In this paper, a previous experimental investigation on physical refining of steel melts by filtration was numerically studied. To be specific, the filtration of non-metallic alumina inclusions, in the size range of 1-100 & mu;m, was stimulated from steel melt using a square-celled monolithic alumina filter. Computational fluid dynamics (CFD) studies, including simulations of both fluid flow and particle tracing using the one-way coupling method, were conducted. The CFD predicted results for particles in the size range of & LE;5 & mu;m were compared to the published experimental data. The modeled filtration setup could capture 100% of the particles larger than 50 & mu;m. The percentage of the filtered particles decreased from 98% to 0% in the particle size range from 50 & mu;m to 1 & mu;m.

Ceramic foam filters (CFFs) are used to remove inclusions and/or solid particles from molten metal. In general, the molten metal poured on the top of a CFF should reach a certain height to form the pressure (metal head) required to prime the filter. For estimating the required metal head and obtaining the permeability coefficients of the CFFs, permeability experiments are essential. Recently, electromagnetic priming and filtration of molten aluminum with low and high grades of CFF, i.e., 30, 50 and 80 pore per inch (PPI) CFFs, have been introduced. Since then, there has been interest in exploring the possibility of obtaining further inclusion entrapment and aluminum refinement by using electromagnetic force to prime and filter with stacked CFFs. The successful execution of such trials requires a profound understanding concerning the permeability parameters of the stacked filters. Such data were deemed not to exist prior to this study. As a result, this study presents experimental findings of permeability measurements for stacks of three 30, three 50 and three 80 PPI commercial alumina CFFs from different industrial batches and compares the findings to numerically modelled data as well as previous research works. Both experimental and numerical findings showed a good agreement with previous results. The deviation between the experimentally and numerically obtained data lies in the range of 0.4 to 6.3%.

Ceramic foam filters (CFFs) are used to remove solid particles and inclusions from molten metal. In general, molten metal which is poured on the top of a CFF needs to reach a certain height to build the required pressure (metal head) to prime the filter. To estimate the required metal head, it is necessary to obtain permeability coefficients using permeametry experiments. It has been mentioned in the literature that to avoid fluid bypassing, during permeametry, samples need to be sealed. However, the effect of fluid bypassing on the experimentally obtained pressure gradients seems not to be explored. Therefore, in this research, the focus was on studying the effect of fluid bypassing on the experimentally obtained pressure gradients as well as the empirically obtained Darcy and non-Darcy permeability coefficients. Specifically, the aim of the research was to investigate the effect of fluid bypassing on the liquid permeability of 30, 50, and 80 pores per inch (PPI) commercial alumina CFFs. In addition, the experimental data were compared to the numerically modeled findings. Both studies showed that no sealing results in extremely poor estimates of the pressure gradients and Darcy and non-Darcy permeability coefficients for all studied filters. The average deviations between the pressure gradients of the sealed and unsealed 30, 50, and 80 PPI samples were calculated to be 57.2, 56.8, and 61.3 pct. The deviations between the Darcy coefficients of the sealed and unsealed 30, 50, and 80 PPI samples found to be 9, 20, and 31 pct. The deviations between the non-Darcy coefficients of the sealed and unsealed 30, 50, and 80 PPI samples were calculated to be 59, 58, and 63 pct.

This paper presents experimental results of pressure drop measurements on 30, 50, and 80 pores per inch (PPI) commercial alumina ceramic foam filters (CFF) and compares the obtained pressure drop profiles to numerically modeled values. In addition, it is aimed at investigating the adequacy of the mathematical correlations used in the analytical and the computational fluid dynamics (CFD) simulations. It is shown that the widely used correlations for predicting pressure drop in porous media continuously under-predict the experimentally obtained pressure drop profiles. For analytical predictions, the negative deviations from the experimentally obtained pressure drop using the unmodified Ergun and Dietrich equations could be as high as 95 and 74 pct, respectively. For the CFD predictions, the deviation to experimental results is in the range of 84.3 to 88.5 pct depending on filter PPI. Better results can be achieved by applying the Forchheimer second-order drag term instead of the Brinkman-Forchheimer drag term. Thus, the final deviation of the CFD model estimates lie in the range of 0.3 to 5.5 pct compared to the measured values.

Ceramic Foam Filters (CFFs) are widely used to filter solid particles and inclusions from molten metal in metal production, particularly in the aluminum industry. In general, the molten metal is poured on the top of a ceramic foam filter until it reaches a certain height, also known as metal head or gravity head. This is done to build the required pressure to prime the filter media and to initiate filtration. To predict the required metal head, it is necessary to obtain the Darcy and non-Darcy permeability coefficients of the filter. The coefficients vary upon filter type. Here, it is common to classify CFFs based on grades or pore per inches (PPI). These CFFs range from10 to100 PPI and their properties vary in everything from cell and window size to strut size. The 80-100 PPI CFFs are generally not practical for use by industry, since the priming of the filters by a gravitational force requires an excessive metal head. However, recently a new method has been developed to prime such filters by using electromagnetic Lorentz forces. This allows the filters to be primed at a low metal head.

To continue the research work, it was deemed necessary to measure the pressure gradients of single and stack of commercial alumina ceramic foam filters and to obtain the permeability characteristics. Therefore, efforts have been made to validate the previously obtained results, to improve the permeametry experimental setup, and to obtain Darcy and non-Darcy permeability coefficients of single 30, 50, and 80 PPI filters and stacks of filters. Furthermore, the experimentally obtained pressure gradients were analyzed and compered to the mathematically and analytically estimated pressure gradients.

The studies showed that, in permeametry experiments, the sample sealing procedure plays an important role for an accurate estimation of the permeability constants. An inadequate sealing or an un-sealed sample results in an underestimation of the pressure drop, which causes a considerable error in the obtained Darcy and non-Darcy permeability coefficients. Meanwhile, the results from the single filter experiments showed that the permeability values of the similar PPI filters are not identical. However, the stacks of three identical filters gave substantially the same measured pressure drop values and roughly the same Darcy and non-Darcy coefficients as for the single filters.

The permeability coefficients of the filters are believed to be best defined and calculated by using the Forchheimer equation. The well-known and widely used Ergun and Dietrich equations cannot correctly predict the pressure drop unless a correction factor is introduced. The accuracy of the mathematically estimated pressure drop, using COMSOL Multiphysics® 5.1, found to be dependent on the drag term used in the Brinkman-Forchheimer equation.  Unacceptable error, as high as 84 to 89 percent for the 30, 50 and 80 PPI single filters, compared to the experimentally obtained pressure gradient values were observed when the literature defined Brinkman-Forchheimer drag term was used. However, when the same second order drag term (containing the non-Darcy coefficient) as defined in the Forchheimer equation was used, the predicted pressure gradient profiles satisfactorily agreed with the experiment data with as little as 0.3 to 5.5 percent deviations for the 30, 50 and 80 PPI single filters.

CFFs are used to filter liquid metal in the aluminum industry. CFFs are classified in grades or pores per inch (PPI), ranging from 10-100 PPI. Their properties vary in everything from pore and strut size to window size. CFFs of 80-100 PPI are generally not practical for use by industry, as priming of the filters by gravitational forces requires an excessive metal head. Recently, co-authors have invented a method to prime such filters using electromagnetic Lorentz forces, thus allowing filters to be primed with a low metal head. In the continuation of this research work, an improved experimental setup was developed in the present study to validate previous results and to measure the permeability of different filters, as well as a stack of filters. The study of permeability facilitates estimation of the required pressure drop to prime the filters and the head required to generate a given casting rate.

CFF s are used to filter liquid metal in the aluminum industry. CFFs are classified in grades or pores per inch (PPI), ranging froml0-100 PPI. Their properties vary in everything from pore and strut size to window size. CFFs of 80-100 PPI are generally not practical for use by industry, as priming of the filters by gravitational forces requires an excessive metal head. Recently, co-authors have invented a method to prime such filters using electromagnetic Lorentz forces, thus allowing filters to be primed with a low metal head. In the continuation of this research work, an improved experimental setup was developed in the present study to validate previous results and to measure the permeability of different filters, as well as a stack of filters. The study of permeability facilitates estimation of the required pressure drop to prime the filters and the head required to generate a given casting rate.

Electromagnetic fields can influence the behavior of liquid metals in commercial Ceramic Foam Filters (CFF's). In the present study 9 inch industrial CFF's of high grade with 50 and 80 pores per inch (ppi) have been investigated. The main objective was to prime the 9 inch industrial scale CFF's with a standard aluminum casting alloy (3XXX - alloy) by the use of various magnetic field strengths (max. 0.12 T) induced by a coil. The obtained results were compared with reference gravity experiments. The influence of the electromagnetic Lorentz forces on the obtained results was calculated with 2D Finite Element Modeling (FEM) using the COMSOL® software. The fluid flow characteristics inside the CFF were considered and are part of another publication within the group.

Electromagnetic fields can influence the behavior of liquid metals in commercial Ceramic Foam Filters (CFF's). In the present study 9 inch industrial CFF's of high grade with 50 and 80 pores per inch (ppi) have been investigated. The main objective was to prime the 9 inch industrial scale CFF's with a standard aluminum casting alloy (3XXX-alloy) by the use of various magnetic field strengths (max. 0.12 T) induced by a coil. The obtained results were compared with reference gravity experiments. The influence of the electromagnetic Lorentz forces on the obtained results was calculated with 2D Finite Element Modeling (FEM) using the COMSOL® software. The fluid flow characteristics inside the CFF were considered and are part of another publication within the group.

Particle concentration and size distribution in the melt can give important information regarding the filtration efficiency and the quality of the aluminium. LiMCA (Liquid Metal Cleanliness Analyser) system, used in primary and secondary production of aluminium, provides in-situ data for granulometric and total density information on the inclusion content, but has problems quantifying particles < 20 μm in size. To be able to determine the required cleanliness with particle counts down' to 10 μm for modern alloys a novel method for automated quantitative results has been developed. Results are obtained using a user friendly technique based on the ImagePro® Plus 7.0 software. The different image processing steps adopted for automated quantification of the particle count in a size range from 2 to 50 μm is described and discussed together with the obtained results. The automated technique has been benchmarked elsewhere with a manual particle count reviling an error of ∼3% on the overall filtration efficiency.