H. Henein

Experimental study of transverse bed motion in rotary kilns

Slumping and rolling beds have been studied extensively in a continuous pilot kiln and batch rotary cylinders. Solids investigated include nickel oxide pellets, limestone, sand, and gravel. The effect of variables such as rotational speed, bed depth, cylinder diameter, particle size, and particle shape on bed motion has been determined. For a given material, the different modes of bed motion can be delineated conveniently on a Bed Behavior Diagram which is a plot of bed depthvs rotational speed. The scaling of bed behavior with respect to particle size and cylinder diameter requires similarity of Froude number modified by(D/dp)1/2, and pct fill. Measurements of key variables characterizing slumping and rolling beds have also been made.

The modeling of transverse solids motion in rotary kilns

Mathematical models have been developed to predict the conditions giving rise to the different forms of transverse bed motion in a rotary cylinder: slumping, rolling, slipping, cascading, cataracting, and centrifuging. Model predictions of the boundaries between these modes of bed motion compare well with previously reported measurements, and can be represented conveniently on a Bed Behavior Diagram which is a plot of pct fill against Froude number (or bed depthvs rotational speed). The location of the boundaries is shown to depend on material variables which characterize frictional conditions in the bed. For the slumping/rolling boundary these are primarily the shear angle and the limiting wedge angle which defines the solids involved in a slump. For the slipping/slumping and slipping/rolling boundaries the governing material variables are the bed/wall friction angle and the upper angle of repose and dynamic angle of repose, respectively. Similarly, the location of the other boundaries related to cascading and cataracting is determined by the dynamic angle of repose. Complete Bed Behavior Diagrams have been calculated for solids having different particle size and particle shape rotated in cylinders having different diameters.

Finite difference heat-transfer modeling for continuous casting

A number of solution strategies for heat-flow models of continuous casting processes are developed and compared for interfacing with optimization algorithms. These two-dimensional (2-D) slice models include nonlinear thermodynamic and transport properties. As a result of the compari-son, a number of modifications are applied to enhance the accuracy of the simulation as well as the efficiency of the solution. Here, we found that by applying the Kirchhoff transformation to the heat-flow equations and by iterative adjustment of temperature-dependent properties, fewer calculations are required per time-step and larger time-steps can be taken. Consequently, this leads to approximately a twentyfold reduction in computation time when using 2-D slice models to determine three-dimensional, steady-state temperature fields. As a result, a complex, 2-D heat-flow model for a cast strand can be solved in 2 to 5 minutes on a Micro VAX II and is thus suitable for incorporation into a systematic optimization procedure and for process simu-lation in real time.

An analysis of radial segregation for different sized spherical solids in rotary cylinders

Rotary reactors such as ball mills, driers, mixers, blenders, and kilns are extensively used in chemical and metallurgical industries to process billions of tons/year of granular solids. Longitudinal and radial segregation of the feed and/or product due to size, shape, and density differences of the feed or product are often encountered in these unit operations. This paper presents some observations of radial segregation due to size differences of spherical solids for slumping, rolling, cascading, cataracting, and centrifuging beds. The mechanism by which spherical beads segregate radially is a combination of percolation and flow. The kinetics of radial segregation for rolling beds is observed and quantified using high speed photography. The rate of segregation is shown to be a function of the cylinder rpm but is independent of bed depth. Furthermore, the larger the size ratio of coarse to fines, the faster the rate of segregation. The scale-up of the segregation process is investigated using two cylinders, 0.2 m ID × 0.2 m L and 0.4 m ID × 0.4 m L. Both the size ratio of the solids and the Froude number are the scale-up criteria for the rate of radial segregation. The former is related to the packing characteristics of loosely packed beds when rolling and to the effective radius of rotation of particles when cataracting or centrifuging.

Single fluid atomization through the application of impulses to a melt

Abstract Impulse Atomization (IA) is a single fluid atomization technique that is capable of producing droplets of controlled size having a relatively tight distribution and a predictable cooling rate. The process has been successfully employed to produce a wide range of metal droplets including Pb–Sn alloys, aluminum alloys, copper alloys, low carbon steel and tool steel. Atomization characteristics determined from load cell measurements, video imaging and particle size analysis are discussed as a function of process characteristics. It is shown that atomization occurs by Raleigh instability and that only primary atomization of the stream is in effect. The rate of cooling of a moving molten droplet has been modeled and experimentally validated using this atomization technique. The droplet Nusselt number ranges nearly from 2 to 10, indicating that thermal conduction from the droplet to the gas is an important mechanism by which the droplet loses heat. Comparison of droplet microstructure of IA and gas atomized powders reveals that for the same size powder of the identical alloy, IA generates a finer microstructure or solidifies with a higher cooling rate. This is attributed to two-way thermal coupling (between gas and melt spray) in gas atomization being greater than in IA. These atomization and heat flow characteristics clearly demonstrate a number of unique features of this technique as well as its flexibility to meet different processing requirements for production and research.

Functionally graded metal/ceramic composites by tape casting, lamination and infiltration

Abstract Alloying aluminum with magnesium and a the use of a N2 atmosphere led to the spontaneous infiltration of ZrO2 preforms with a graded porous structure and the successful fabrication of functionally graded Al–Mg/ZrO2 components. The infiltration process was controlled by an incubation phenomenon resulting from the time needed to destabilize an oxide film present on the molten aluminum droplet. Oxide destabilization depended critically on the presence of Mg, a N2 gas atmosphere and preform microstructure. It is proposed that the destabilization process occurs by a complex reaction localized at the triple point formed between the liquid, solid and gas. The driving force for oxide removal depends on the amount of triple point area present at the infiltration front. A simple model is developed which indicates that this triple point area increases with a decrease in the scale of porosity present in the preform. The model is used to explain the dependence of the incubation time on preform microstructure.

An experimental study of segregation in rotary kilns

An experimental study has been conducted to elucidate the mechanism of radial segregation in the bed of a rotary kiln and to determine the size and composition of the segregated core. Bed-behavior diagrams were obtained, and slumping and rolling characteristics were measured, for two sand mixtures and one limestone mixture containing fines. The fines resulted in significant changes in bed behavior, but no effects were observed on either static or dynamic angle of repose, shear angle, slumping frequency, or active-layer thickness. Therefore it was concluded that the fines do not segregate according to the differential flow of particles down the surface of the bed. Instead percolation, in which the fines pass through the voids in the coarser solids, appears to be the operative mechanism. The size of the segregated core was measured by sampling the bed of solids. Predictions of core width based on geometric considerations and a core composition having the minimum void fraction agreed reasonably well with measurements. A second region of segregation was found adjacent to the wall. The two zones of segregation can influence deleteriously kiln performance. Solids in the central core are not exposed to the hot freeboard gases and therefore may only partially react. The zone of fines at the kiln wall can reduce wall/bed friction to the extent that the bed slips against the wall.

Solidification Study of Aluminum Alloys using Impulse Atomization: Part I: Heat Transfer Analysis of an Atomized Droplet

Abstract Heat transfer models of molten metal droplets moving in a gas stream are used extensively to understand and improve gas atomization systems. In particular, the solidification microstructure of the metal droplets produced during atomization is closely linked with heat flow conditions. The cornerstone of these models is the calculation of the heat exchange between the droplet and gas in an environment with a high temperature gradient. To achieve this, the value of the effective heat transfer coefficient (between the gas and droplet) used in these models is obtained from semi-empirical correlations such as the Ranz- Marshall or Whitaker equations. Unfortunately, most metal atomizing conditions lie outside the experimental envelope in which these correlations were derived. Hence, the object of this paper is two fold: firstly, to develop a reliable and controlled experimental technique by which the transfer of heat from a high temperature droplet to a significantly cooler gas can be assessed and secondly, to determine the validity of both the Ranz-Marshall and Whitaker correlations under these conditions. An experimental technique was developed to conduct a series of quench tests using AA6061 aluminum and AZ91D magnesium droplets falling in a cool nitrogen and argon atmosphere, respectively. A heat transfer model was formulated to account for large droplet gas temperature gradients typically found in metallurgical processing operations. It was determined that a modified Whitaker correlation provided the best agreement with the experimental data given that the Reynolds and Prandtl numbers were evaluated at the free stream gas temperature and the gas conductivity in the Nusselt number at the droplet surface temperature. Les modèles de transfert de chaleur de gouttelettes de métal fondu se déplaçant dans un courant de gaz sont amplement utilisés pour comprendre et améliorer les systèmes de pulvérisation au gaz. En particulier, la microstructure de solidification de gouttelettes de métal produites lors de la pulvérisation est liée de près aux conditions d’écoulement de chaleur. La pierre angulaire de ces modèles est le calcul d’échange de chaleur entre la gouttelette et le gaz dans un environnement à gradient élevé de température. Pour obtenir cela, la valeur du coefficient actuel de transfert de chaleur (entre le gaz et la gouttelette) utilisée dans ces modèles est obtenue à partir de corrélations semi-empiriques telles que les équations de Ranz-Marshall ou de Whitaker. Malheureusement, la plupart des conditions de pulvérisation de métal se trouvent en dehors de l’enveloppe expérimentale dans laquelle ces corrélations ont été dérivées. Par conséquent, ce document a deux objectifs: premièrement, développer une technique expérimentale fiable et contrôlée par laquelle le transfert de chaleur à partir d’une gouttelette à haute température vers un gaz significativement plus froid peut être établi et, deuxièmement, déterminer la validité tant des corrélations de Ranz-Marshall que de Whitaker sous ces conditions. On a développé une technique expérimentale afin de conduire une série de tests de trempe en utilisant des gouttelettes d’aluminium AA6061 et de magnésium AZ91D tombant dans une atmosphère froide d’azote et d’argon, respectivement. On a formulé un modèle de transfert de chaleur en tenant compte des gradients étendus de température gouttelette-gaz que l’on trouve typiquement dans les opérations de traitement métallurgique. On a déterminé qu’une corrélation de Whitaker modifiée fournissait le meilleur accord avec les données expérimentales étant donné que les nombres de Reynolds et de Prandtl ont été évalués à la température du gaz d’écoulement libre et la conductivité du gaz du nombre de Nusselt, à la température de surface de la gouttelette.

Solidification Study of Aluminum Alloys Using Impulse Atomization: Part ii. Effect of Cooling Rate on Microstructure

Abstract A single fluid atomization technique (Impulse Atomization) was used to determine the effect of solidification time on the development of a solidification microstructure for a nominal Al-4.5%Cu alloy, AA6061 and AA6111 aluminum alloys. The advantages of this atomization technique for studying solidification phenomena in metals are the small sample size (diameter of atomized droplets is in the 300 – 1000 μm range) and the predictable heat transfer conditions which can be achieved with this process. By varying the size of an individual droplet produced by the atomizer and the type of cooling gas atomization, solidification times ranging from 0.033 seconds to 0.68 seconds were obtained in this study. The cell spacing of the solidified metal droplets was measured and an equation relating this microstructural characteristic to solidification time was developed for each alloy. The microstructure/solidification time relations were compared with published literature values and the validity of this technique for studying a solidification microstructure was discussed. The experimentally determined relations obtained in this study show good agreement with theoretical equations relating dendrite arm spacing with solidification time. On a utilisé une technique de pulvérisation à fluide unique (pulvérisation à impulsion) pour déterminer l’effet de la durée de solidification sur le développement de la microstructure de solidification d’un alliage nominal d’Al-4.5% Cu et des alliages d’aluminium AA6061 et AA6111. Les avantages de cette technique de pulvérisation pour l’étude des phénomènes de solidification de métaux sont la petite taille de l’échantillon (le diamètre des gouttelettes pulvérisées est de l’ordre de 300 à 1000 μm) et les conditions prévisibles de transfert de chaleur que l’on peut obtenir avec ce procédé. Dans cette étude, on a obtenu des durées de solidification variant de 0.033 seconde à 0.68 seconde, en variant la taille d’une gouttelette individuelle produite par le pulvérisateur et le type de gaz de refroidissement de la pulvérisation. On a mesuré l’espacement de cellule des gouttelettes solidifiées de métal et on a développé, pour chaque alliage, une équation reliant cette caractéristique de la microstructure à la durée de solidification. On compare les relations microstructure/durée de solidification avec les valeurs publiées de la littérature et on discute de la validité de cette technique pour l’étude de la microstructure de solidification. Les relations déterminées expérimentalement obtenues dans cette étude montrent un bon accord avec les équations théoriques reliant l’espacement des branches de dendrite avec la durée de solidification.

Modelling the leaching kinetics of a sphalerite concentrate size distribution in ferric chloride solution

Abstract A phenomenological rate equation for the dissolution of particle assemblages is used to model the batch leaching behavior of multisize, sphalerite flotation concentrates in FeCl 3 solutions. Two concentrates have been studied. The analysis is basically the same for the two, except for the rate constant dependence on particle size. The model used is simply an extended surface-limiting expression, integrated over a number of particle size classes, with allowance made for size dependence of ore grade. Extensive experimental verification of the models was accomplished through leaching experiments on material containing from two to fifteen discrete size fractions combined as equal weights, normal and bimodal size distributions. Results indicate that a reasonable prediction of the leaching rate may be obtained through the use of an extension to the surface-limiting model and that reliance on more complex models, for some applications, may not be necessary.