Discrete-element-method (DEM) simulations were shown to be useable for selecting screens with simple geometries (e.g., wirewrap screens) under prepack test conditions (Mondal et al. 2011). Metal mesh screens (MMSs) (commonly referred to as premium screens) can have multiple (sometimes sintered) layers with different weave patterns. Any model for estimating sand production in such screens must account for these complexities. The objective of this paper is to explore different techniques to create accurate representations of complex filter media and to use them in numerical simulations and/or analytical models for improved screen selection. In this paper, we used computed-tomography (CT) images of real-screen coupons to construct 3D replicas of two MMS types: plain square mesh (PSM) and plain Dutch weave (PDW). An entirely computer-based method of creating 3D screen assemblies is also presented. These virtual screens were compared and validated against those generated from the CT images. We conducted DEM simulations of prepack tests through these multilayer MMS assemblies. The retention efficiency of MMSs for different particle sizes was also calculated for better slurry-test modeling. We quantitatively showed that the effective size of MMS assemblies can be significantly lower than the nominal rating of the filter layer. This paper augments the prepack-test and slurry-test models for mesh-type screens presented earlier (Mondal et al. 2012; Chanpura et al. 2013). Determination of the retention efficiency of MMSs, which is currently characterized by a single filter-cut point, is also a step forward in quality assurance and quality control of these screens. Finally, we demonstrate that one can conduct in-depth screen design and optimization studies with virtual meshes in a consistent way and at a fraction of the cost and time that would be required otherwise by use of conventional experimental techniques.