Laser powder bed fusion (LPBF) additive manufacturing of metals offers an almost unlimited freedom of design and the potential for local microstructure control. The metallurgical microstructure arising from these complex processes is determined by the local temperature evolution and history at a given material point, and thus, by the chosen laser scan pattern and the part geometry. Consequently, to enable a digital process design for components with tailored microstructure, computational models for microstructure prediction would be desirable that consistently resolve the scan pattern, i.e., the physical laser beam path, while capturing geometries on the scale of real parts. While existing models accounting for spatially resolved crystal structures, e.g., based on the phase field method, offer very detailed insights into underlying physics-based phenomena of crystal formation and dissolution, the associated computational costs typically limit these approaches to small representative volumes. In this talk, a recently developed physics-based and data-supported phenomenological microstructure model for Ti-6Al-4V is presented. The model predicts spatially homogenized phase fractions of the most relevant microstructural species, namely the stable -phase, the stable s-phase as well as the metastable martensitic m-phase, in a physically consistent manner. The predictive ability of the model is demonstrated by means of continuous cooling transformation (CCT) diagrams, showing that experimentally observed characteristics such as critical cooling rates are correctly predicted by the proposed microstructure model. Moreover, due to the continuum representation of phase fractions instead of resolved crystal structures, this approach offers the general suitability for part-scale simulations. In particular, to provide the required temperature information, the combination of this microstructure model with a recently proposed, highly efficient computational model for scan-resolved simulations of LPBF on the scale of real parts will be considered as promising aspect of our future research.