Model & Simulate in Filtration

FilterDict®

FilterDict is a powerful module for the simulation of particulate filtration in filter media and filter elements to complete filters. Additionally, predefined settings with user input allow the simulation of cross-flow filtration and the simulation of flow in the complete filter with housing.

FilterDict® is a registered trademark of Math2Market GmbH.

FilterDict is a pioneer software tool, widely used in academic and industrial settings, and the only software on the market providing this functionality on standard hardware (16GB 64bit Windows or Linux computers).

Recently, M2M has become active in simulating room air filtration systems and their optimal room placement to increase particle removal efficacy and possibly reduce the risk of viral infection.

Examples of applications for simulations

  • Aerosol filtration in a nonwoven filter 
  • Soot filtration through a honeycomb structure of a Diesel Particulate Filter (DPF) or Gasoline Particulate Filter (GPF)
  • Oil filtration multi-pass test bench
  • Micronic particle filtration and clogging phenomena in microfluidic devices
  • Particle tracking through a pleat element
  • Clogging of a high-efficiency particulate air (HEPA) filter pleat
  • Room air purifiers and their placement in the room
Technical Report

Filtration modeling and simulation with GeoDict®, from filter media to filter element

Azimian M., Linden S., Cheng L., Wiegmann A., 2020: Filtration modeling and simulation with GeoDict®, from filter media to filter element, Math2Market GmbH, technical report, ​https://doi.org/10.30423/report.m2m-2020-03

Read Technical Report

FilterDict® for filter media

The filter media can be any kind of porous media, typically nonvowen media, woven media (metal, textile, or plastic), foam, membrane (e.g. desalination), or cellulose paper, but FilterDict can also simulate filtration through soil or sand, and others. Any of these filter media can be analyzed, optimized and/or newly designed. Even for nanofibrous filter media, FilterDict simulates filtration accurately by incorporating the slip effect.

With the FilterDict module, filter efficiency computations requiring a single flow field computation can be achieved for realistic nonwoven models (300 x 300 x 600 computational cells) in a few hours. Filter life time simulations may take several days due to the need for re-computing the flow fields.

The prediction of filter efficiency and pressure drop evolution by FilterDict has been extensively validated in a variety of industrial and scientific research projects. The predictions agree with measurements to within 10-20%.

Initial Filter Efficiency - FilterDict determines the filter efficiency and most penetrating particle size (MPPS) of a filter medium by tracking particles through a 3D filter structure. The simulation of the particle movement can include interception, inertial impaction, Brownian motion and electrostatic attraction. Various physical models are available to model the contact of a particle with the filter medium (Caught on first touch, Van der Waals forces / Hamaker constant, Sieving) and allow to model different filtration regimes. User-defined functions allow full control of these model parameters.

Initial Pressure Drop - Using the flow solvers from the FlowDict module, FilterDict determines the initial pressure drop over the filter media. Fast flow rates are modeled with the Navier-Stokes, slow flow rates with Stokes equation.

Filter Life Time and Capacity - FilterDict can simulate the clogging of a filter in a Single Pass or Multi Pass experiment. Pressure drop evolution over time and the dust holding capacity is determined. The simulation covers the full range from depth to cake filtration.

Technical Report

Innovative design, analysis and optimization of woven filter media through experimental and computational methods

Azimian M., Mantler A., Meyer F., Edelmeier F., Becker J., Wiegmann A., 2020: Innovative design, analysis and optimization of woven filter media through experimental and computational methods, Math2Market GmbH & HAVER & BOECKER OHG, technical report, ​https://doi.org/10.30423/report.m2m-2020-02

Read Technical Report


Technical Report

Improved flow simulations on nanofibrous filter media by incorporating the slip effect

Cheng L., Linden S., Azimian M., Wiegmann A., 2019: Improved flow simulations on nanofibrous filter media by incorporating the slip effect, Math2Market GmbH, technical report, ​https://doi.org/10.30423/report.m2m-2019-01

Read Technical Report

Examples of Applications for Filter Media

FilterDict® for pleated filters and complete filters with housing

The ideal filter has a large Dust Holding Capacity (DHC) and high filter efficiency while maintaining a low pressure drop. These three properties usually depend on each other, meaning that improving one or two of them will usually undermine the third property.

With the goal of increasing the DHC, FilterDict simulates the clogging of a filter in a single pass experiment on the unresolved media (such as single or multiple pleats, pleat elements and diesel particulate filter structures)

Filter elements

For filter elements such as pleated filters, FilterDict® determines filter life-time, pressure drop evolution and fractional efficiencies over time, and DHC.
The filter media of the pleats can be a nonwoven or woven filter media. The pleated filters can be optimized and/or newly designed by choosing flat sheet filter media and specifying the geometry and the number of pleats.

The implemented pass-through model is implemented considers the filtration possibility of each type of particle in different types of unresolved porous media.

The simulation covers the entire filtration regime from depth filtration to cake filtration. The simulation with GeoDict is validated by comparing the results with experiments and adjusting the numerical parameters.

Complete Filters with Housing

For complete filters with housing, the geometry of the parts of the complete filter in CAD format (*.stl and/or .obj) is imported and converted into a GeoDict voxel (3d pixel) using the ImportGeo-CAD module. The user defines inlet and outlet for the flow and the characteristics of the filter parts (flowing fluid, solids, porous parts). The characteristics of the filter media at micro scale are also defined (e.g., permeability).

The filter media of the pleats inside the complete filter can be a nonwoven or woven filter media.

The lifetime single-pass simulation determines the changes in pressure drop as a function of time, the fractional filtration efficiencies, and the deposited dust per batch.

Honeycomb Structures

For honeycomb structures, such as in Diesel Particulate Filters (DPF) und Gasoline Particulate Filters (GPF), FilterDict is used to design and optimize the filters for lower pressure drop, higher filter efficiency and longer life time.

The simulation steps ( Azimian et al., 2018) include

  • modeling the ceramic filter media (GrainGeo),
  • simulating the air flow through the filter media (FlowDict) and the transport and deposition of soot particles (FilterDict),
  • the conversion of deposited particles into a porous media, and
  • determining the soot layer's packing density and viscous flow resistivity.

The simulation of gas filtration at the molecular level can be simulated with the AddiDict module.

Examples of Applications for pleated filters and complete filters with housing

Additional modules needed?

ModulesApplications
GeoDict Base includes the basic functionality of GeoDict [necessary]
FlowDictcomputes flow fields as well as the pressure drops or mean velocities [necessary]
ImportGeo-Vol import and segmentation of the µCT images and generation of the 3D microstructure models based on them [optional]
Digitales Materialdesignfor filter media: creating and modeling 3D microstructure models in GeoDict [optional]
PleatGeofor pleat elements: generation of unresolved porous media  [optional]
GridGeofor honeycomb structure models: generation of unresolved porous media  [optional]
ImportGeo-CADfor complete filter with housinge:  Import of STL or OBJ files of the housing and convert them to voxel data for the simulation  [optional]