Fraunhofer Institute for Industrial Mathematics ITWM Hall 7 / M15

In many areas of filtration application, woven filters are the preferred media type due to defined filtration properties, durability and mechanical strength. The latter feature is particularly important to counter the deformation caused by the fluid flow through the porous medium during operation.

When searching for the optimal combination of yarn material(s) and weave design (yarn strength, weave pattern) for a given application, an experimental approach using prototypes can become time-consuming and costly. In addition, empirical knowledge is of limited use when new yarn materials or material combinations are to be considered. Suitable simulation techniques can help to significantly accelerate this stage of product development and optimization.

To this end, a specialized software tool is used for both the design of the woven filter fabric and the prediction of its mechanical strength. Known quantities such as the weaving pattern, yarn diameter(s) and the yarn materials – especially their mechanical properties – serve as input data for the computation. Instead of performing 3D simulations for finding properties like tensile strength and flexural rigidity of the fabric, the simulation is sped up tremendously by using highly efficient beam models.

The simulation results are validated by comparing them to experimental data obtained by mechanical testing.

The results are applied to the coupled simulation of the deformation of the filter fabric under stationary flow on the macroscopic scale. On this length scale, a direct numerical simulation that resolves the open areas and the yarn in the grid and uses a fluid-structure interaction applied to the solid yarn would require a lot of computational resources and time. Instead, a multiscale approach as in [1] is taken: By performing CFD simulations on the microscopic length scale (i.e. the scale of the mesh holes and yarns), the flow resistance of the fabric is obtained for different strain rates. This is used for the simulation on the macroscopic scale, where the filter fabric is modelled as a (poro-)elastic shell with the effective flow resistance and mechanical properties obtained from the mesh-scale simulations. Based on the operating conditions (volumetric flow rate) a CFD simulation computes an initial pressure distribution, which represents the mechanical load on the filter fabric. This data is used for simulation of the deformation of the fabric. For the new state, another CFD simulation updates the pressure distribution. This cycle is repeated until a steady state in terms of deformation is reached.

The present paper presents the approach and the results in detail, particularly the improvements compared to previous methods...

The biggest challenge when developing filtering face pieces (FFP) is to guarantee the required level of protection while keeping breathing resistance as low as possible. For instance, the FFP2 classification according to the DIN EN 149 standard requires that at an air flow rate of 95 l/min, particle penetration must not exceed 6% and breathing resistance must not exceed 2.4 mbar [1].

The aim of the research project SULA (“Sicherer und leichter atmen”, engl.: ‘safer and easier breathing’) is to significantly reduce both penetration and breathing resistance by improving the manufacture and processing of the nonwoven fabric. Expertise from industry (Reifenhäuser Reicofil, IMSTec) and application-oriented research (Fraunhofer ITWM) is being pooled to investigate the interaction of the individual processes and their effects on the product.

For a meltblown nonwoven to meet the specifications, the fibers are electrostatically charged. In this project, the focus is on the hydrocharging technology, which is applied already during fiber production and thus promises an even distribution of the electric charge in the nonwoven.

In Part 1 of the contribution, we aim to analyze the impact of various process parameters on the charging characteristics of nonwovens, particularly focusing on the differences between electrostatic charging and hydrocharging, as this is the main driver to reduce the pressure drop resulting in lower breathing resistance for the face mask.

To increase the productivity of the filter layer the Multi-Row process is in favor to the well-established Single-Row process. Nonwovens produced by these methods exhibit distinct permeabilities and charges, affecting the quality of FFP2 masks.

Our goal is to enhance our understanding of the positive effect of hydrocharging in comparison with electrostatic charging through simulation and to adapt the superior performance of the hydrocharged single row process to the multi row process with increased productivity.

The entire meltblown process, including hydrocharging, is complex. Therefore, we are starting with modeling and validating individual aspects of the process. Initially, the water nozzle was modeled in a static airflow, simulated, and its performance validated against the nozzle manufacturer's measurements. The simulation results closely matched the actual measurements, both in terms of droplet diameter distributions and velocities. Subsequently, we modeled the high-turbulence airflow without considering hydrocharging effects and validated this model using measurements of Reifenhäuser Reicofil. Later, we incorporated the interaction between water droplets and the airflow into the model and conducted further validations. We measured water quantities at various positions of the water nozzles behind the air jet and compared these with our simulation results.

In the next steps we want to include all the effects of Meltblown process to get insight into hydrocharging process. With this knowledge we want to optimize ...

The biggest challenge when developing filtering face pieces (FFP) is to guarantee the required level of protection while keeping breathing resistance as low as possible. For instance, the FFP2 classification according to the DIN EN 149 standard requires that at an air flow rate of 95 l/min, particle penetration must not exceed 6% and breathing resistance must not exceed 2.4 mbar [1].

The aim of the research project SULA (“Sicherer und leichter atmen”, engl.: ‘safer and easier breathing’) is to significantly reduce both penetration and breathing resistance by improving the manufacture and processing of the nonwoven fabric. Expertise from industry (Reifenhäuser Reicofil, IMSTec) and application-oriented research (Fraunhofer ITWM) is being pooled to investigate the interaction of the individual processes and their effects on the product.

For a meltblown nonwoven to meet the specifications, the fibers are electrostatically charged. In this project, the focus is on the hydrocharging technology, which is applied already during fiber production and thus promises an even distribution of the electric charge in the nonwoven. The corresponding experimental investigations and the associated modeling and simulations are presented in Part 1 of the contribution [2].

This second part is dedicated to the characterization of the electret media produced and the model-based prediction of their protective effect and breathing resistance. The characterization includes basic properties such as thickness and surface weight. In addition, a workflow is being developed for the automatic detection of the nonwoven’s fiber diameter distribution, based on suitable sets of SEM images. In terms of performance, air permeability measurements as well as standard tests for the filter efficiency are performed for both flat sheet samples and mask prototypes. These experimental investigations are carried out for different aerosols (paraffin oil, sodium chloride) and volumetric air flow rates. The influence of the charging is studied by comparing the filtration efficiency for discharged samples of the filter materials with the charged counterparts. To quantify the effectiveness of the charging process, electrostatic fieldmeter measurements are performed on both sides of the flat sheet samples.

These data form the basis for the identification and calibration of models for the prediction of air flow resistance and fractional efficiency of the materials. The modeling allows for

It is well known that filter media are exposed to compression during the manufacturing process. The transportation via conveyor belt rolls compacts the medium over the entire surface, changing the filter performance, i.e. efficiency, pressure drop and dust holding capacity. Further processing steps such as pleating compress the filter media locally at the tops and bottoms. The nonuniform permeability leads to a reorientation of the flow field so that the filter surface is not utilized optimally. Experimental measurement series are very helpful for the design and optimization of the manufacturing process and the nonwoven fabric. Nonetheless, they are associated with large efforts in time and cost.

This work provides an overview of the influence of compression on the filter performance of flat filter media. The pressure drop across the media as well as the fractional efficiency are predicted by models. The approach is based solely on measured material characteristics of the uncompressed nonwoven and requires no additional experimental effort for the compressed material (see Figure 1).

The data basis for the analysis of pressure drop is provided by five nonwovens, which differ by their fields of application and media structure to ensure a broad validity of the presented model. SEM images as well as measurements of the thickness, the area weight and the correlation between pressure and velocity characterize the uncompressed media. The prediction of the pressure drop for multiple compression levels is based on the Darcy-Forchheimer law [1] whereby relevant model parameters (permeability, inertial loss coefficient) are scaled with suitable models [2] referring to the change of material thickness and the corresponding reduction of porosity.

Based on one of the five filter media, which is an electret filter, the experiments are extended to include the fractional efficiency. Single fiber efficiency (SFE) models [3] are a reasonable choice for modeling porous media that consist of fibers. The mechanisms of diffusion, direct interception, inertial impaction, electrophoresis and dielectrophoresis are distinguished. Numerous models are available in the literature (see e.g. [4]), although the accuracy of the models can vary depending on the filter medium or moreover the manufacturing process. A suitable combination of these models paired with weighting factors describes the initial efficiency and can be extended by considering only the changes of material thickness and porosity.

The proposed methods allow for ...

Stainless steel wire meshes are known for their mechanical strength, corrosion resilience and thermal stability. These properties make them ideal for cleaning highly viscous fluids such as polymer melts. The latter are non-Newtonian and therefore, when simulating the flow through a wire mesh filter, the shear-thinning behavior must be considered properly to obtain accurate results for the pressure drop and filter efficiency.

As an example, the present talk studies the flow of polypropylene (PP) melt through wire meshes under the conditions during the production of meltblown filter media. A two-step multiscale approach is used to describe the non-Newtonian flow phenomenon in polymer melt flows both at the micro- and macroscopic scales. To reconstruct the microgeometry of the filter element, a parametric solid representation of the individual wires and their contact regions [1] is used.

On the macroscopic scale of the filter element, a direct numerical simulation of the filter weave, i.e. resolving all the pore spaces in the computational grid, would be very costly in terms of computational effort. Instead, a representative elementary volume (REV) of the weave is chosen and represented in a computational grid with sufficiently fine resolution. Based on the non-Newtonian viscosity law w.r.t. temperature and shear rate and experimental data, CFD simulations [2] are carried out for different temperatures and flow rates.

The results are used to obtain the effective flow resistance of the wire mesh which allows for an upscaling of the simulation to the macroscale. In addition, the velocity field of a Newtonian fluid which causes the same pressure drop for the given flow rate is compared with the local velocities of the melt. The difference is relevant for micro-scale simulations of the transport and deposition of solid particles.

The talk also presents several examples for the multiscale simulation of PP melt flow in wire mesh filters and discusses the results...