Hironobu Imakoma

An effective medium treatment of the transport properties of a Voronoi tesselated network

The present work stresses the significance of the effective medium theory in the computation of the macroscopic transport coefficients from the microgeometry of porous media. The porous ‘‘material’’ is simulated as a two‐dimensional network of interconnected slits of irregular shape and a random distribution using the Voronoi–Delaunay tesselation technique. The calculation procedure for the macroscopic transport coefficients is based on two concepts, the first one being the approximation of the microscopic field by a smooth field (SFA), and the second one being the average of the network random parameters to a mean/effective value by the effective medium theory (EMT). For the latter we apply an improved version of the EM equation derived for regular lattices by Kirkpatrick [Rev. Mod. Phys. 45, 4 (1973)]. This equation takes into account the irregularity in the slit‐length distribution and is applicable on both regular and irregular lattices. The EMT/SFA results of the improved version for ordinary diffusi...

Transport properties of porous media from the microgeometry of a three-dimensional voronoi network

In this work a way of calculating effective transport coefficients from the microgeometry of a porous medium is presented. The model material consists of a random packing of uniform spheres, and by applying the Voronoi—Delaunay tessellation technique the void between the spheres is simulated as a network of cylindrical pores. The tessellation yields all the necessary information for the structural characterization, such as the pore diameter, pore angle and pore length distribution functions and the topological interconnection. The effective transport coefficients of ordinary diffusion, Knudsen flow and viscous flow are calculated numerically by mass balancing at each network node and over all nodes of the system. The results obtained agree very well with the experimental ones, especially for ordinary diffusion. For Knudsen and viscous flow, inaccuracies in the estimation of the pore overlapping volume cause a relative error between the numerical and experimental results of the order of 16%–33%.

Principle and Applications of Drying Characteristic Functions

ABSTRACT This article focuses on the concept of a ‘drying characteristic function,’ which is an effective way to correlate drying rate curves for convective drying of homogeneous nonporous, hygroscopic porous and non-hygroscopic porous materials. The characteristic function, obtained bv a certain transformation of a drying rate curve, is independent of drying conditions and hence characterizes the transport kinetics in the material. The principle and some applications of the functions are reviewed. The first application is estimation of dryinge rate curves. Because any drying rate curve can be transformed into the characteristic function and vice versa, the drying rate can be estimated for various drying conditions from a single drying experiment. Another aoolication is determination of the moisture diffusivity. Using the ‘flux ratio method’ an analytical expression of the characteristic function can be obtained for any aiven moisture diffusivity. The exprcssion enables one to determine the moisture diffu...

Approximate isothermal drying curves of hygroscopic porous materials with given desorption isotherm

Abstract The drying curves of hygroscopic porous materials can usually only be predicted by solving the mass transfer equation (isothermal case) numerically with the corresponding limit conditions because of the strong dependence of the apparent diffusion coefficient on the moisture content and the influence of the desorption isotherm on the drying rate. A simple approximate method which gives an analytical expression for the drying curve under isothermal conditions for both drying periods (penetration period and regular regime) for any given diffusion coefficient and desorption isotherm, both defined as a function of the moisture content, is presented here. In this method, the observation that the drying rate normalized by an appropriate steady-state flux hardly depends on the form of the diffusion coefficient is used.

Characterization of porous media by means of the Voronoi-Delaunay tessellation

Abstract The Voronoi-Delaunay tessellation technique for subdividing the space between entities is used in this paper in order to characterize a porous medium. The model material consists of a bundle of cylinders with uniform diameter, randomly arranged so that they are closest to each other. The resulting network consists of quadrilaterals of different width and height and infinite depth. The width is defined as the closest distance between two adjacent cylinders (narrowest neck). On the other hand, the height is defined as the length of a Voronoi polygons edge (distance between two nodes). The height vector lies perpendicular to the width vector. The coordination number, denoting the number of edges meeting at a node, was always found to be 3. The porosity of the packing was 0.47 and of the resulting network 0.41. The simulation consists of about 150 realizations of a model network, which, on average, accommodated about 89 cylinders, 180 nodes and 270 slits. Topologically, the slits do not show any preference for a specific orientation, so that in a large-scale network all possible orientations of the slits will be equally represented. The relation between the slit width and the slit height, of great importance for calculating dynamic properties, is very satisfactorily approximated by a linear expression. A way to predict the saturation-dependent transport coefficient of ordinary diffusion, Knudsen diffusion and viscous flow is proposed. The network showed a percolation threshold at a number based saturation of 0.67, the same as for the honeycomb regular lattice. The tortuosity factor was found to be 2.95.

Estimation of drying rate curves for nonhygroscopic porous slabs using the characteristic function for the regular regime

Abstract A new method to predict convective drying is proposed for slabs of nonhygroscopic porous material with uniformly varying body temperature. The method enables the drying rate and the body temperature to be estimated for various drying conditions from a single experiment on drying. No information about the dependence of the moisture diffusivity on the moisture content is required. The basis of the method is that a period exists during which the relationship between a modified flux parameter and a modified mean moisture content is independent of the drying conditions. This period is defined as the regular regime, and the relationship is called the characteristic function for the regular regime (CFRR). The CFRR, representing the characteristic of the material to be dried, is determinable from a single experiment on convective drying. The present method extracts information for desired drying conditions directly from the CFRR. The practicability of this method has been verified by applying it to experiments on the drying of unglazed alumina.

DETERMINATION METHOD OF VARIABLE DIFFUSION COEFFICIENTS OF HYGROSCOPIC MATERIAL FROM DRYING RATE CURVES OF THE SINGLE PARTICLE

ABSTRACT The concentration dependency of diffusion coefficients of hygroscopic materials can usually only be calculated by cumbersome experimental techniques. Taking the diffusion rate in the fictious steady state with the same mean moisture concentration as in the regular regime of the drying process of a spherical hygroscopic particle (which means the drying period not influenced by initial moisture distributions) into account, a simple method is proposed to estimate the dependency of diffusion coefficients on the moisture concentration for hygroscopic materials from drying rate curves of the single particle.

Solidification by Convective Drying of Particle Layer Wetted with Dilute Agar Gel

Abstract: Solidified porous slab is formed through convective drying of glass particle layer wetted with aqueous dilute agar gel. Measured critical mean moisture content increases with increasing initial moisture or agar content. The agar gel moves in viscous flow caused by capillary pressure during drying. A new drying model based on the receding evaporation plane model is proposed. Drying period is divided into surface and internal evaporation periods. Wet slab consists of dried and wet zones during the internal evaporation period, while the wet slab consists of wet zone only during the surface evaporation period. In the new model, the evaporation rate from the wet zone in the internal drying period is estimated with the linear driving force (LDF) approximation in the field of adsorption engineering. Critical moisture content, that is, mean moisture content between the surface and internal periods, is estimated with a mass balance on the surface. Simulated results by the new drying model with reasonable fitting parameters agree very well with measured drying data.

Measurement of absorbed power of glass particle layer on moisture content and drying rate by combined convective and microwave drying

Abstract: Dependency of absorbed power by microwave on the local moisture content in a glass particle layer was measured with a new method; that is, heating the wet layer. The heating experiment was performed using a laboratory-scale combined convective and microwave heater/dryer that was manufactured by modifying a domestic microwave oven at 2.45 GHz. The measured result was strongly dependent on the local moisture content and showed a maximum and a minimum within the measured range of the moisture content. This dependency can be explained by the assumption that moisture in the wet layer behaves as a mass of the free water. The combined drying rate of the wet layer measured with the heater/dryer was simulated with both the power dependency and the experimental convective-only drying rate. Power dependency on temperature is as important as the moisture content in the simulation. Simulated results agree very well with experimental ones.

Effective medium approximation of 3‐D Voronoi networks

This article deals with the determination of the effective (apparent) diffusivity from the topological characteristics of microscopically inhomogeneous materials. The porous medium is modeled as a three‐dimensional network of interconnected pores by means of the Voronoi–Delaunay tesselation. The calculation procedure is based on an improved version of the effective medium theory and the smooth field approximation. We investigate the case where the pores in the effective network are transformed to pores of a uniform cross‐sectional area, and the case where the pore conductances are all replaced by an effective one. The results are compared with the numerical network solution. It is shown that the effective cross‐section area approach reproduces the ‘‘exact’’ calculated transport coefficient with an error of less than 9%, whereas the effective uniform conductance approach fails considerably, displaying an error of 62%.