IT for Engineering (it4e) GmbH Hall 8 / C28

DNSlab is a software for analysis and simulation of mass transport and deposition especially in microporous filter media. The computation of gas or liquid flow fields in 3D models of the media is used for the determination of the permeability, the filter efficiency and filter cake build-up and backwashing, amongst others.

DNSlab provides Finite Differences (FD) and Lattice-Boltzmann (LB) schemes to compute flow fields. The FD schemes are advantageous for slow creeping flows, the LB schemes for faster inertial flows. DNSlab’s FD schemes feature low memory consumption and ultimate numerical stability for creeping flows, as they always converge for any 3D geometry.

The contribution shows the recent FD scheme improvements in DNSlab, which enable the users to run simulations faster with larger, more representative 3D models, and less computing resources. A typical nonwoven model is used to compare the performance of the different FD scheme versions. The simulations are run on a practicable low-cost gaming PC.

The 2024 version has the lowest computing time, the lowest memory requirement and gives the most precise result on the non-coarsened voxel grid, meaning that it performs...

DNSlab is a software for analysis and direct numerical simulation of fluid and particle flows by means of 3D models especially of microscale porous structures.

The new DNSlab Python Interface was introduced with the first DNSlab release in 2024, Release2024a. It enables DNSlab users to automize and customize the pre- and postprocessing of DNSlab applications. The interface is provided in the form of a package of classes and functions which are imported in the user’s python code, and which enable to read and write DNSlab input and output data. The basic working principle is that by the user’s python code, DNSlab runs are prepared, executed and evaluated. In particular for the evaluation, Python’s comprehensive plotting capabilities are applicable.

Numerical simulation of filtration processes is used to investigate and improve filtration characteristics especially on the microscale. Coupling of computational fluid dynamics (CFD) with the discrete element method (DEM) is a simulation approach which allows to model all essential microscale effects which occur in solid-liquid filtration processes. The software DNSlab® provides different numerical methods that can be used for such simulations, namely the Lattice Boltzmann Method (LBM) and Finite Differences (FD) methods for the fluid dynamics, coupled with the DEM for the particle modeling.

In previous research [Puderbach, 2021] the LBM was used to describe the filtration behavior of solid spheres in solid-liquid separation processes using 2-way coupling of fluid and particle interactions. The investigations show good agreement of the experimental and numerical results. The 2-way coupling with the LBM allows a very accurate simulation of the particle motion in the fluid, but due to the 2-way coupling it is a very computationally intensive and time-consuming method.

The new 1½-way coupling approach is a further development by omitting too detailed computations which have only a neglectable effect on the filtration characteristic [Schmidt, 2021]. Unlike as for the 2-way coupling, the CFD for the 1½-way coupling doesn’t have to be done by the LBM, the resource-saving FD method can also be used. The approach still allows to simulate the particle deposition and detachment, especially by backwashing the filter cake, but requires significantly less computing resources.

The following figure shows the comparison of simulation and measurement for the separation of glass particles from water at a metal wire mesh [Puderbach2021