This research project will focus on a detailed investigation of micro-scale flows occurring during composites manufacturing with special emphasis on the filling of preform-containing mold cavities with low-viscosity resins, as encountered in the area of liquid molding of composites. This work will provide a clearer understanding of the detailed nature of micro-scale flow patterns that are specific to liquid molding - and thus guide future process design efforts. Methods: For computation we will use in-house parallel boundary element codes to solve the equations of Stokes flow across 2D and 3D microstructures that are reasonably realistic local representations of commercial fabrics used in liquid molding. The major challenge stems from the sheer size of the computational problem, since such representations can easily contain many thousands of individual fibers and millions of degrees of freedom; this will be addressed by parallel computation using the Boundary Element Method. Outcomes: A clearer understanding of the nature and extent of micro-scale flows occurring in liquid molding. Structure-Permeability correlations. Intellectual merit: The application of highly advanced numerical techniques to the given problem. The potential to result in significant advances in the state-of-the-art of modeling of liquid molding processes, including elimination of defects. Broader impacts: The development of structure-property correlations in composite and porous materials and cross-fertilization of methods and techniques between the areas of manufacturing and computational analysis of micro-scale fluid flow. The training of future engineers in these methods and techniques. The results of this work will be of direct relevance to other applications involving flow across structured fiber arrays, such as reaction engineering, biotechnology and physiological flows, filtration and membrane separations, drying of paper and wood etc.