Computational modeling is an emergent approach to complement and expand the quality-by-design (QbD) approach through research, development and applications of experiment informed and validated multiphysics models and predictive simulation tools for de-risking the scale-up of biomass feedstock preprocessing and handling unit operations (grinding, sieving, feeding, hopper discharge, etc.). This presentation will overview the recent progress and challenges in the development of science-based, multiphysics-resolved digital twins of feedstock preprocessing and handling unit operations. A variety of numerical models for simulating biomass particle size reduction, particle size classification, and granular flow & transport from the laboratory scale to the pilot scale will be presented. The best practices of using those numerical models regarding their fidelity and computational costs will be discussed as well.