International Journal of Heat and Mass Transfer xxx (2011) xxxxxx

Contents lists available at ScienceDirect

International Journal of Heat and Mass Transfer

Momentum and heat transfer phenomena of spheroid particles at moderate Reynolds and Prandtl numbers
Nanda Kishore , Sai Gu
Energy Technology, School of Engineering Sciences, University of Southampton, Southampton SO17 1BJ, United Kingdom

a r t i c l e

i n f o

a b s t r a c t
The momentum and heat transfer phenomena of spheroid particles in an unbounded Newtonian uid have been numerically investigated by solving governing conservation equations of the mass, the momentum and the energy. The numerical solution methodology has been benchmarked by performing comparisons between present results with those reported in the literature. Further, extensive new results have been obtained to elucidate effects of pertinent dimensionless parameters such as the Reynolds number (Re), the Prandtl number (Pr) and the aspect ratio (e) on the ow and heat transfer behaviour of spheroid particles in the range of parameters: 1 6 Re 6 200; 1 6 Pr 6 1000 and 0.25 6 e 6 2.5. Regardless of the value of the Reynolds number, the total and individual drag coefcients of oblate spheroids (e < 1) are smaller than those of spheres (e = 1) and opposite trend has been observed for prolate spheroids (e > 1). Irrespective of values of Reynolds and Peclet numbers, the average Nusselt number is large for prolate particles as compared to spheres and opposite trend has been observed for the case of oblate particles. Major contribution of this work is the development of simple correlations for the total drag coefcient and the average Nusselt number of unconned isolated spheroid particles based on present numerical results which can be used in new applications. 2011 Elsevier Ltd. All rights reserved.

Article history: Received 30 August 2010 Accepted 25 January 2011 Available online xxxx Keywords: Spheroid Aspect ratio Reynolds number Drag coefcient Prandtl number Nusselt number Heat transfer coefcient

1. Introduction The momentum and heat transfer behaviour of spherical and non-spherical particles are of fundamental importance in many chemical, environmental, food, mineral and process industries. Some of these applications include ow in xed and uidized bed reactors, fermentation, heterogeneous catalytic reactors, wastewater treatment, slurry transportation of particulates, mineral slurries, coal-oil slurries transport, etc. In such envisaged applications, the stratication of solid-liquid suspensions is undesirable to ensure the trouble free transportation of the suspensions. In some food applications, during the thermal treatment of foodstuffs, the density-based stratication is detrimental to maintain the nal food quality. On the other hand, such stratication is highly undesirable in order to maintain the homogeneity of the product such as fruit juices, soups, pharmaceutical products, liquid detergents, cleaning aids, etc. Therefore, regardless of the necessity of the stratication of suspensions, the knowledge of the settling velocity (or drag forces) and the heat transfer coefcient (or Nusselt number) are important key parameters in design of such solid-liquid contacting equipments [14]. On the other hand, in many process industries, one often encounters both regular and
Corresponding author. Tel.: +44 23 80598301.
E-mail addresses: [email protected], [email protected] (N. Kishore). 0017-9310/$ - see front matter 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.ijheatmasstransfer.2011.02.001

irregular particles; some of the regular shaped particles include spheres, cylinders, spheroids, ellipsoids, cubes, cones, etc. In general, these particles sediment as clusters; however, in the design of aforementioned contacting equipments, one often require the drag coefcient and the Nusselt number correlations which are based on single particle approach. Furthermore, the settling behaviour of unbounded isolated particles provides the insight about the physics of the problem and would lead to the simplication of complex multi-particle systems [2]. In this context, the momentum and heat transfer phenomena of spheres and cylinders have been most thoroughly studied as compared to other regular shaped particles [13]. On the other hand, many experimental observations indicate that the spheroid shape is a very good approximation for non-spherical particles [1]. Despite this fact, limited attention has been focussed to investigate the momentum of and heat/mass transfer from unconned spheroid particles beyond the creeping ow regime. Therefore, the aim of this study is to numerically investigate the momentum and heat transfer phenomena of isolated unconned spheroid particles in the range of parameters: 1 6 Re 6 200; 1 6 Pr 6 1000 and 0.25 6 e 6 2.5. 2. Previous work The pertinent literature on the motion of unconned and conned spheroid particles has been thoroughly reviewed in our

Please cite this article in press as: N. Kishore, S. Gu, Momentum and heat transfer phenomena of spheroid particles at moderate Reynolds and Prandtl numbers, Int. J. Heat Mass Transfer (2011), doi:10.1016/j.ijheatmasstransfer.2011.02.001

N. Kishore, S. Gu / International Journal of Heat and Mass Transfer xxx (2011) xxxxxx

recent work [4]; however, some of these studies are briey recounted here. The creeping motion past oblate and prolate spheroid particles has been investigated analytically [5,6] and by using semi-analytical and/or semi-numerical techniques [712]. Most of these studies report drag coefcients of spheroid particles as a function of shape correction factor to the existing well known Stokes drag of spherical particles in the zero-Reynolds number regime. Similarly, few numerical results are also available on the ow phenomena of thin oblate particles (aspect ratio 0.050.2) at intermediate Reynolds numbers [1322]. Wang et al. [23] have experimentally studied the sedimentation characteristics of spherical and hemispherical particles with faces upward and downward while settling in a viscous uid. However, none of these studies reported correlations for the drag coefcient of spheroid particles beyond the creeping ow regime. Militzer et al. [24] have compiled the existing literature data on the drag coefcient of spheres and spheroids; and developed a correlation in the intermediate range of the Reynolds number and the aspect ratio. However, this correlation overestimates the drag coefcient of spheres (e = 1) both at low and high Reynolds numbers. Recently, Kishore and Gu [4] have proposed a drag model for the case of conned spheroid particles in the moderate range of Reynolds numbers up to 200. In summary, no reliable correlation is available for the drag coefcient of unconned spheroid particles over wide range of conditions based on purely numerical results. On the other hand, pertaining to heat transfer characteristics of spheroid particles, large gap exist in the literature. Only a few analytical and numerical results are available either for small but nite Prandtl numbers < 10 or for very large Peclet numbers using the boundary layer approximation. Norminton and Blackwell [25] have obtained approximate series solutions in descending powers of time for the transient heat ow from isothermal spheroid particles. Masliyah and Epstein [26] have numerically determined heat and mass transfer rates from spheroids of aspect ratio 0.2 at intermediate Reynolds numbers and Schmidt number of 0.7. They have proposed several correlations for average Nusselt number which can not be readily used because of inherent complications. Pitter et al. [27] have numerically investigated the effect of the forced convection on the mass transfer from a thin oblate spheroid (aspect ratio 0.05) of ice in the air (Schmidt number 0.7) at Reynolds numbers up to 20. Yovanovich [28] has presented a semi-empirical correlation for the laminar forced convection heat and mass transfer from isothermal spheroid particles into an extensive uid at Prandtl number less than unity. Comer and Kleinstreur [29] have numerically studied the laminar axisymmetric thermal ow past oblate spheroids of aspect ratio 0.21 at intermediate Reynolds numbers and for limited Prandtl number of 0.7 and 7. Zheng and List [30] have carried out wind tunnel experiments to investigate the latitudinal and overall convective heat transfer of rotating spheres and spheroids of aspect ratio between 0.5 and 1 with non-uniform surface temperatures at very high Reynolds numbers of order of magnitude 104 and limited thermal diffusivity. Feng and Michaelides [31] have examined the unsteady heat transfer from spheroid particles in the limit of very small Peclet numbers using a perturbation technique and obtained rates of heat and mass transfer in Laplace and time domains. Wen and Jog [32] have numerically studied effects of the Reynolds number, the aspect ratio, surface and far eld temperatures and variable thermal properties on the ow eld, isotherms, drag coefcients and average Nusselt numbers at intermediate Reynolds numbers. Juncu [33] has numerically studied the unsteady heat transfer from spheroid particles with aspect ratio 0.10.9 at intermediate Reynolds numbers for limited values of Prandtl numbers 1 and 10. In summary, despite a few works on the heat transfer phenomena of spheroid particles are reported in the literature; no reliable correlations for heat transfer coefcient

of unbounded isolated spheroid particles based on purely numerical results are available over wide range of the aspect ratio, Reynolds and Peclet numbers. Therefore, on the basis of aforementioned discussions, it is safe to conclude that there is a strong necessity for reliable drag and heat transfer correlations of isolated unconned spheroid particles over wide ranges of the aspect ratio, Reynolds and Peclet numbers. Thus, this work is aimed to ll this gap in the literature. 3. Problem statement and description Consider a spheroid particle heated at a constant temperature TS, and located in a long cylindrical tube. An incompressible Newtonian uid is entering the tube at velocity V with pressure P0 and temperature T0 and leaving the tube with pressure P. The spheroid particle is located along the central axis of the tube at a distance of Lu from the inlet and at a distance of Ld from the outlet. The diameter of the cylindrical tube is Dt and the length of the tube L is the sum of Lu and Ld. The heat is transferring from the surface of the spheroid particle to the surrounding unbounded Newtonian uid which is owing at T0. In order to obtain the unbounded ow conditions, a large value of Dt has been chosen such that the extra retardation effects due to the presence of the imaginary tube wall are negligible. This is further ascertained by adopting the moving wall boundary condition for the imaginary tube wall. Similarly, large values of Lu and Ld have been chosen such that entry and exit effects are negligible. In other words, the computational domain is a cylinder, with boundary conditions and the size of the domain is chosen so that the ow around the spheroid particle is unaffected by the presence of boundaries. Furthermore, it is assumed that all the physical properties are constant; and the ow is two dimensional, steady and axisymmetric. These assumptions are consistent with other relevant studies [1422,2629,3233]. The size of the domain has been non-dimensionalized using the polar diameter of the spheroid particle as the scaling parameter. The governing equations are the conservation of the mass, the momentum and the energy, as written below:  Continuity equation:

oV x oV y ox oy

0:

 x-Component of momentum equation:

!   oV x oV x oP o2 V x o2 V x : q Vx Vy l ox ox oy ox2 oy2
 y-Component of momentum equation:

!   oV y oV y oP o2 V y o2 V y : q Vx Vy l oy ox oy ox2 oy2
 Energy equation:

!   oT oT o2 T o2 T k ; qC p V x V y ox oy ox2 oy2

where V is the velocity vector, q is the density of the uid, l is the viscosity of the uid and P is the pressure. The appropriate boundary conditions for this problem can be prescribed as follows:  At the inlet, the uniform axial velocity is prescribed:

V x V;

V y 0;

T T0:

 At the surface of the particle, the no slip condition is prescribed:

V x 0;

V y 0;

T T S:

Please cite this article in press as: N. Kishore, S. Gu, Momentum and heat transfer phenomena of spheroid particles at moderate Reynolds and Prandtl numbers, Int. J. Heat Mass Transfer (2011), doi:10.1016/j.ijheatmasstransfer.2011.02.001

N. Kishore, S. Gu / International Journal of Heat and Mass Transfer xxx (2011) xxxxxx

 At the outlet, the pressure outlet boundary condition has been used so that the conditions upstream of the exit shall not be inuenced:

op 0; ox

oV x 0; ox

oV y 0; ox

T T0:

 Along the imaginary tube wall, moving wall boundary condition has been used. The wall is imposed to move with a velocity equal to the inlet velocity:

V x V;

V y 0;

T T0:

 Along the axis of the tube, the symmetry boundary condition has been used:

V y 0;

oV x 0; oy

oT 0: oy

The governing differential Eqs. (1)(4) along with above mentioned boundary conditions (Eqs. (5)(9)) provide the complete theoretical frame to map the entire ow domain in terms of Vx, Vy, P and T. These equations have been numerically solved using Fluent as described in the next section. This commercial software uses conservation equations and boundary conditions in the Cartesian coordinates to solve any ow and heat/mass transfer (with or without chemical reaction) problem for any complex geometry. Finally, the fully converged steady state velocity and temperature elds are further used to evaluate individual and total drag coefcients; and heat transfer characteristics such as local and average heat transfer coefcients via the calculation of Nusselt number. In this work, results have been reported as functions of pertinent dimensionless parameters such as the Reynolds number (Re), the Prandtl number (Pr), the Peclet number (Pe), the total drag coefcient (Cd), the Nusselt number (Nu) and the aspect ratio (e); and are dened as follows:  Reynolds number:

software based on computational uid dynamics (CFD), Fluent; in conjunction with a mesh generating software, Gambit. The detailed numerical procedure has been presented elsewhere [4]; hence only salient features are recapitulated here. Semi-implicit method for pressure-linked equations (SIMPLE) solution algorithm along with quadratic upstream interpolation for convective kinematics (QUICK) scheme for convective terms has been chosen. The solution procedure initialized from the inlet and continued calculations until the steady state velocity and pressure elds have been obtained in the entire computational domain for different combinations of the Reynolds number and the aspect ratio. So obtained steady velocity eld is further used to solve the energy equation for different combinations of the Reynolds number, the Prandtl number and the aspect ratio in the similar manner. The tolerance residual value for the complete convergence has been set to 1015. The fully converged steady velocity, pressure and temperature elds are further used to evaluate the near object kinematics such as individual and total drag coefcients; and local and average Nusselt numbers as described in the previous section. 4.1. Domain independence The domain size will have signicant effect on the ow eld at low Reynolds numbers (due to domination of viscous forces) for particles of large aspect ratio. Thus, the domain independence study has been carried out at Re = 1 for a prolate spheroid particle of aspect ratio e = 2.5 and are presented in Table 1. This table summarizes individual and total drag coefcients for three different sizes of the domain, i.e., 60 40, 90 60 and 120 80. From this table, it is clear that by increasing the size of the domain beyond 90 60, the improvement in either individual or total drag coefcient is <0.4% but at the expense of large increase in CPU time. Thus a domain of size 90 60 with Lu = 30, Ld = 60, and Dt/2 = 60 is found to be optimum. 4.2. Grid independence The size of the grid shows signicant effect on the ow eld at high Reynolds numbers due to the domination of convective terms. Thus the grid independence study has been carried out at Re = 200 for a prolate spheroid particle of aspect ratio e = 2.5 and are presented in Table 2. This table shows summary of effect of grid on the total drag coefcient using four different grids and it is found that all four grids produce drag coefcients within 0.05% of each other. Thus, a mesh with 300 node points on the surface of the particle, 90 and 240 node points on Lu and Ld, respectively and 60 node points on both inlet and outlet edges have been chosen as optimum grid. 5. Results and discussion To evaluate effects of pertinent dimensionless parameters such as the Reynolds number (Re), the Prandtl number (Pr) and the particle aspect ratio (e) on momentum and heat transfer phenomena of spheroid particles, the following range of conditions have been considered: Re = 1, 10, 20, 50, 100, 200; Pr = 1, 10, 50, 100, 500, 1000; and e = 0.25, 0.5, 1, 1.5, 2, 2.5.

Re

2bV q

10

 Prandtl number:

Cpl Pr : k
 Peclet number:

11

Pe Re Pr:
 Total drag coefcient:

12

Cd

2F d

qV 2 Ap

C dp C df :

13

 Nusselt number:

Nu

h2b : k

14

 Aspect ratio:

e b=a;

15

where a and b are equatorial and polar radii of the spheroid particle, Fd is the drag force acting on the particle, Ap is the area of the particle; Cdp and Cdf are pressure and frictional components of the total drag coefcient, Cp is the specic heat of the uid, h is the convective heat transfer coefcient and k is the thermal conductivity. 4. Numerical methodology The numerical solution of this problem is obtained by solving momentum and energy equations separately using commercial

Table 1 Domain independence study of ow past a spheroid of e = 2.5 at Re = 1. Domain size 60 40 90 60 120 80 Lu 20 30 40 Ld 40 60 80 Dt 40 60 80 Cdp 35.444 35.067 34.927 Cdf 25.819 25.575 25.492 Cd 61.263 60.642 60.419

Please cite this article in press as: N. Kishore, S. Gu, Momentum and heat transfer phenomena of spheroid particles at moderate Reynolds and Prandtl numbers, Int. J. Heat Mass Transfer (2011), doi:10.1016/j.ijheatmasstransfer.2011.02.001

N. Kishore, S. Gu / International Journal of Heat and Mass Transfer xxx (2011) xxxxxx

Table 2 Grid independence study of ow past a spheroid of e = 2.5 for an optimum domain of 90 60 at Re = 200. Number of nodes on edges Surface 250 300 300 350 Lu 60 90 120 120 Ld 180 240 300 300 Tube 90 90 90 100 Inlet 60 60 60 70 Outlet 60 60 60 70 Total cells 193772 315852 408716 485824 2.097 2.097 2.097 2.096 Total drag coefcient (Cd)

5.1. Validation Extensive comparisons of drag coefcients of unconned spherical and spheroid particles; and of conned spherical particles over wide ranges of the Reynolds number have been shown in our previous work [4]. Hence, in this study, additional comparisons pertaining to the present average Nusselt numbers have only been shown in Table 3. This table shows the comparison between the present values of the average Nusselt number of spherical particles with those of Kishore et al. [34] and Feng and Michaelides [35] for different values of the Reynolds and Peclet numbers. The agreement between the present and literature values for any combination of Re and Pe is excellent. Furthermore, differences of this order of magnitude are not at all uncommon in such numerical studies and are frequently ascribed to the numerical procedure, grid adopted, etc. [36,37]. Thus, our previous experience [4] and the present benchmarking studies inspire the condence on the reliability and accuracy of the present solver and it is believed that the present results are accurate within 24%. 5.2. Flow and drag phenomena Fig. 1 represents combined effects of the Reynolds number and the aspect ratio on streamline patterns in the vicinity of different spheroid particles. The upper half of the central axis shows streamlines at Re = 20 (solid lines) and lower half depicts those at Re = 200 (dashed lines). It can be seen that for spheres (e = 1; Fig. 1(c)) and oblate spheroids (e < 1; Fig. 1(a)(b)), there is no ow separation at Re = 20 (upper halves solid lines); however, for prolate particles (e > 1; Fig. 1(d)(f)), recirculation wake exist in the rear of the particle and its size increases as the value of the aspect ratio increases. For oblate spheroid of aspect ratio e = 0.25 (Fig. 1(a)), there is no ow separation even at Re = 200 (lower half dashed lines) due to the small frontal area of the particle which is acting as obstacle in the ow. On the other hand, for the spheroid of aspect ratio e = 0.5 (Fig. 1(b)), there is small recirculation of the ow at the rear end of the particle. As the value of the aspect ratio increases further, the frontal area also increases which causes the ow to separate from the surface of particle and eventually the size of the

recirculation wake increases. Thus, in summary, the length of the recirculation wake in the rear end of the particle increases as the value of the Reynolds number and/or the value of the aspect ratio increases. Fig. 2 shows combined effects of the Reynolds number and the aspect ratio on individual and total drag coefcients; and on the relative contribution between pressure and frictional components of the total drag coefcient. Regardless of the value of the aspect ratio, as the value of the Reynolds number increases, the pressure drag (Fig. 2(a)), the friction drag ((Fig. 2(b)) and the total drag (Fig. 2(c)) coefcients gradually decreases, whereas the ratio between the pressure and friction drag coefcient increases (Fig. 2(d)). For all values of the aspect ratio, at small values of the Reynolds numbers, due to the domination of viscous forces and because of poor pressure recovery, individual and total drag coefcients are large. However, as the value of the Reynolds number increases the domination of viscous forces decreases and convective terms domination increases; hence, both individual and total drag coefcient decreases. For oblate spheroids, quantitatively the pressure recovery is better and the effect of the viscous force on the oblate particle is small. Therefore, as compared to spheres, both individual and total drag coefcients of oblate spheroids are small and opposite trends have been observed for prolate spheroids. For all values of the aspect ratio, at small values of the Reynolds number, because of the domination of the viscous forces, the viscous drag coefcient is large as compared to the pressure drag coefcient. Thus, from Fig. 2(d), at small values of the Reynolds number, the value of Cdp/Cdf is small and as the value of the Reynolds number increases, because of the increased contribution of convective forces over viscous forces, the ratio between pressure to friction drag coefcients increases. On the basis of present numerical results, the following simple correlation for the total drag coefcient of spheroid particles has been proposed which can be used in new applications:

Cd

i 24e0:49 h 1:05 0:152Re0:687 e0:671 : Re

16

Table 3 Comparison of present average Nusselt numbers of spheres with literature. Re Pe = 1 1 10 100 Pe = 100 1 10 100 Pe = 100 1 10 100 Present 2.323 2.369 2.401 5.809 6.462 7.685 11.314 12.735 15.758 Kishore et al. [34] 2.323 2.402 5.813 7.696 11.318 15.764 Feng and Micahelides [35] 2.315 2.362 2.384 5.865 6.552 7.695 11.628 13.201 15.946

The above equation reproduces present numerical results with an average error of 5.8% which rises to a maximum of 18.7%. However, if results for e = 0.25 are excluded, then the average error drops to less than 4%. Furthermore, by substituting e = 1, this correlation reduces to slightly modied form of the classical drag model due to Schiller and Naumann [38] for the case of solid spheres. This indicates that the present results compare very well with existing literature correlations. 5.3. Heat transfer phenomena Fig. 3(a)(c) shows combined effects of the Reynolds number and the Prandtl number on isotherm contours in the vicinity of prolate particles of aspect ratio, e = 2. The upper half (solid lines) above the central line depict isotherm contours at Re = 20 whereas the lower half (dashed lines) represent those at Re = 200 for different Prandtl numbers ranging from 1 to 1000. In the case of Pr = 1 (Fig. 3(a)), for Re = 20 (upper half solid lines), isotherm contours

Please cite this article in press as: N. Kishore, S. Gu, Momentum and heat transfer phenomena of spheroid particles at moderate Reynolds and Prandtl numbers, Int. J. Heat Mass Transfer (2011), doi:10.1016/j.ijheatmasstransfer.2011.02.001

N. Kishore, S. Gu / International Journal of Heat and Mass Transfer xxx (2011) xxxxxx

Fig. 1. Streamline contours of spheroid particles at Re = 20 (upper half solid lines) and Re = 200 (lower half dashed lines).

10

(a)
10

(b) e = 0.25
= 0.5 =1 = 1.5 =2 = 2.5

10

10

Cdp

10

Cdf
10
0 0 1 2

10

-1

10

-2

10

-1

10

10

10

10

10

10

Re
10
2

Re
(d)
4

(c)

3 10
1

C /Cdf dp
10
0

Cd

10

-1

10 0

10 1

10 2

50

100

150

200

Re

Re

Fig. 2. Effect of aspect ratio on: (a) pressure drag, (b) friction drag, (c) total drag, and on (d) ratio between pressure and friction drag coefcients at different Reynolds numbers.

are less dense indicating poor temperature gradients. As the value of the Prandtl number increases to 100 (see Fig. 3(b)), due to the increased contribution of the convection, isotherm contours are dragged along the ow direction. Further increasing the Prandtl number to 1000 (see Fig. 3(c)) causes isotherms to come much close to the surface of particle because of the presence of the recirculation wake in the rear end of the particle and hence promoting the rate of heat transfer. In the case of Re = 200 (lower half dashed lines), for Pr = 1 (Fig. 3(a)), as compared to the upper half (Re = 20), moderate number of isotherms are concentrated around

the particle indicating moderate heat transfer because of contributions from both conduction and convection. As the value of the Prandtl number increases further up to 1000 (lower half; for the case of Re = 200), the more number of isotherm contours concentrate close to the particle with large temperature gradient indicating large enhancement in the rate of heat transfer. Unlike in the case of Re = 20, for the case of Re = 200 the isotherm contour value of 333.333 K is very close to the surface with contour value of 600 K. The enhancement in the rate of heat transfer is attributed to both the recirculation zone (large Reynolds number) and

Please cite this article in press as: N. Kishore, S. Gu, Momentum and heat transfer phenomena of spheroid particles at moderate Reynolds and Prandtl numbers, Int. J. Heat Mass Transfer (2011), doi:10.1016/j.ijheatmasstransfer.2011.02.001

N. Kishore, S. Gu / International Journal of Heat and Mass Transfer xxx (2011) xxxxxx

Fig. 3. Isotherm contours of prolate spheroid particles of e = 2 (ac) and oblate spheroid particles of e = 0.5 (df) at Re = 20 (upper half solid lines) and Re = 200 (lower half dashed lines) for different Prandtl numbers.

dominated effect of convection heat transfer. Similarly, Fig. 3(d)(f) shows combined effects of the Reynolds number and the Prandtl number on isotherm contours in the vicinity of oblate spheroid particles of aspect ratio e = 0.5. Here too similar observations can be made as in the case of prolate particles of e = 2. Fig. 4 shows the effect of the aspect ratio on the average Nusselt number of different spheroid particles as functions of Reynolds and Peclet numbers. Regardless of the value of the Reynolds number

and the aspect ratio, as the value of the Peclet number increases, the average Nusselt number increases. For xed values of Reynolds and Peclet numbers, as the value of the aspect ratio increases, the rate of heat transfer increases. Similarly, for a xed value of the Peclet number and irrespective of type of spheroid particle, as the Reynolds number increases, the heat transfer coefcient increases because of the recirculation wake that occurs at the rear end of the particle.

10

(a) Re = 1
e = 0.25
= 0.5 =1 = 1.5 =2 = 2.5

10

(b) Re = 10

10

(c) Re = 20

Nu avg

Nu avg

10

10

Nu avg
2 3 4

10

10 0 0 10

10

10

10

10 0 1 10

10

10

10

100 1 10

10

10

10

10

Pe

Pe

Pe
10
3

(d) Re = 50
10
2 2

(e) Re = 100

(f) Re = 200

10

10

Nu avg

Nu avg

10

10

Nu avg
10
1 2 3 4 5

10

10

10

10

10

10

10

10

10

10

10

10 2 10

10

10

10

10

Pe

Pe

Pe

Fig. 4. Effect of aspect ratio on average Nusselt number for different Reynolds and Peclet numbers.

Please cite this article in press as: N. Kishore, S. Gu, Momentum and heat transfer phenomena of spheroid particles at moderate Reynolds and Prandtl numbers, Int. J. Heat Mass Transfer (2011), doi:10.1016/j.ijheatmasstransfer.2011.02.001

N. Kishore, S. Gu / International Journal of Heat and Mass Transfer xxx (2011) xxxxxx

Finally, on the basis of present numerical results, the following simple correlation for the average Nusselt number of spheroid particles has been proposed which can be used in new applications:

n o Nuavg 2e0:3 Pr0:4 0:4Re0:5 e0:83 0:06Re2=3 e0:1 :

17

The above correlation reproduces present numerical results (216 data points) with an average error of 7.98% which rises to a maximum of 28.1%. Further, by substituting e = 1, the above correlation reduces to the well known correlation of Whitaker [39] for the average Nusselt number of spheres. Here too, present results are comparing very well with existing literature correlations for spherical particles. But, unfortunately no correlations for average Nusselt number of unconned spheroid particles are available to compare the present model in the wide range of Reynolds and Peclet numbers. 6. Conclusion The momentum and heat transfer characteristics of spheroid particles in the steady axisymmetric ow regime have been numerically investigated using Fluent. For e 6 1, there is no ow separation for Reynolds number 620; however, as values of the aspect ratio and/or the Reynolds number, length of recirculation wake in the rear of spheroid particles increases. The individual and total drag coefcients of prolate spheroid particles are larger than those of solid spheres and opposite trends have been observed for oblate spheroid particles. For xed values of the Reynolds number and the aspect ratio, more isothermal contours are concentrated near the prolate particles as compared to spherical particles and opposite trends have been observed for oblate spheroid particles. Regardless of values of Reynolds and Peclet numbers, the average Nusselt number is large for prolate particles as compared to spheres and opposite trend has been observed for the case of oblate spheroid particles. Finally, on the basis of present numerical results, simple correlations for total drag coefcients and average Nusselt numbers of spheroid particles have been proposed which can be used in new applications. Acknowledgements Authors gratefully acknowledge the nancial support from Royal Commission for the Exhibition of 1851 Research Fellowship Award, EU FP7 SIMUSPRAY Project (Grant No. 230715), UK EPSRC Grant EP/G034281/1 and Leverhulme-Royal Society Africa Award. References
[1] R. Clift, J.R. Grace, M.E. Weber, Bubbles, Drops and Particles, Academic Press, New York, 1978. [2] J. Happel, H. Brenner, Low Reynolds Number Hydrodynamics, Nijhoff, Dordrecht, Netherlands, 1983. [3] E.E. Michaelides, Particles, Bubbles and Drops: Their Motion, Heat and Mass Transfer, World Scientic, Singapore, 2006. [4] N. Kishore, S. Gu, Wall effects on ow and drag phenomena of spheroid particles at moderate Reynolds numbers, Ind. Eng. Chem. Res. 49 (2010) 9486 9495. [5] A. Oberbeck, Uber die stationare ussigkeitsbewegungen mit berbucksichtigung der inneren Reibung, J. Reine Angew. Math. 81 (1876) 62 80. [6] L.E. Payne, W.H. Pell, The stokes ow problem for a class of axially symmetric bodies, J. Fluid Mech. 7 (1960) 529549.

[7] D.R. Breach, Slow ow past ellipsoids of revolution, J. Fluid Mech. 10 (1961) 306314. [8] I. Proudman, J.R.A. Pearson, Expansions at small Reynolds numbers for the ow past a sphere and a circular cylinder, J. Fluid Mech. 2 (1957) 237262. [9] T. Aoi, The steady ow of viscous uid past a xed spheroidal obstacle at small Reynolds numbers, J. Phys. Soc. Jpn. 10 (1955) 119129. [10] M. Sugihara-Seki, The motion of an ellipsoid in tube ow at low Reynolds numbers, J. Fluid Mech. 324 (1996) 287308. [11] R.S. Alassar, H.M. Badr, Steady ow past an oblate spheroid at small Reynolds numbers, J. Eng. Math. 36 (1999) 277287. [12] S. Blaser, Forces on the surface of small ellipsoidal particles immersed in a linear ow eld, Chem. Eng. Sci. 57 (2002) 515526. [13] Y. Rimon, H.J. Lugt, Laminar ow past oblate spheroids of various thickness, Phys. Fluids 12 (1969) 24652472. [14] J.H. Masliyah, N. Epstein, Numerical study of steady ow past spheroids, J. Fluid Mech. 44 (1970) 493512. [15] J.K. Comer, C. Kleinstreuer, A numerical investigation of laminar ow past nonspherical solids and droplets, J. Fluid Eng. Trans. ASME 117 (1995) 170175. [16] R.L. Pitter, H.R. Pruppacher, A.E. Hamielec, A numerical study of viscous ow past a thin oblate spheroid at low and intermediate Reynolds numbers, J. Atmos. Sci. 30 (1973) 125134. [17] C.C. Chang, B.H. Liou, R.L. Chern, An analytical and numerical study of axisymmetric ow around spheroids, J. Fluid Mech. 234 (1992) 219246. [18] W.Z. Li, Y.Y. Yan, J.M. Smith, A numerical study of the interfacial transport characteristics outside spheroidal bubbles and solids, Int. J. Multiphase Flow 29 (2003) 435460. [19] A. Tripathi, R.P. Chhabra, T. Sundararajan, Power law uid ow over spheroid particles, Ind. Eng. Chem. Res. 33 (1994) 403410. [20] A. Tripathi, R.P. Chhabra, Drag on spheroidal particles in dilatant uids, AIChE J. 41 (1995) 728731. [21] Q. Xu, E.E. Michaelides, A numerical study of the ow over ellipsoidal objects inside a cylindrical tube, Int. J. Numer. Methods Fluid 22 (1996) 10751087. [22] A.A. Zamyshlyaev, G.R. Shrager, Fluid ows past spheroids at moderate Reynolds numbers, Fluid Dyn. 39 (2004) 376383. [23] J. Wang, H. Qi, C. You, Experimental study of sedimentation characteristics of spheroidal particles, Particuology 7 (2009) 264268. [24] J. Militzer, J.M. Kan, F. Hamdullahpur, P.R. Amyotte, A.M. Al Taweel, Drag coefcient for axisymmetric ow around individual spheroid particles, Powder Technol. 57 (1989) 193195. [25] E.J. Norminton, J.H. Blackwell, Transient heat ow from constant temperature spheroids and the thin circular disk, Q. J. Mech. Appl. Math. XVII (1964) 6572. [26] J.H. Masliyah, N. Epstein, Numerical solution of heat and mass transfer from spheroids in steady axi-symmetric ow, Prog. Heat Mass Transfer 6 (1972) 613632. [27] R.L. Pitter, H.R. Pruppacher, A.E. Hamielec, A numerical study of the effect of forced convection on mass transfer from a thin oblate spheroid of ice in air, J. Atmos. Sci. 31 (1974) 10581066. [28] M.M. Yovanovich, General expression for forced convection heat and mass transfer from isopotential spheroids, in: AIAA 26th Aerospace Sciences Meeting, Nevada, AIAA88-0743, January 1114, 1988. [29] J.K. Comer, C. Kleinstreuer, Computational analysis of convection heat transfer to non-spherical particles, Int. J. Heat Mass Transfer 38 (1995) 31713180. [30] G. Zheng, R. List, Convective heat transfer of rotating spheres and spheroids with non-uniform surface temperatures, Int. J. Heat Mass Transfer 39 (1996) 18151826. [31] Z.G. Feng, E.E. Michaelides, Unsteady heat and mass transfer form a spheroid, AIChE J. 43 (1997) 609614. [32] Y. Wen, M.A. Jog, Variable property, steady, axi-symmetric, laminar, continuum plasma ow over spheroid particles, Int. J. Heat Fluid Flow 26 (2005) 780791. [33] G. Juncu, Unsteady heat transfer from an oblate/prolate spheroid, Int. J. Heat Mass Transfer 53 (2010) 34833494. [34] N. Kishore, R.P. Chhabra, V. Eswaran, Mass transfer from ensembles of Newtonian uid spheres at moderate Reynolds and Peclet numbers, Chem. Eng. Res. Des. 85 (2007) 12031214. [35] Z.G. Feng, E.E. Michaelides, Heat and mass transfer coefcients of viscous spheres, Int. J. Heat Mass Transfer 44 (2001) 44454454. [36] P.J. Roache, Perspective: a method for uniform reporting of grid renement studies, J. Fluid Eng. Trans. ASME 116 (1994) 405413. [37] N. Kishore, R.P. Chhabra, V. Esaran, Sedimentation in emulsions of mono-size droplets at moderate Reynolds numbers, Chem. Eng. Res. Des. 84 (2006) 1180 1193. [38] L. Schiller, A.Z. Naumann, Uber die grundlegenden Berechungen bei der Schwerkraaufbereitung, Ver. Deut. Ing. 77 (1933) 318320. [39] S. Whitaker, Forced convection heat transfer correlations for ow in pipes, past at plates, single cylinders, single spheres and ow in packed beds and tube bundles, AIChE J. 18 (1972) 361371.

Please cite this article in press as: N. Kishore, S. Gu, Momentum and heat transfer phenomena of spheroid particles at moderate Reynolds and Prandtl numbers, Int. J. Heat Mass Transfer (2011), doi:10.1016/j.ijheatmasstransfer.2011.02.001