Coupling of CFD and structural mechanics simulation for the prediction of manufacturing effects on filter media

In many cases, industrial filter media are designed as multi -layered combinations of different materials such as supporting or spacing meshes, protection fleeces and filtering nonwovens. When performing simulations on the macroscopic length scale of an entire filter element, layered media are usually treated as a homogenized material s, described by effective properties like flow resistivity (permeability), mechanical strength and filtering efficiency which can be obtained by experiments. There is a growing interest in a deeper understanding of the impact of the manufacturing methods on the effective properties of multi- layered media. For instance, when a supporting mesh and a filtering non -woven are combined under mechanical compression, the mesh wires leave imprints (pits) in the nonwoven layer (e.g. during calendering or pleating), which will lead to strong local variations in both permeability and filtering efficiency. In order to predict these changes, it is necessary to model the impact of the compression and to couple structural mechanics simulations with CFD. To this end, we present a multi-scale and multi-physics simulation approach for a representative region of the multi-layered medium on the mesoscopic scale or so-called media level. Based on known material properties of the media components before the manufacturing, a structural mechanics simulation of the compression of the multi-layered medium is performed. For the resulting distribution of the fiber volume fraction in the non-woven layer(s), the local permeability is updated. The subsequent simulation of the fluid flow yields the effective permeability of the medium after the manufacturing. Even more challenging is the modeling of the impact on the filtration efficiency of the medium. For this, we present a simulation study of a compressed non- woven material on the microscopic level (pore scale). It is seen that this approach reduces the computational cost by orders of magnitude in comparison to Direct Numerical Simulation (DNS).