Unsteady heat transfer from an equilateral triangular cylinder in the unconfined flow regime

AMIT DHIMAN* and RADHE SHYAM

DEPARTMENT OF CHEMICAL ENGINEERING INDIAN INSTITUTE OF TECHNOLOGY ROORKEE, 247 667, INDIA

*ADDRESS CORRESPONDENCE TO Dr. AMIT KUMAR DHIMAN Assistant Professor Department of Chemical Engineering Indian Institute of Technology Roorkee, 247 667, India E-MAIL: dhimuamit@rediffmail.com TEL.: +91-1332-285890 (OFFICE), +91-9410329605 (MOBILE)

1

Abstract In the present study, heat transfer from a two-dimensional triangular cylinder is considered in the unsteady flow regime. In particular, the effects of Reynolds numbers on the heat transfer characteristics of a long equilateral triangular cylinder (heated) are investigated for the range of conditions Re = 50 to 150 (in the steps of 10) and Prandtl number = 0.71 (air) in the unconfined unsteady cross-flow. In order to simulate this situation, the computational grid is created by using commercial grid generator GAMBIT and the numerical computations are carried out by using FLUENT (6.3). The SIMPLE method is used to solve Navier-Stokes and energy equations along with the appropriate boundary conditions. The second order upwind scheme is used to discretize the convective terms, while the central difference scheme is used to discretize the diffusive terms in the governing equations. The present results are in an excellent agreement with the literature values. The temperature isotherms and temporal history of Nusselt number are presented. The local as well as time-averaged Nusselt numbers are calculated. The time-averaged Nusselt number increases with increasing Reynolds number for the fixed value of the Prandtl number. Finally, the present numerical results are used to develop the simple heat transfer correlation for the range of conditions covered here. Keywords: Triangular cylinder; Reynolds number; Prandtl number; Nusselt number; unsteady regime; Forced convection. Introduction The fluid flow over bluff bodies (e.g., cylinders) has been the topic of intense research for a century or so, because of its importance in various engineering applications such as electronic cooling, heat exchange systems, cooling towers, design of vortex flow meters and flow dividers, probes and sensors, etc. On the other hand, in spite of such wide applications, limited information is available in the open literature as far as forced convection flow and/or heat transfer from non-circular obstacles such as triangular, trapezoidal and rectangular cylinders are concerned [1 - 17]. This has motivated us to examine the heat transfer around an obstacle of triangular cross-section in the unsteady flow regime. For instance, as far as known to us, the fluid flow and/or heat transfer across a triangular cylinder are investigated only by a few researchers in both unbounded [6 – 10] and bounded [11 - 13] configurations. The studies dealing with the heat transfer from a long unconfined square cylinder can be found in [18 - 22]. In contrast, a healthy literature on the heat transfer from the unconfined obstacle of
2

Chattopadhyay [11] has done 2-D numerical study on the augmentation of heat transfer in a channel due to triangular prism at very high Reynolds numbers in the order of 10. [12] carried out the 2-D forced convection in a channel with a built-in triangular prism for the Reynolds number = 20 – 250 and Prandtl number of 0. [8] carried out 2-D numerical simulations for the fluid flow over an equilateral triangular bluff body for a blockage ratio of 1/15. Dalal et al. Therefore. De and Dalal [13] investigated the flow and heat transfer across an equilateral triangular cylinder placed in a 2-D confined space with blockage ratios of 1/12 ≤ β ≤1/3 for the range 80 ≤ Re ≤ 200 and Pr = 0. They reported the critical Reynolds number of 36.9.000. They reported the variation of drag coefficient. However. Abbassi et al. Recently. De and Dalal [9] carried out 2-D numerical investigations of the fluid flow across an equilateral triangular cylinder in the unconfined space in the Reynolds number range 10 ≤ Re ≤ 250. Strouhal number.71 . it can be summarized here that there is only one limited study due to Dalal et al. Wesfreid et al. More recently. triangular prisms and flat plates by using finite element method in the two-dimensional unbounded flow regime.71 . They studied global modes in wake flow and the value of critical Reynolds number for triangular cylinder is reported to be 39.circular cross-section can be found elsewhere [23 . as far as known to us. Jackson [6] examined the onset of periodic behavior over variously shaped bodies such as circular and elliptical cylinders. the 3 .35]. [10] proposed a novel finite-volume formulation for unsteady solutions on complex geometries. They presented a computational study of 2-D laminar flow and heat transfer past a triangular cylinder in free stream for the range 10 ≤ Re ≤ 200 and Pr = 0.71 for a fixed blockage ratio of 1/4. lift coefficient and Nusselt number with respect to Reynolds number. Zielinska and Wesfreid [7] performed 2-D numerical simulation of the flow through a triangular cylinder to obtain global modes of wake flow for the range of conditions: Reynolds number = 34 – 50 and blockage ratios = 1/5 and 1/15 in the bounded configuration. This study is concerned with the heat transfer from a long equilateral triangular obstacle in the unsteady unconfined flow regime. The time-averaged Nusselt number increases quasi-linearly with increasing value of the Reynolds number. They found an enhancement of about 85 % in the heat transfer at a Reynolds number of 250. As far as heat transfer from a long triangular obstacle is concerned. [10] is available in the open literature on the heat transfer from an unconfined equilateral triangular cylinder in the 2-D unsteady flow regime. very limit information is available on the variation of the local Nusselt number and no information in available on the temporal variation of the Nusselt number with Reynolds number.2 for the onset of the oscillations. Thus.

Continuity Equation ∂Vx ∂Vy + =0 ∂y ∂x (1) 4 . however. as shown in Fig. Tw ( > T∞ ). The heat transfer characteristics around the triangular obstacle are presented by instantaneous isotherm profiles and by temporal history of the Nusselt number at various values of the Reynolds number for the fixed value of the Prandtl number. Here. The thermo-physical properties of the streaming fluid are assumed to be independent of the temperature and the viscous dissipation effects are also neglected. The dimensionless equations of the continuity.71 ).(4). x. In particular. Finally. Problem statement and mathematical formulation In the present study. the heat transfer from a long equilateral triangular cylinder is investigated for the Reynolds number range Re = 50 − 150 (in the steps of 10) for air ( Pr = 0. the Reynolds and Prandtl numbers are defined as Re = dU ∞ ρ / µ and Pr = µ c p / k . and the non-dimensional downstream distance between the rear surface of the triangular cylinder and the exit plane ( X d / d ) is taken as 20.aim of this work is to provide the detailed systematic study on the heat transfer from a cylinder of triangular cross-section in the unconfined (periodic) flow regime. Here. the unconfined flow condition is simulated by creating an artificial long channel. with the total non-dimensional length of the computational domain ( L1 / d ) of 32 in the axial direction. The variation of the local Nusselt number on the surfaces of the triangular cylinder is shown. the non-dimensional upstream distance. respectively. the two-dimensional and unsteady (periodic) flow of an incompressible fluid (with a uniform velocity. the simple heat transfer correlation is obtained in order to calculate the time-averaged Nusselt number for the range of conditions covered here. as the computational domain has to be finite.momentums and energy balances are given by equations (1) . U ∞ and uniform temperature. 1. T∞ at the inlet) across a long equilateral triangular cylinder (of side d ) is considered. The * surface of the triangular obstacle is maintained at a constant temperature. X u / d is taken as 12. The height of the computational domain ( L2 / d ) is used as 30 in the lateral direction.and y.

Vx = 0 . • On the top and bottom boundaries: The symmetric boundary condition used by FLUENT is used.momentums and energy equations (1) – (4) along with above noted boundary conditions are solved for the fluid flow and heat transfer over a long equilateral triangular obstacle to obtain velocity. pressure and temperature fields. Numerical details 5 . This is similar to Neumann boundary condition as ∂Vx / ∂x = 0. Shrouhal number and Nusselt number. The extensive details of the flow parameters (e. which assumes a zero diffusion flux for all flow variables.e.x-Momentum equation ∂Vx ∂ ( Vx Vx ) ∂ ( Vy Vx ) ∂p 1  ∂ 2 Vx ∂ 2 Vx  + + =− + +   ∂t ∂x ∂y ∂x Re  ∂x 2 ∂y 2  y-Momentum equation ∂Vy ∂t + ∂ ( Vx Vy ) ∂x + ∂ ( Vy Vy ) ∂y 2 2 ∂p 1  ∂ Vy ∂ Vy =− + +  2 ∂y Re  ∂y 2  ∂x (2)     (3) Energy equation ∂θ ∂ ( Vxθ ) ∂ ( Vyθ ) 1  ∂ 2θ ∂ 2θ  + + = +   ∂t ∂x ∂y RePr  ∂x 2 ∂y 2  The following boundary conditions in their dimensionless forms are used (Fig. which are utilized to calculate the engineering parameters such as drag and lift coefficients.and y. i. Vy = 0 and θ = 1 .g. • At the exit boundary: The default outflow boundary condition in FLUENT. rms values of drag and lift coefficients and Strouhal number. etc. The above continuity. This study is concerned with the heat transfer from an equilateral triangular cylinder in the unconfined unsteady flow regime. etc. 1): • (4) At the inlet boundary: The isothermal fluid is flowing uniformly from left to right. i. Vy = 0 and θ = 0 . timeaveraged drag and lift coefficients.) are presented in our recent study [36]... • On the surface of the triangular obstacle: No-slip condition is used.e. ∂Vy / ∂x = 0 and ∂θ / ∂x = 0 . x. Vx = 1 . is used..

x. The values at the finite volumes can be summed up to get the value over entire domain. 0. The grid is generated by using GAMBIT. it is assumed that volume is made up of large number of small control volumes. the solution is assumed to have converged. as shown in Fig. FLUENT uses finite volume method. 2. Also. X d = 20 d and L2 = 30 d (Table 1). Table 1 also presents the grid independence results for the value of the Reynolds number of 100 and 6 .3). etc. several non-uniform orthogonal grids (with 100 and 200 grid points on each side of the triangular obstacle and having a first grid point at distances of 0. The number of grid points and their distribution is an important matter in such unsteady laminar flow over cylinders. The second order upwind scheme is used to discretize convective terms. equations can be discretized to be written in algebraic form to be solved. Choice of numerical parameters A thorough study has been carried out in order to fix the computational grid size. upstream distance. δ = 0.0015 d and 0. downstream distance and the height of the computational domain for the flow parameters as reported in our recent study [36].In the present work. according to which. the grid structure consists of non-uniform orthogonal cells having 116500 cells with 200 grid points on the each side of the triangular cylinder and a very fine grid of cell size. The governing equations are valid over all such control volumes. which are regular parallelepiped.and y-components of Navier-Stokes and energy equations to determine the flow and heat transfer parameters here. A convergence criterion of 10-20 is used for continuity. The fluid flow and heat transfer over a long unconfined equilateral triangular obstacle is solved by commercial CFD solver FLUENT (6. In this study. the corresponding heat transfer calculations are carried out in order to fix the grid size. The resulting algebraic equations are solved by Gauss-Siedel iterative scheme. upstream and downstream distances and the height of the computational domain for the heat transfer parameters in the unconfined unsteady regime. while the diffusive terms are discretized by central difference scheme.004 d . The computational grid structure used in this work is presented in Fig.001 d ) are examined for X u = 12 d . because of the complex phenomena existing in this type of flow due to separation and vortex shedding. when it shows more than 10 constant periodic cycles in the time history profiles.0015 d is used near the triangular cylinder and larger size grids are used away from the triangular cylinder. 2. For the grid resolution study. Hence.

0015 d with respect to the finest grid size of 0.. Results and discussion In this study. The percentage differences in the values of the time-averaged Nusselt number are found to be less than 0.0015 d and 200 grid points on the surface of the cylinder for X d = 20 d for Re = 100 and Pr = 0. In order to determine the effect of the upstream distance on the heat transfer parameters.4 % for the grid with 200 grid points on the each side of the triangular cylinder and having a first grid point at a distance of 0. the upstream distance of 12 d is used here.0015 d and 200 grid points on each surface of the triangular cylinder for X u = 12 d and X d = 20 d . The percentage relative differences in the values of the timeaveraged Nusselt number are found to be less than 0. Thus.. i.150 (in the steps of 10) and Prandtl number = 0. In summary. Thus.001 d ) for the Reynolds number of 100 and Prandtl number of 0.e. the percentage relative differences in the values of the timeaveraged Nusselt number are found to be only less than 0. X u = 9 d and 12 d with δ = 0. The domain independence study is carried out for two values of the height of the computational domain. in this work. the two-dimensional numerical simulations are carried out for the range of conditions: Reynolds number = 50 . However. the computational domain height of 30 d is used with δ = 0.Prandtl number of 0.71. L2 = 20 d and 30 d for δ = 0. Validation of results 7 .3 % for the value of Reynolds number of 100 and Prandtl number of 0.0015 d is used in all the computations reported in this work.71.0015 d . i. The percentage differences in the values of the time-averaged Nusselt number are found to be less than 0.25 %. L2 =30 d for the grid size of 116500 with 200 grid points on the each surface of the triangular cylinder and δ = 0. the grid with 200 grid points on the surface of the triangular cylinder and δ = 0. two values of the upstream distance are used.0015 d with 200 grid points prescribed on each surface of the triangular cylinder.71. the following parameters have been used for the generation of results presented here as X d =20 d .001 d .e.71.71 (air). X u =12 d .004 d with respect to the finest grid size ( δ = 0.2 % for the grid with 200 grid points on the each side of the triangular cylinder and having a first grid point at a distance of 0. Therefore.

100 and 150. (a) will be repeated after Fig. which span over the whole period. The following subsections present the details about the temperature fields around the triangular obstacle. 100 and 150.55 % and less than 3. For instance. which result in high temperature gradients as compared to the rear surface of the triangular cylinder.The comparison of present flow results and that of De and Dalal [9] of time-averaged drag coefficient and Strouhal number for the Reynolds number of 50. An excellent agreement between the present results and that of De and Dalal [9] can be seen in this table. which is very similar to Karman vortex street for all the values of the Reynolds number covered in this work. An excellent agreement can be seen here between the results of two studies. the present heat transfer results are compared with Dalal et al. respectively.0 % for the values of the Reynolds number of 50.6 % and less than 1.7 % and less than 1. etc.65 % for the values of the Reynolds number of 80. variation of the Nusselt number with time (i. Note that in Figs. 100 and 150. This validates the present numerical solution procedure. Figures 3 .5 (e . respectively.. less than 1. 3 – 5. However. temporal history of the Nusselt number) and the variation of local and time-averaged Nusselt number with Reynolds number.5 %. (d) for the next cycle. respectively.1 %. Also. However.d) present the representative instantaneous temperature profiles around the triangular cylinder for the Reynolds numbers of 50. Isotherm Patterns The effects of Reynolds numbers on the heat transfer around the long equilateral triangular obstacle are presented via instantaneous isotherm profiles for the fixed value of the Prandtl number of 0. less than 1.3 % for the values of the Reynolds number of 50. For the validity of the heat transfer results. the percentage deviations in the values of the timeaveraged drag coefficient are found to be less than 3. Fig. the magnified views of isotherms in the vicinity of the triangular cylinder can be seen in Figs.e. [10] in Table 3. the distribution of isotherms on the rear surface of the triangular 8 .h). The percentage deviations in the values of the time-averaged Nusselt number are found to be less than 1. It can also be seen in these figures that the top and bottom surfaces of the triangular obstacle have the maximum clustering of isotherms. 100 and 150 is shown in Table 2. 100 and 150 for the four successive moments of time.5 (a . less than 2. 3 . as the Reynolds number increases.71. the corresponding deviations in the values of the Strouhal number are found to be less than 1. It is clear from these figures that a temperature street is formed behind the triangular obstacle.7 %.

the corresponding magnified views of the time history of the Nusselt number at different values of the Reynolds 9 . which span over the whole period. for Reynolds numbers of 50. a local maximum in the value of the Nusselt number on the rear surface of the triangular cylinder can also be seen.5). in Fig. 6 (a .1). 3 . Local Nusselt number In this study. The local Nusselt number at each corner of the triangular cylinder increases with increasing value of the Reynolds number. 6 and is increasing with increasing value of the Reynolds number. Similar to the top surface. It is found that the grid with 200 cells on the one side of the triangular obstacle is sufficiently fine to obtain results that are essentially grid independent (subsection 3. However. this maxima in the value of the Nusselt number on the rear surface of the triangular obstacle fluctuates from the top-rear corner to bottom-rear corner and vice versa for the four successive moments of time. On the other hand. It can be seen from these figures that there exists a local minimum in the value of the Nusselt number on the top surface (near the top-rear corner) of the triangular cylinder (Fig. 120 and 150 are shown in Figs. where n is the cylinder surface normal direction. Also. 6 (d . 7 (a . the local minimum in the Nusselt number on the bottom surface exists near the bottom-rear corner of the triangular cylinder. the local Nusselt number ( Nu ) is defined as −∂θ / ∂n . The variation of the instantaneous local Nusselt number for the four successive moments of time. 6) for all the four successive moments of time considered here at different values of the Reynolds number. which span over the whole period. Time history of Nusselt number The temporal variations of the Nusselt number for the values of the Reynolds numbers of 50.f). 100 and 150 is shown in Figs. special attention is paid to the role of the grid size on the value of the local Nusselt number at each corner of the long triangular cylinder.cylinder affects more than that of the top and bottom surfaces of the triangular cylinder with time due to vortex shedding for the range of conditions covered here (Figs. 80. It is also clear from these figures that the distribution of local Nusselt number on the rear surface affects more than that of the top and bottom surfaces of the triangular obstacle with time due to vortex shedding in the unconfined unsteady regime.c). This is due the symmetrically shedding of vortices from the top-rear and bottom-rear corners of the triangular obstacle. The corresponding enlarged views are presented in Figs.e). especially at higher values of the Reynolds number. Due to the singularity at the corner of the triangular obstacle. 100.

5924 computed results for the range of conditions as 50 ≤ Re ≤ 150 and Pr = 0.number are presented in Figs. 5) has the maximum deviation of less than 0. Initially.5 % with the present In the present study. It is also clear from these figures that Nusselt number increases with increasing value of the Reynolds number.factor varies approximately linearly with Reynolds number for the fixed value of the Prandtl number for the range of conditions covered here.j) with time for all the cases considered here. as presented in Fig.71.factor with Reynolds number. The rms values of the Nusselt number are also calculated here. The time-averaged Nusselt number is calculated by averaging 10 cycles beyond the time the asymptotic shedding frequency of the Karman vortex is attained.3638 × Re0. the unsteady heat transfer across a long equilateral triangular cylinder has been simulated for the range of conditions: Reynolds number = 50 – 150 (in the steps of 10) and Prandtl number = 0. the average Nusselt number is calculated by averaging the local Nusselt number over the surfaces of the long equilateral triangular obstacle.e) and finally stabilizes (Figs. Further.71 .j). the instantaneous values of the Nusselt number are calculated at each time step and plotted versus time. The detailed heat transfer characteristics around the triangular obstacle are presented by instantaneous temperature profiles and by the temporal variation of the Nusselt number with Reynolds number. The time-averaged Nusselt number increases monotonically with increasing value of the Reynolds number for the Prandtl number of 0. The rms value of the Nusselt number increases with increasing value of the Reynolds number. the present results are also presented in terms of the usual Colburn heat transfer factor ( jh ) to show the functional dependency of the flow and heat transfer parameters [20 – 22. The time-averaged values of the Nusselt number are presented in Fig. 7f . Here. 8 (a). Figure 9 shows the variation of the jh . Conclusions − (5) The above correlation (Eq. The jh . 8 (b). 7a . these time-averaged values of the Nusselt number are utilized to obtain the simple heat transfer correlation for the range of physical parameters covered here. Nu = 0. The local Nusselt number at each corner of the triangular cylinder 10 .71 (air) in the unbounded unsteady flow regime. Average Nusselt number In this study. 7 (f . In order to give the further insights. 35]. the value of the Nusselt number fluctuates (Figs.

m cylinder surface normal direction local Nusselt number (= hd / k ) time-averaged Nusselt number (= h d / k ) − L1 L2 n Nu Nu _ Nurms p root mean square (rms) value of Nusselt number 2 )] pressure [= p* /( ρU ∞ Pr Re St t T * Tw Prandtl number [= µ c p / k ] Reynolds number [= dU ∞ ρ / µ ] Strouhal number time [= t * /(d / U ∞ ) ] time period constant wall temperature at the surface of the cylinder. W/m2 K Colburn factor for heat transfer [= Nu/(Re× Pr1/ 3 ) ] − CD cp h h _ − jh k thermal conductivity of the fluid. W/m2 K average convective heat transfer coefficient. m time-averaged drag coefficient specific heat of the fluid. J/kg K local convective heat transfer coefficient. simple heat transfer correlation is obtained for the range of conditions studied here.increases with increasing value of the Reynolds number. K uniform velocity of the fluid at the inlet. K temperature of the fluid at the inlet. Nomenclature d side of an equilateral triangular cylinder. m height of the computational domain. m/s T∞ U∞ 11 . Finally. W/m K length of the computational domain. The time-averaged Nusselt number increases monotonically with increasing value of the Reynolds number for the fixed value of the Prandtl number.

Cheremisinoff P. 1998.. 31. S. E. Handbook of Applied Polymer Processing Technology. vol.. Y. [2] Goujon-Durand. 99-112. P. pp. Downstream evolution of the Benard-von Karman instability.. and Kang.Vx component of velocity in the x-direction [= Vx* / U ∞ ] component of velocity in the y-direction [= Vy* / U ∞ ] stream-wise coordinate (= x* / d ) downstream distance of the cylinder. 1994. Numerical simulation of wake flow behind trapezoidal bluff bodies. S.. pp. and Wesfreid. 12 .. Numerical Methods Fluids. 1251-1254. [4] Chung. 12. Laminar vortex shedding from a trapezoidal cylinder with different height ratios.. P. N. N. P. Chapter 1. P.. J. in: Cheremisinoff N.. m upstream distance of the cylinder. and Wang. 308-313. R. Hydrodynamics of non-spherical particles in non-Newtonian fluids. J. T. pp. vol. Pa s density of the fluid. [5] Kahawita. (Eds. 1996.. Marcel Dekker. m Subscripts bottom surface of the triangular cylinder rear surface of the triangular cylinder top surface of the triangular cylinder surface of the triangular cylinder inlet condition Superscript r t w ∞ * dimensional variable References [1] Chhabra.). H. 2000.. Phys. Physical Review E. Early stages of an impulsively started steady laminar flow past tapered trapezoidal cylinders. R. 2002. vol. 26. kg/ m3 smallest cell size. vol. [3] Lee. Computers Fluids. pp. Fluids. 50. Y. m transverse coordinate (= y* / d ) Greek symbols Vy x Xd Xu y θ * temperature [= (T * − T∞ ) /(Tw − T∞ ) ] µ ρ δ b dynamic viscosity of the fluid. 1181-1023.. S. Int. J. Jenffer.

Int. pp... 182.. 1982. [10] Dalal. 6. A. A. 116. R.. [15] Igarashi. Sci. Heat Mass Transfer. vol. 1987. S. Global mode behavior of the stream wise velocity in wakes. J. A. E. 2006. A.. P. [11] Chattopadhyay. Numerical study of laminar forced convection fluid flow and heat transfer from a triangular cylinder placed in a channel. 2004. 1990. Fluids. Numerical investigation of forced convection in a plane channel with a built-in triangular prism. A. Eng. Numer.. S.1. 247 – 269. vol... A. vol. pp. 30. J. [8] Wesfreid. A. pp. G.. Heat and fluid flow across a square cylinder in the twodimensional laminar flow regime. Phys. vol. Int. T. A. and Nasrallah. and Chhabra. Thermal Sci.. A. Chem. vol. pp. and Dalal. [19] Sharma. J. P.. pp. and Zielinska. vol. 1343-1357. J. 2007. J. V... W. [18] Paliwal..5). 5315 – 5329. 1996. 1418-1424.. [14] Davis. 646-656. J. vol.[6] Jackson. vol. vol. 40. Int. Numerical Methods Fluids. 2008. J. pp. Numerical simulation of unconfined flow past a triangular cylinder. E. Heat Transfer. Power-law fluid flow past a square cylinder: momentum and heat transfer characteristics. J. vol. B. 45. pp. On the spatial structure of global modes in wake flow.. J. [17] Nitin.. and Eswaran. pp. Fluid flow and heat transfer around rectangular cylinders (in case of a width/height ratio of a section of 0. 238 – 259. Turki. 43.. Goujon-Durand. P. vol. pp. and Moore. V. Thermal Sci. 58. vol. F.. [13] De. Sci.... and Biswas.. Numerical simulation of flow around rectangular cylinders. pp. J. V. A numerical Study of vortex shedding from rectangles. B.. 46.. H. [12] Abbassi. B.. 893 – 901. vol. Heat Transfer A. J. Eswaran. pp. Physics II France. Sharma. R. 52. and Dalal.. C. A finite-volume method for Navier-Stokes equations on unstructured meshes. Eng. Heat Transfer B. H. K. Int. 13 . S. 23-45. A finite-element study of the onset of vortex shedding in the flow past variously shaped bodies. Augmentation of heat transfer in channel using a triangular prism. 475 – 506. pp. B. 2003. Fluid Mech. 501-505. vol. 707 – 720. pp.. 54. Aerodyn. Wind Ind.. 171 – 180. [16] Okajima.. and Wesfreid. J. Chhabra. Num. 7. and Eswaran. 1995.33 . K. 2001.. 1987. pp. 129. A. S.. J. 2005. J.. R. 801-821.. 649-658... E. 33.. 2007. [7] Zielinska. Fluid Mech. Non-isothermal flow of a power-law fluid past a rectangular obstacle (of aspect ratio 1 × 2) in a channel: drag and heat transfer. Int. [9] De.

M. Distribution of heat transfer coefficient around circular cylinder in cross flow at Reynolds numbers from 20 to 500.. Heat Mass Transfer. 1975. 17. Heat Transfer. F.. P. Chhabra. Breuer. Comput.. J. vol. vol. 41. D. A.. F. and Durst. [28] Karniadakis. Numer. pp.. 2009.. pp. and Eckert. M. Heat Mass Transfer. G. G. E. 4. and Eswaran. 199–264. [31] Nakamura.. Heat Mass Transfer. A. Sukhatme S. vol. Heat Transfer A. [29] Lange. [26] Krall.... K. 31–81. Local heat transfer around a cylinder at low Reynolds number.... 1998. 11. R. 2004. 2007. [22] Sahu. pp. Trans.. vol. 31. C. 49. vol. Heat Transfer A. V. R. pp. J. pp.. R. [23] Morgan. Int. vol. Effects of Reynolds and Prandtl numbers on heat transfer across a square cylinder in the steady flow regime.-M. 20. pp. Appl. vol. 74. Int. Heating effect on steady and unsteady horizontal laminar flow of air past a circular cylinder. vol. Biswas. 839 – 850. Phys.. T. vol. G. P. 16. R. Adv. 273 – 275. M. Variation of Nusselt number with flow regimes behind a circular cylinder for Reynolds numbers from 70 to 30000. J.. Numer. H. V.. V. Momentum and heat transfer from cylinders in laminar cross flow at 10-4 ≤ Re ≤ 200. Computation of unsteady momentum and heat transfer from a fixed circular cylinder in laminar flow. J. K. vol. Mech. [32] Shi. Effects of Reynolds and Prandtl numbers on heat transfer from a square cylinder in the unsteady flow regime. and Breuer. 47. and Soehngen. R. 2003. A. E. and Eswaran. F. ASME. Int. L... T. 2006. Gerlach. vol. Sharma. Heat transfer to power-law fluids from a heated square cylinder. 5169 – 5173. and Igarashi. S. Heat Transfer Eng. pp... [27] Sarma. Numerical simulation of forced convection heat transfer from a cylinder in cross-flow. J. Local heat transfer from a horizontal cylinder to air in cross-flow: influence of free convection and free stream turbulence. [24] Ahmad. [30] Baranyi. E. [25] Eckert. vol. pp. G.. J. A. 52.. [21] Dhiman. 51 – 56.. 52. 717 – 731... Steady-state numerical solution of the Navier-Stokes and energy equations around a horizontal cylinder at moderate Reynolds numbers from 100 to 500. 1977.. vol. P. 1996. V. J. 95. Durst. Heat Transfer.. pp. pp. P. and Eswaran. K. 1988. pp.. 1952. pp.[20] Dhiman. 343 – 347. Int. 107 – 118. Heat Mass Transfer. 2004. 185 – 201. 13 – 25... 1973. pp. E. 3409 3430 . 14 . Int. J. T. Chhabra. Fluids. K. Heat Mass Transfer. 4331 – 4345. A. R. The overall convective heat transfer from smooth circular cylinders. Chhabra.

. V. P.. I. vol.. 50. 15 . 2005. Unsteady fluid flow over an equilateral triangular cylinder in cross-flow. A. S. pp. Chem. 2009. 2007. vol. R. pp. A. P. Eng. Res. A. 43. S. and Lysenko. 977 – 990. Flow and forced convection heat transfer in cross-flow of non-Newtonian fluids over a circular cylinder. and Eswaran. Shashi. under review. A. High Temperature. M. P.. A. 5815 – 5827. Radhe.. pp. Chem. [36] Dhiman.. Heat Mass Transfer. vol. Des... J. A. R. J. Shyam. Kumar. Leontiev... 746 – 759. [34] Isaev.. Kudryavtsev. T. Steady forced convection heat transfer from a heated circular cylinder to power-law fluids. 44. Amit.[33] Soares.. [35] Bharti. R. Numerical simulation of unsteady state heat transfer under conditions of laminar transverse flow past a circular cylinder. A. Res. Ind.. and Chhabra.. Ferreira. Baranova. 2005. Eng. Int. N. D. Kumar. Chhabra.

j) magnified views for Re = 50. 120 and 150 Variation of (a) time-averaged Nusselt number and (b) rms value of Nusselt number with Reynolds number Variation of the Colburn heat transfer factor ( jh ) with Reynolds number for the fixed value of the Prandtl number 16 .Figure Captions Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Figure 9 Schematics of the unconfined flow across an equilateral triangular cylinder Non-uniform computational grid structure Instantaneous isotherm profiles at Reynolds number of 50 Instantaneous isotherm profiles at Reynolds number of 100 Instantaneous isotherm profiles at Reynolds number of 150 Instantaneous local Nusselt number around the triangular cylinder surfaces (a c) complete and (d .e) complete and (f . 100. 80.f) magnified views for Re = 50. 100 and 150 Time history of Nusselt number (a .

5961 5.001 d Time-averaged Nusselt number 5.71 Total number of cells in the domain 116500 100 116500 116500 Spacing of the first grid point from the cylinder 0.004 d 0.0015 d 0.5741 Re 17 .5843 5.Table 1: Grid independence study for the value of the Reynolds number of 100 and Prandtl number of 0.

Table 2: Validation of present flow results with literature values for Re = 50. 100 and 150 Present work Re St De and Dalal [9] CD _ St CD _ 50 100 150 0.1916 0.1505 0.5420 1.2015 1.8750 18 .2041 1.8937 0.1966 0.5257 1.7316 1.1455 0.7607 1.

0447 7. [10] for Re = 80. [10] 4.8711 4.Table 3: Validation of time-averaged Nusselt number with the results of Dalal et al.5843 5.67 7.954 5. 100 and 150 for Pr = 0.71 Source Re = 80 Re = 100 Re = 150 Present work Dalal et al.31 19 .

V x = 0.Tw U∝ d T∝ L2 y x Xu L1 Xd Figure 1 20 . V y = 0.Equilateral Triangular cylinder.

30 28 26 24 22 20 18 16 14 12 10 8 6 4 2 0 y 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 x Figure 2 21 .

5 15 14.15 13.15 0.22 0.29 0.65 0.15 (b) (f) y 0.15 0.43 0.15 14 12 10 12 20 18 16 14 16 18 20 22 24 26 28 30 32 14 13.5 15 14.43 0.5 16.07 0.22 0.5 15 14.5 y 0.07 0.22 y 14 0.5 12 10 12 20 18 16 14 16 18 20 22 24 26 28 30 32 14 13.5 0.5 16.07 14 13 14 15 16 (d) (h) 0.15 13 14 x 15 16 Figure 3 22 .5 16 (e) 0.07 16 (c) (g) 0.5 11 12 11 12 0.5 16 15.20 18 16 (a) 16.5 16.43 0.07 11 12 13 14 15 0.5 16 15.5 12 10 12 14 16 18 20 22 x 24 26 28 30 32 14 13.5 0.65 0.36 15.22 y 0.29 0.5 11 12 13 14 15 16 0.07 0.43 0.15 0.36 0.5 15 14.15 14 12 10 12 20 18 16 14 16 18 20 22 24 26 28 30 32 14 0.29 0.22 0.36 0.29 0.36 0.5 16 15.22 0.

29 0.22 y 14 0.15 0.5 16 15.07 16 0.29 y 14 x 15 16 Figure 4 23 .22 0.43 0.15 0.43 0.15 y 14 14.29 0.36 11 12 13 14 15 16 (b) (f) y 0.22 0.5 15 14.22 0.5 15 (e) 0.5 16 15.5 15 14.15 (d) (h) 0.29 0.07 11 12 13 14 15 0.15 0.29 0.5 16.5 16 15.5 11 12 13 14 0.5 12 10 12 14 16 18 20 22 x 24 26 28 30 32 14 13.5 16.36 0.15 0.5 12 10 12 20 18 16 14 16 18 20 22 24 26 28 30 32 14 13.36 0.22 0.5 16.22 0.5 15 14.07 14 11 12 13 14 15 16 (c) (g) 0.20 18 16 (a) 16.07 0.36 0.29 0.5 12 10 12 20 18 16 14 16 18 20 22 24 26 28 30 32 14 13.29 0.07 0.5 12 10 12 20 18 16 14 16 18 20 22 24 26 28 30 32 14 13.07 0.5 16 15.29 0.

5 16 15.07 0.07 0.5 16.07 0.5 16 15.43 0.5 16.29 11 12 13 14 15 0.36 0.15 y 14 12 10 12 20 18 16 14 16 18 20 22 24 26 28 30 32 14 13.29 0.07 0.5 15 14.15 0.22 0.29 0.07 0.07 15 14.46 0.22 0.15 0.15 0.36 0.07 15.36 0.15 0.36 0.15 0.5 16.29 0.20 18 16 (a) 16.29 0.15 0.07 11 12 13 14 15 16 (c) (g) 0.5 0.07 y 14 12 10 12 14 16 18 20 22 x 24 26 28 30 32 14 13.15 y 14 12 10 12 20 18 16 14 16 18 20 22 24 26 28 30 32 14 13.22 0.5 0.5 (e) 0.22 0.15 0.29 0.5 y 14 12 10 12 20 18 16 14 16 18 20 22 24 26 28 30 32 14 13.5 16 15.07 16 (d) (h) 0.5 11 12 13 14 x 15 16 Figure 5 24 .5 15 14.5 15 14.5 0.15 0.22 0.5 16 11 12 13 14 15 16 (b) (f) 0.22 0.

35 11 (a) Re = 50 30 25 t=0 t=T/4 t=2T/4 t=3T/4 (d) Re = 50 10 9 8 7 t B r Nu 20 15 10 5 0 0.5 (c) Re=150 (f) Re=150 Nu A 0.5 t B 1 1.5 b A 3 t Distance along cylinder surfaces B 1 1.5 r C 2 b 2.5 b A 3 1 0.5 t B 1 1.5 (b) Re=100 (e) Re=100 Nu 25 20 15 10 5 A 60 55 50 45 40 35 30 25 20 15 10 5 0 0 0.5 b A 3 t B 1 1.5 1 A b 6 5 4 3 2 C A 50 45 40 35 30 t B 1.5 r C 2 b 2.5 r C 2 2.5 Distance along cylinder surfaces t B 1 1.5 r C 2 b 2.5 Figure 6 25 .5 r C 2 2.5 r C 2 2.5 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0.5 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 0.

7031 3.5 7.3 6.7028 3.1 6 5.58 0 50 100 150 200 150 6.55 0 5.01 t 150 200 7 150 160 170 t 180 190 200 Figure 7 26 .24 Nu 5.65 3.2 0 50 100 7.8 (b) Re = 80 4.59 (h) Re = 100 Nu 5.22 6.85 4.585 5.4 6.09 7.87 4.7 5.05 Nu 6.21 200 210 220 230 240 250 260 270 280 290 300 7.67 3.3.9 4.07 (j) Re = 150 (e) Re = 150 7.7 4.703 3.02 7.66 3.6 6.35 5.7032 3.6 4.03 7.05 6.63 0 4.2 7 6.6 5.7034 3.7029 3.8 5.75 4.71 3.7033 3.25 160 170 180 190 200 210 220 (i) Re = 120 (d) Re = 120 6.2 6.7035 3.08 7.873 390 400 410 420 430 440 (f) Re = 50 (a) Re = 50 Nu 3.45 50 100 150 200 150 160 170 180 190 200 210 (c) Re = 100 5.65 4.15 5.9 5.06 7.872 4.23 6.25 5.871 4.869 (g) Re = 80 Nu 4.7026 50 100 150 200 250 300 350 400 450 380 4.69 3.64 3.68 3.7027 3.04 7.55 5.8 0 50 100 150 200 250 300 6.7 3.

5 7 6.015 Nurms 0.5 6 (a) Nu 5.5 4 3.005 0 50 60 70 80 90 100 110 120 130 140 150 Re Figure 8 27 .5 50 60 70 80 90 100 110 120 130 140 150 0.01 0.5 5 4.7.025 (b) 0.02 0.

065 0.075 0.055 0.08 0.0.06 0.07 jh 0.085 0.05 50 60 70 80 90 100 110 120 130 140 150 Re Figure 9 28 .

Sign up to vote on this title
UsefulNot useful