Depth filters are employed for various purposes, for the purification of liquids as well as the cleaning of gases. The technique is also used in different industries, such as energy conversion, waste water treatment, beverage technology, and air conditioning. With current manufacturing methods, precisely-defined gradient media for depth filters can be built. However, it remains unclear which gradients of local filtration properties are desirable for different applications. Therefore, we aim to contribute to a general design method for depth filter media with the present study. A multi-scale approach is introduced in which pore-scale simulations are combined with corresponding continuum models. The present work is based on two previous studies [1,2] conducted by separate research groups which combine their findings here.
Methodologically, pore-scale simulations (software: GeoDict, supplier: Math2Market) of depth filtration are conducted from which the parameters of continuum models are determined. A representative domain size for all pore-scale simulations is determined in a preliminary study. On the continuum scale, predictions are made with respect to the local separation properties of the filters, as described by the filter coefficient (software: Matlab, supplier: Mathworks). Based on optimal control solutions, the local filter coefficient is determined so that deposition within the filter media is homogeneous along the filter depth because inhomogeneous clogging is a known problem of depth filters. In the optimization, the constraint of a given overall filtration efficiency is applied. Subsequently, the optimized filter-coefficient trajectories are translated back to the pore scale and the results are validated. The described strategy is conducted for a fibrous filter medium (I), composed of fibers of two different diameters, as well as for a packed-bed filter which consists of a binary mixtures of spherical particles (II). For both case studies, it is found that...
Session: F11 - Numerical Methods for Filter Media Characterization and Improvement
Day: 24 October 2019
Time: 13:00 - 14:15 h