G, V, and hatch spacing on microstructure morphology investigated. Accurate prediction of solidification microstructures is essential for optimizing laser powder bed fusion (LPBF) process. We present a coupled multiscale framework combining finite volume melt pool simulations with a two-dimensional microscopic phase-field (PF) model. The process-scale model captures transient heat transfer and Marangoni convection, providing local solidification parameters including temperature gradient (G) and growth velocity (V) that directly drive microstructure simulations. Single- and double-track LPBF experiments were conducted on 316L stainless steel under comparable thermal conditions providing model validation. It is shown that increasing laser energy density under conduction mode results in deeper melt pool penetration, enhanced Marangoni convection, cell size enlargement, and more defined cellular structures, collectively promoting melt pool stability and dimensional growth. Results reveal distinct behaviors for successive scan tracks: the second track develops a wider, shallower melt pool due to residual heat and asymmetric boundaries, leading to reduced cooling rates. Simulations capture planar-to-cellular transition and predict primary arm spacings consistent with Hunt and Kurz–Fisher models. Comparison with micrographs confirms that the model reproduces spatial variations in grain morphology and solute segregation. This work demonstrates that a physics-informed CFD–PF coupling, together with a pseudo-binary alloy approximation, enables reliable yet computationally efficient microstructure prediction in LPBF.