Laser Powder Bed Fusion (LPBF) is a fast-developing metal additive manufacturing process offering unique capabilities including geometric freedom, flexibility, and part customization. The process induces complicated thermal histories with high temperature gradients and cooling rates, leading to rapid solidification microstructures with anisotropic properties as different from those produced conventionally. In addition, the LPBF parts exhibit to a large extent of in-sample and sample-to-sample variabilities in the microstructure and consequently part performance. The high variability in the microstructure and properties is considered the major obstacle against the widespread adoption of LPBF as a viable manufacturing technique. Therefore, a more in depth understanding and control of the solidification microstructure is needed to achieve the LPBF fabricated parts with desired properties. Since the solidification microstructure is highly influenced by the thermal input, it is essential to have an accreditable thermal model first. Therefore, a portion of this dissertation was devoted to developing an accurate thermal model through various methods including code-to-code verification and experimental validation. The materials used in this portion include Ti-6Al-4V, NiTi-SMA (Shape Memory Alloy). Next, a multi-scale multi-physics modeling framework which couples a finite element (FE) thermal model to a non-equilibrium phase field (PF) model was developed to investigate the rapid solidification microstructure during LPBF. The framework was utilized to predict the spatial variation in the morphology, size and micro-segregation in the single-track deposition of binary NiNb alloy during LPBF and a very good agreement with the experimental measurements was achieved.
