M. Ferer

PORE-LEVEL MODELING OF IMMISCIBLE DRAINAGE: VALIDATION IN THE INVASION PERCOLATION AND DLA LIMITS

Motivated by a wide-range of applications from ground water remediation to carbon dioxide sequestration and by difficulties in reconciling experiments with previous modeling, we have developed a pore-level model of two-phase flow in porous media. We have attempted to make our model as physical and as reliable as possible, incorporating both capillary effects and viscous effects. After a detailed discussion of the model, we validate it in the very different limits of zero capillary number and zero-viscosity ratio. Invasion percolation (IP) models the flow in the limit of zero capillary number; results from our model show detailed agreement with results from IP, for small capillary numbers. Diffusion limited aggregation (DLA) models the flow in the limit of zero-viscosity ratio; flow patterns from our model have the same fractal dimension as patterns from DLA for small viscosity ratios.

Patchy cleaning of rigid gas filters - Transient regeneration phenomena : comparison of modelling to experiment

Abstract Rigid ceramic filter media, widely used for the removal of particles from gas streams at elevated temperatures tend to show patchy cleaning when the filter regeneration is incomplete [Filtr. Sep. (1989) 187]. In order to investigate the regeneration behaviour and the operational performance of partially regenerated, rigid gas cleaning filter media over many filtration cycles, experiments were performed in a filter test rig. The regeneration behaviour of the filter sample was characterized by the overall regeneration efficiency, the local frequency of regeneration, and the number and size of regenerated filter areas. Using only four adjustable parameters, our modelling results compare favourably with our experimental results, at room temperature. This favourable comparison of the regeneration behaviour between modelling and experiment is achieved only if it is assumed that cohesive and adhesive bonds, which are broken during filter regeneration, do not heal during the next filtration cycle. Assuming otherwise would cause (i) the dust cake to be removed at the same positions during every regeneration and (ii) the patch size to increase from cycle to cycle instead of decreasing as seen in the experiment. Therefore, the model development was guided by our extensive experimental results. This agreement of modelling with experiment indicates that the modelling has real predictive capabilities for operational filter cleaning. Both filter conditioning and dust cake compression significantly influence the operational performance of partially regenerated filter media.

Analysis of Operational Filtration Data Part III: Re-entrainment and Incomplete Cleaning of Dust Cake

ABSTRACT Operational data for the hot-gas filtration system of the Integrated Gasification and Cleanup Facility (IGCF) at the Federal Energy Technology Center are carefully analyzed. A model that includes re-entrainment of filter-cake fragments following cleaning of the candle filters, as well as incomplete removal of cake from the filters by the cleaning backpulses of compressed air, is developed and used to interpret the data. The parameters of the model are evaluated with a least-square error fit procedure. It is shown that inclusion of the re-entrainment model significantly improves the fit to the data during the short interval (∼50 s after each backpulse) during which re-entrainment is predicted to occur by particle-transport modeling. The re-entrainment/incomplete-cleaning model is used to evaluate the buildup of filter-cake thickness and the filter-cake permeability. Both effects were significant for the data analyzed.

A simple model of the adhesive failure of a layer: Cohesive effects

A fine-scale model is developed for the removal of an adhesive layer by a uniform stress. The initial motivation of this modeling project was a description of the removal of a layer of filter cake from cylindrical filters by backpulse cleaning. The model includes the bonding forces of adhesion between the layer and a substrate, as well as the forces of cohesion between imaginary “gridblocks’’ within the layer. For stresses greater than a threshold value, some of the layer is removed, with the fraction removed depending upon the stress, the average adhesive and cohesive forces, and the distribution of these forces about their average. The cohesive forces reduce the threshold well below the average strength of the adhesive force, because they increase the stress near broken adhesive bonds. The cohesive forces also sharpen the threshold in the cleaning pressure significantly, so that the threshold is very much sharper than the distribution of adhesive strengths. For moderate filter cake thickness (moderately...

Spatial distribution of avalanches in invasion percolation: their role in fingering

For two decades, invasion percolation (IP) has provided a simple model of ‘drainage’ where a non-wetting fluid is injected into a porous media saturated with a wetting fluid, in the limit where capillary forces dominate and viscous forces are negligible. IP produces a characteristic fingering with a fractal dimension close to that of ordinary critical percolation. Avalanches (also called ‘bursts’ or ‘Haines jumps’) have been observed. In this paper, we focus on the practical issues relating to the causes of the fingering and of the low saturations of injected fluid. We show that the saturation and the average position of the injected fluid exhibit standard fractal scaling behavior. However, the fractional flow of the injected fluid does not allow an average analysis because of the noise arising from the avalanches, even for the million site systems investigated in this paper. In studying the spatial distribution of these avalanches, we find a size cutoff depending upon the position of the avalanches; this is characteristic of the finite size of the system and signals that the systems have not achieved self-organized criticality. Furthermore, we show that the average size of these avalanches, 〈sa〉, increases with their average distance, 〈x〉, from the outlet as 〈sa〉≈〈x〉1.1. As a result, larger avalanches will tend to occur at the end of longer fingers causing preferential growth of the long fingers at the expense of the shorter fingers.

A new stereolithography experimental porous flow device

A new method for constructing laboratory-scale porous media with increased pore-level variabilities for two-phase flow experiments is presented here. These devices have been created with stereolithography directly on glass, thus improving the stability of the model created with this precision rapid construction technique. The method of construction and improved parameters are discussed in detail, followed by a brief comparison of two-phase drainage results for air invasion into the water-saturated porous medium. Flow through the model porous medium is shown to substantiate theoretical fractal predictions.

Crossover from fractal to compact flow from simulations of two-phase flow with finite viscosity ratio in two-dimensional porous media

The effect of the viscosity ratio (M\ensuremath{\equiv}

Modeling of Backpulse Filter Cleaning: The Small Particle Filter Cake Fragments

{\mathrm{\ensuremath{\mu}}}_{\mathit{D}}

Critical-scaling theory for ternary systems with polar tielines

/

Critical‐scaling theory of binary excess enthalpies and its use to determine liquid–liquid equilibrium phase compositions

{\mathrm{\ensuremath{\mu}}}_{\mathit{I}}