Heat transfer coefficients gas-solid

Investigator Type of correlation Phases involved Model associated Kunii and Levenspiel [2] Particle-to-gas heat transfer coefficient Gas-solid Gas in plug flow through the bed... 

Investigator Type of correlation Phases involved Region associated Rowe and Claxton [82] Gas-to-particle heat transfer coefficient Gas-solid Central spouting region... 

Figure 14 Heat transfer variations with gas velocity and solids concentration, (a) Dependence of heat transfer coefficient on gas velocity for three types of particles. (From Molerus, 1992), (b) dependence of heat transfer coefficient on solids concentration in a CFB. (From Shi et al., 1998.)... 

A pseudo-convective heat-transfer operation is one in which the heating gas (generally air) is passed over a bed of solids. Its nse is almost exchisively limited to drying operations (see Sec. 12, tray and shelf dryers). The operation, sometimes termed direct, is more aldu to the coudnctive mechanism. For this operation, Tsao and Wheelock [Chem. Eng., 74(13), 201 (1967)] predict the heat-transfer coefficient when radiative and conductive effects are absent by... 

One manner in which size may be computed, for estimating purposes, is by employing a volumetric heat-transfer concept as used for rotary diyers. It it is assumed that contacting efficiency is in the same order as that provided by efficient lifters in a rotaiy dryer and that the velocity difference between gas and solids controls, Eq. (12-52) may be employed to estimate a volumetric heat-transfer coefficient. By assuming a duct diameter of 0.3 m (D) and a gas velocity of 23 m/s, if the solids velocity is taken as 80 percent of this speed, the velocity difference between the two would be 4.6 m/s. If the exit gas has a density of 1 kg/m, the relative mass flow rate of the gas G becomes 4.8 kg/(s m the volumetric heat-transfer coefficient is 2235 J/(m s K). This is not far different from many coefficients found in commercial installations however, it is usually not possible to predict accurately the acdual difference in velocity between gas and soRds. Furthermore, the coefficient is influenced by the sohds-to-gas loading and particle size, which control the total solids surface exposed to the gas. Therefore, the figure given is only an approximation. 

The value of tire heat transfer coefficient of die gas is dependent on die rate of flow of the gas, and on whether the gas is in streamline or turbulent flow. This factor depends on the flow rate of tire gas and on physical properties of the gas, namely the density and viscosity. In the application of models of chemical reactors in which gas-solid reactions are caiTied out, it is useful to define a dimensionless number criterion which can be used to determine the state of flow of the gas no matter what the physical dimensions of the reactor and its solid content. Such a criterion which is used is the Reynolds number of the gas. For example, the characteristic length in tire definition of this number when a gas is flowing along a mbe is the diameter of the tube. The value of the Reynolds number when the gas is in streamline, or linear flow, is less than about 2000, and above this number the gas is in mrbulent flow. For the flow... 

The above equations for heat transfer apply when there is no heat generation or absorption during the reaction, and the temperature difference between the solid and the gas phase can be simply defined tliroughout the reaction by a single value. Normally this is not the case, and due to the heat of the reaction(s) which occur tlrere will be a change in the average temperature with time. Furthermore, in tire case where a chemical reaction, such as the reduction of an oxide, occurs during the ascent of tire gas in the reactor, the heat transfer coefficient of the gas will vary with tire composition of tire gas phase. 

Heat transfer between gas and sohds is exceedingly hard to measure because it is so rapid. Although the coefficient is low, the available surface area and the relative specific heat of solid to gas are so large that temperature equilibration occurs almost instantaneously. Experiments on injection of argon plasmas into fluidized beds have shown quenching rates of up to fifty million degrees Kelvin per second. Thus, in a properly designed bed, gas to solids heat transfer is not normally a matter of concern. 

An important mixing operation involves bringing different molecular species together to obtain a chemical reaction. The components may be miscible liquids, immiscible liquids, solid particles and a liquid, a gas and a liquid, a gas and solid particles, or two gases. In some cases, temperature differences exist between an equipment surface and the bulk fluid, or between the suspended particles and the continuous phase fluid. The same mechanisms that enhance mass transfer by reducing the film thickness are used to promote heat transfer by increasing the temperature gradient in the film. These mechanisms are bulk flow, eddy diffusion, and molecular diffusion. The performance of equipment in which heat transfer occurs is expressed in terms of forced convective heat transfer coefficients. 

Wall-to-bed heat-transfer coefficients were also measured by Viswanathan et al. (V6). The bed diameter was 2 in. and the media used were air, water, and quartz particles of 0.649- and 0.928-mm mean diameter. All experiments were carried out with constant bed height, whereas the amount of solid particles as well as the gas and liquid flow rates were varied. The results are presented in that paper as plots of heat-transfer coefficient versus the ratio between mass flow rate of gas and mass flow rate of liquid. The heat-transfer coefficient increased sharply to a maximum value, which was reached for relatively low gas-liquid ratios, and further increase of the ratio led to a reduction of the heat-transfer coefficient. It was also observed that the maximum value of the heat-transfer coefficient depends on the amount of solid particles in the column. Thus, for 0.928-mm particles, the maximum value of the heat-transfer coefficient obtained in experiments with 750-gm solids was approximately 40% higher than those obtained in experiments with 250- and 1250-gm solids. 

Increasing system temperature causes hgc to decrease slightly because increasing temperature causes gas density to decrease. The thermal conductivity of the gas also increases with temperature. This causes h to increase because the solids are more effective in transferring heat to a surface. Because hgc dominates for large particles, the overall heat transfer coefficient decreases with increasing temperature. For small particles where dominates, h increases with increasing temperature. 

In contrast to the strong effect of gas properties, it has been found that the thermal properties of the solid particles have relatively small effect on the heat transfer coefficient in bubbling fluidized beds. This appears to be counter-intuitive since much of the thermal transport process at the submerged heat transfer surface is presumed to be associated with contact between solid particles and the heat transfer surface. Nevertheless, experimental measurements such as those of Ziegler et al. (1964) indicate that the heat transfer coefficient was essentially independent of particle thermal conductivity and varied only mildly with particle heat capacity. These investigators measured heat transfer coefficients in bubbling fluidized beds of different metallic particles which had essentially the same solid density but varied in thermal conductivity by a factor of nine and in heat capacity by a factor of two. 

Temperature of the fluidized bed is another parameter that could influence the heat transfer coefficient. Increasing bed temperature affects not only the physical properties of the gas and solid phases, but also increases radiative heat transfer. Yoshida et al. (1974) obtained measurements up to 1100°C for bubbling beds of aluminum oxide particles with 180 pm diameter. Their results, shown in Fig. 6, indicate an increase of over 100% in the heat transfer coefficient as the bed temperature increased from 500 to 1000°C. Very similar results were reported by Ozkaynak et al. (1983) who obtained measurements for bubbling beds of sand particles (dp = 1030 pm) at temperatures up to 800°C. 

The first type of model considers the heat transfer surface to be contacted alternately by gas bubbles and packets of closely packed particles. This leads to a surface renewal process whereby heat transfer occurs primarily by transient conduction between the heat transfer surface and the particle packets during their time of residence at the surface. Mickley and Fairbanks (1955) provided the first analysis of this renewal mechanism. Treating the particle packet as a pseudo-homogeneous medium with solid volume fraction, e, and thermal conductivity (kpa), they solved the transient conduction equation to obtain the following expression for the average heat transfer coefficient due to particle packets,... 

The interaction of parametric effects of solid mass flux and axial location is illustrated by the data of Dou et al. (1991), shown in Fig. 19. These authors measured the heat transfer coefficient on the surface of a vertical tube suspended within the fast fluidized bed at different elevations. The data of Fig. 19 show that for a given size particle, at a given superficial gas velocity, the heat transfer coefficient consistently decreases with elevation along the bed for any given solid mass flux Gs. At a given elevation position, the heat transfer coefficient consistently increases with increasing solid mass flux at the highest elevation of 6.5 m, where hydrodynamic conditions are most likely to be fully developed, it is seen that the heat transfer coefficient increases by approximately 50% as Gv increased from 30 to 50 kg/rrfs. 

The data of Fig. 20 also point out an interesting phenomenon—while the heat transfer coefficients at bed wall and bed centerline both correlate with suspension density, their correlations are quantitatively different. This strongly suggests that the cross-sectional solid concentration is an important, but not primary parameter. Dou et al. speculated that the difference may be attributed to variations in the local solid concentration across the diameter of the fast fluidized bed. They show that when the cross-sectional averaged density is modified by an empirical radial distribution to obtain local suspension densities, the heat transfer coefficient indeed than correlates as a single function with local suspension density. This is shown in Fig. 21 where the two sets of data for different radial positions now correlate as a single function with local mixture density. The conclusion is That the convective heat transfer coefficient for surfaces in a fast fluidized bed is determined primarily by the local two-phase mixture density (solid concentration) at the location of that surface, for any given type of particle. The early observed parametric effects of elevation, gas velocity, solid mass flux, and radial position are all secondary to this primary functional dependence. 

The parametric effect of bed temperature is expected to be reflected through higher thermal conductivity of gas and higher thermal radiation fluxes at higher temperatures. Basu and Nag (1996) show the combined effect (Fig. 23) which plots heat transfer coefficients as a function of bed temperature for data from four different sources. It is seen that for particles of approximately the same diameter, at a constant suspension density (solid concentration), the heat transfer coefficient increases by almost 300% as the bed temperatures increase from 600°C to 900°C. 

Some researchers have noted that this approach tends to underestimate the lean phase convection since solid particles dispersed in the up-flowing gas would cause enhancement of the lean phase convective heat transfer coefficient. Lints (1992) suggest that this enhancement can be partially taken into account by increasing the gas thermal conductivity by a factor of 1.1. It should also be noted that in accordance with Eq. (3), the lean phase heat transfer coefficient (h,) should only be applied to that fraction of the wall surface, or fraction of time at a given spot on the wall, which is not submerged in the dense/particle phase. This approach, therefore, requires an additional determination of the parameter fh to be discussed below. 

One method of improving G/S contacting consists of showering solids in dilute suspension from the top into an upflowing gas stream. Experiments verified that gas/solid heat transfer coefficient for such a system is essentially the same as for the discrete particles, and that pressure drop for gas flow is extremely low. 

Normally, the heat-transfer rate is between 5 and 25 times that for the gas alone. Bed-to-surface-heat transfer coefficients vary according to the type of solids in the bed. Group A solids have bed-to-surface heat-transfer coefficients of approximately 300 J/(m2s-K) [150 Btu/(h-ft2-°F)]. Group B solids h ave bed-to-surface heat-transfer coefficients of approximately 100 J/(m2- s-K) [50 Btu/(h-ft2-°F)], while group D solids have bed-to-surface heat-transfer coefficients of 60 J/(m2-s-K) [30 Btu/(hft2 oF)]. 

The fluid bed has generally established itself as the preferred equipment for the crystallization and for the cooling section of the SSP plant. High heat-transfer coefficients enable the pellets to be heated and cooled very quickly, pellet agitation can be achieved without dust generation, and the direct contact between gas and solid enables a de-dusting effect. 

In a modified system in which a suspension of solids is conveyed through the heat transfer section, the heat transfer coefficient is greater than that obtained with liquid alone, though lower than that obtained at the same concentration in a fluidised system. Similar conclusions have been reached by Jepson, Poll, and Smith(95) who measured the heat transfer to a suspension of solids in gas. 

With gas-solids systems, the heat-transfer coefficient to a surface is very much dependent on the geometrical arrangement and the quality of fluidisation furthermore, in many cases the temperature measurements are suspect. Leva 100 plotted the heat transfer coefficient for a bed, composed of silica sand particles of diameter 0.15 mm fluidised in air, as a function of gas rate, using the correlations put forward as a result of ten different investigations, as shown in Figure 6.23. For details of the experimental conditions relating to these studies reference should be made to Leva. For a gas flow of 0.3kg/m2 s, the values of the coefficient ranged from about 75 W/m2 K when calculated by the formula of... 

---