Aijrfans14 231

American International Journal of Research in Formal, Applied & Natural Sciences

Available online at http://www.iasir.net

ISSN (Print): 2328-3777, ISSN (Online): 2328-3785, ISSN (CD-ROM): 2328-3793 AIJRFANS is a refereed, indexed, peer-reviewed, multidisciplinary and open access journal published by International Association of Scientific Innovation and Research (IASIR), USA (An Association Unifying the Sciences, Engineering, and Applied Research)

Effect of Fin Configuration on Heat Transfer of an Annulus Tube 1

Dipti Prasad Mishra, 2Kailash Mohapatra Department of Mechanical Engineering, Birla Institute of Technology, Mesra, Ranchi, India 2 Department of Mechanical Engineering, Raajdhani Engineering College, Bhubaneswar, India 1

Abstract: Conservation equations of mass, momentum, energy have been solved numerically with two equations based k- model to determine fluid flow and heat transfer characteristics of a finned annulus tube.

The outer wall of outside tube is subjected to a constant heat flux and the outer wall of the inner tube is adiabatic. It is found from the numerical investigation that the heat transfer to the air increases with fin number present between the annulus space and also the heat transfer increases with the fin height. It is also obtained from the numerical experiment that trapezoidal shaped fin has a higher heat transfer rate compare to other shape.

Key words: Nusselt number, annulus tube, friction factor, heated surface, wall temperature

NOMENCLATURE C1: constant value of 1.44 C2: constant value of 1.92 D 1: outer diameter of annulus tube D 2: inside diameter of inner tube f: friction factor H: height of the fin h: heat transfer coefficient k: turbulent kinetic energy L: length of the tube n: number of fins Nu: Nusselt number p: pressure Prt: turbulent Prandtl number q: heat flux Re: Reynolds number

Sij :

1  U i U j   2  x j xi

T: t: U:

temperature time velocity

  

ui u j : 2 t Sij w: x:

 width of the fin axial location

Greek symbols μ: dynamic viscosity μt: turbulent viscosity ε: rate of dissipation σk: turbulent Prandtl number for k σε: turbulent Prandtl number for ε : a scalar variable either k or  : thermal diffusivity Subscripts avg: average B: bulk x: local wx: axial surface distance inf: ambient υ: kinematic viscosity ρ: density

I. INTRODUCTION Internally finned tubes are extensively used in many industrial applications to enhance heat transfer. Specifically in heat exchangers annular tubes with fins plays a very vital role in transferring the heat from the heated wall to fluid. In many industries such as power plants, chemical plants and petroleum industries attention is also focussed to reduce the size of the heat exchanger without sacrificing the heat transfer rate which can be achieved by changing the number, shape and size of the fins.

Experimental investigation performed by Braga and Saboya [1] with isothermal wall condition of inner tube and insulated outer wall having 20 numbers of fins. It was found that Nusselt number was exclusive function of Reynold number of the air flow and it was not depending on the thermal conductivity of the fin material. Heat transfer and pressure drop characteristics of a double pipe structure have been investigated experimentally by Yu et al. [2]. Friction factor and Nusselt number have been correlated with Reynolds number for blocked inner tube and unblocked inner tube. Experimental investigations were performed for turbulent flow through annular ducts with 560 internal pin fins to determine average heat transfer coefficients and friction factors where air was made to flow through the annular channel and water through the inner circular tube [3]. For a fully developed flow they concluded fin efficiency and thermal conductivity of fin material decreases with Reynolds number.

AIJRFANS 14-231; © 2014, AIJRFANS All Rights Reserved

Page 63

Dipti Prasad Mishra et al., American International Journal of Research in Formal, Applied & Natural Sciences, 6(1), March-May 2014, pp.

63-69

Numerical investigations were conducted by Li et al. [4] using three turbulent model namely mixing length model, k- model and full stress model for fully developed flow in annular ducts with four apex angles and four radius ratios. It was concluded numerical results agreed with the available correlation with maximum deviation of 8.3%. Correlations were developed Dirker and Meyer [5] for wide range of annular diameter ratios to predict Nusselt number for different Reynolds number based on hydraulic diameter. The deduced correlations predict the Nusselt number within 3% of experimental values. Numerical investigation conducted by Qiuwang et al. [6] shows there exists an optimal ratio inner core outside diameter to outer tube inner diameter for an annular wavy finned tube for both constant wall temperature and constant heat flux. Experimental investigation on unfinned and finned tube concentric passages have been performed by Kuvvet and Yavuz [7] shows the average Nusselt number is augmented in the finned passage 3.6 times to 18 times compared to unfinned passage. Alijani and Hamidi [8] performed experimental investigation on circular tube with core rod a honeycomb network inserts at a high temperature to determine the heat transfer coefficient and concluded that at uniform wall temperature the heat transfer coefficient of core rod was increased by 227% to 369% and 409% to 679% for honey comb network insert compared with plain tube.

In the present work we intend to analyze the heat transfer and fluid flow characteristics of an annular internal finned tube where the flow assumed to be turbulent and is developing in nature. The longitudinal fins are attached inside annular space between the tubes and the thermal performance has been investigated by varying the fin number, shape and size. We have also tried to match the present computation with the existing experiment of Sobaya and Braga [1] for average Nusselt number. All the computations have been performed using Fluent 14 for different flow conditions to determine the optimum fin condition for maximum thermal performance.

II. MATHEMATICAL FORMULATION The Computational investigation is carried out for an annulus finned tube of outer diameter, D1, inner surface diameter, D2 and length, L as shown in Fig. 1. Air enters into the annulus space of the tube at one end where as

a

b

Fig.1 (a) Longitudinal section of an annular tube with boundary condition applied to it (b) Cross sectional view of the annular tube showing the fins are attached inside the annular space

the other end is exposed to the surrounding atmosphere. Longitudinal fins are placed symmetrically within the annulus space around inner periphery of the tube. The investigation is initiated with the rectangular fins of length L, thickness t and height H as shown in Fig. 1 and subsequently the fin number, shape and sizes have been changed while we keep the mass flow rate of air inside the tube to be constant. The flow field in the domain would be computed by using three-dimensional, incompressible Navier-Stokes equations (2D axisymmetric model for fin less tube) along with the energy equations. The fluid used in the simulation is air, at temperatures of 300 to 700 K, and is treated to be incompressible, at the inlet face of the tube with an inlet velocity below 10 m/s.

A. GOVERNING EQUATIONS The governing equations for the above analysis can be written as: Continuity  ( Ui )0 xi

(1)

Momentum

 D  U i  p    U i U j  (2)          ui u j  Dt xi x j   x j xi    is the fluid density anywhere in the domain and is the function of temperature and taken as per the ideal gas equation and

ui u j is expressed as

AIJRFANS 14-231; © 2014, AIJRFANS All Rights Reserved

Page 64

Dipti Prasad Mishra et al., American International Journal of Research in Formal, Applied & Natural Sciences, 6(1), March-May 2014, pp.

63-69

ui u j  2

t 1  U U j Sij , Sij   i   2  x j xi

  

(3)

Energy Equation D  T     T     Dt xi  prt xi  Conduction Equation   T  k 0 xi  xi  Turbulence kinetic energy - k D   k  t 2 t  1   S      Dt  x j    k

(4)

(5)

 k     x j  Where S is the modulus of the mean rate-of-strain tensor, defined as

S  2Sij Sij

(6)

(7)

Rate of dissipation of k D           C1 t S 2  C2    Dt k   x j

1  t        

      x j 

(8)

t is expressed as

t   C

k2



(9)

k and  are the Prandtl numbers for k and .

The constants used in the above k- equations are the following C1  1.44; C2  1.92; C  0.09; k  1.0;   1.3 , Prt =1 B. BOUNDARY CONDITIONS Fig. 1 shows the boundary conditions for finned tube. The outer tube wall are solid and has been given a no-slip and constant heat flux boundary condition where as the outer wall of inner tube has been applied with adiabatic boundary condition. Pressure outlet boundary condition has been imposed at the outlet of the tube where as velocity inlet boundary condition has been employed at the inlet of the tube. Axi-symmetric boundary conditions have been used in case of the smooth tube. At the pressure outlet boundary, the velocity will be computed from the local pressure field so as to satisfy continuity but all other scalar variables such as, T, k and  are computed from the zero gradient condition, Dash [9]. The turbulent quantities, k and , on the first near wall cell have been set from the equilibrium log law wall function as has been described by Jha and Dash [10, 11] and Jha et al. [12]. The turbulent intensity at the inlet of the tube has been set to 2%, with the inlet velocity being known, and the back flow turbulent intensity at all the pressure outlet boundary have been set to 5%. If there is no back flow at a pressure outlet boundary then the values of k and  are computed from the zero gradient condition at that location.

C. COMPUTATIONS OF SOME IMPORTANT HEAT TRANSFER AND FLUID FLOW PARAMETERS The heat transfer and fluid flow parameters are computed as given in Yu et al. [2] and Islam and Mozumder[13].

Local Nusselt number for the finned heated tube

Nu x 

hx D kair

(10)

Local heat transfer coefficient has been computed as

hx 

q Tsx  TBx

(11)

Where hx, D, kair, Tsx and TBx are respectively, the local heat transfer coefficient, internal nominal diameter of tube, thermal conductivity of air, local tube wall temperature and local bulk temperature of fluid.

TBx  Ti 

q Pn x mc p

(12)

Ti, q, Pn, x, Cp and ṁ are respectively, inlet fluid temperature, wall heat flux, nominal perimeter, axial position, specific heat and mass flow rate of fluid.

m = Vin A

AIJRFANS 14-231; © 2014, AIJRFANS All Rights Reserved

(13)

Page 65

Dipti Prasad Mishra et al., American International Journal of Research in Formal, Applied & Natural Sciences, 6(1), March-May 2014, pp.

63-69

, Vin and A are the density, inlet fluid velocity and area of tube at the inlet respectively.

The Reynolds number has been computed based on nominal diameter of the tube and is given by

Re 

Vin Dh

(14)

 air

air is the kinematic viscosity of air based on inlet condition

Nuavg 

havg D

(15)

kair

havg 

q Tsavg  TBavg

(16)

Where Nuavg, havg, Tsavg and TBavg are respectively, the average Nusselt number, average heat transfer coefficient, average wall temperature and average bulk temperature.

f x  2P

D 1 x Vin2

(17)

Where fx, P,  and x are local friction factor, cumulative pressure drop, density and axial distance along the tube respectively.

III. NUMERICAL SOLUTION PROCEDURE Three-dimensional equations of mass, momentum, energy and turbulence have been solved by the algebraic multi grid solver of Fluent 14 in an iterative manner by imposing proper boundary conditions. First order upwind scheme (for convective variables) was considered for momentum as well as for the turbulent discretized equations. As the density is taken as a function of temperature according to ideal gas equation therefore SIMPLE algorithm with PRESTO (Pressure Staggered Option) scheme has been used for better convergence.

Under relaxation factors (0.3 for pressure, 0.7 for momentum and 0.8 for k and ) were used for the convergence of all the variables. Convergence of the discretized equations were said to have been achieved when the whole field residual for all the variables fell below 10 -3 except energy equation and for energy equation residual was set 10-6.

IV. RESULTS AND DISCUSSION A. MATCHING WITH OTHER COMPUTATION We have tried to compare the existing experimental result of Sobaya and Braga [1] with the present CFD for average Nusselt number. Here in the annulus section the working fluid is air and constant temperature boundary condition is applied over the surface of the inner tube and adiabatic boundary condition is imposed over the outside tube surface. It can be seen from the Fig. 2 the present CFD result matched reasonably well with the existing experimental result with a maximum error of 6%.

70 Experiment (Saboya and Braga, [1]) Present CFD

60

Nuavg

50 40 30 20 10 1

2

3 -4 Rex10

4

5

Fig. 2 Average Nusselt number as a function of Reynolds number: A comparison between existing experiment and present CFD

B. GRID SENSITIVITY TEST D1 = 70 mm, D2 = 50 mm, n = 4, t = 5 mm, q = 3000 W/m2, ṁ = 0.0196 kg/s Cell Number

Outlet Temperature (To/Tinf)

16800 69930

1.138 1.13

237500

1.126

559500

1.1202

1791500

1.119

AIJRFANS 14-231; © 2014, AIJRFANS All Rights Reserved

Page 66

Dipti Prasad Mishra et al., American International Journal of Research in Formal, Applied & Natural Sciences, 6(1), March-May 2014, pp.

63-69 Table 1 Outlet temperature as a function cell number

Grid sensitivity test has been shown in Table – 1 for a sample case of an annulus internal finned tube having rectangular section. The tube length is 1 m, having internal diameter of outer tube of 0.07 m and outer diameter of internal tube of 0.05 m which is subjected to a constant heat flux of 3000 W/m2 and other operating parameter can be seen from the table – 1 itself. Initially coarse grids have been taken and subsequently cells are refined and it is seen that after refining the entire meshes inside the tube from 16, 800 to 2,37, 500, the outlet temperature decreases by 1.5%. After that we refined the meshed near walls of tube and fins twice (a cross sectional view of the pipe can be seen from the Fig. 3) which increases the mesh number to 17,91,500. It was found that after increasing the cell number more than 5,59,500 does not have a significant change in outlet temperature. So finally we have considered 5,59,500 meshes to perform all the computations. However when the shape and of the internal fins are changed we perform the grid test before we report any results.

Fig. 3 Cross sectional view of the internal finned annulus tube showing the meshes inside the tube and around the fins

C. FIN NUMBER Effect of fin number on wall temperature can be seen from Fig. 4. In the present numerical investigation longitudinal fins having rectangular section have been used and other parameters can be seen from the plot itself. It can be seen from the Fig. 4 as the fin number increases the wall temperature reduces and the outlet air 1.18 1.16

1.1225 1.1220

1.12 To/Tinf

Tw/Tinf

1.14

2

q = 3000 W/m . m = 0.0196 kg/s, t = 5 mm D1 = 70 mm, D2 = 50 mm

1.1230

n=4 n=8 n = 12 n = 16 n = 20

1.10 1.08

1.1215 1.1210

2

q = 3000 W/m t = 5 mm . m = 0.0196 kg/s D1 = 70 mm, D2 = 50 mm

1.06 1.04 0.0

0.2

0.4

x/L

0.6

0.8

1.1205 1.1200

1.0

2

Fig. 4 Effect of fin number on wall temperature

4

6

8

10 12 14 Fin Number

16

18

20

22

Fig. 5 Outlet temperature as a function of fin number

temperature increases (can be seen from Fig. 5). Fig. 6 shows that the Nusselt number increases with fin number which suggests more heat transfer to air as the fin number increases. It is also visualised from Fig. 6 the Nusselt number is higher at the entrance of the pipe and it is continuously decreases towards the end. This implies initially at the entrance convection heat transfer is dominated and towards the end conduction heat transfer is more prominent because of growth of boundary layer.

2

240

q = 3000 W/m H = 10 mm, t = 5 mm D1 = 70 mm, D2 = 50 mm

210

.

m = 0.0196 kg/s

n=4 n=8 n = 12 n = 16 n = 20

Nux

180 150 120 90 60 0.0

0.2

0.4

0.6

0.8

1.0

x/L

Fig. 6 Nusselt number as a function of fin number

AIJRFANS 14-231; © 2014, AIJRFANS All Rights Reserved

Page 67

Dipti Prasad Mishra et al., American International Journal of Research in Formal, Applied & Natural Sciences, 6(1), March-May 2014, pp.

63-69

D. FIN HEIGHT 300 1.20 1.18 1.16

D2 = 60 mm

2

D2 = 50 mm D2 = 40 mm D2 = 30 mm

250

q = 3000 W/m . m = 0.0196 kg/s D1 = 70 mm, n = 4

200

t = 5 mm

Nuavg

Tw/Tinf

1.14 1.12 1.10

150 100

1.08

2

q = 3000 W/m , n = 4 t = 5 mm, D1 = 70 mm

1.06

50

.

m = 0.0196 kg/s

1.04 0.0

0.2

0.4

x/L

0.6

0.8

0 0.05

1.0

Fig. 7 Effect of fin height on wall temperature

0.10

0.15

0.20 H/D1

0.25

0.30

Fig. 8 Average Nusselt number as a function fin height

Fin height plays a crucial role in transferring the heat in internal finned annular tube. In the annular tube the outer tube diameter is kept constant with 70 mm where as inside tube diameter varies with fin height. It can be seen from the Fig. 7, increasing the fin height or reducing the inner tube diameter, the wall temperature decreases where as average Nusselt number increases as seen from Fig. 8. This clearly indicates the heat transfer to air increases with the fin height. But it has an adverse effect on air flow as the pressure drop is high and due to that the friction factor becoming more (can be seen from Fig. 9).

D2 = 50 mm q = 3000 W/m

0.08

2

D2 = 40 mm n = 4, H = 10 mm

.

D2 = 30 mm m = 0.0196 kg/s

D1 = 70 mm, t = 5 mm

0.07

fx

0.06

0.05

0.04

0.03 0.0

0.2

0.4

0.6

0.8

1.0

x/L

Fig. 9 Friction factor as a function of fin height

E. FIN SHAPE Rectangular

1.35

84

+ shape Triangular

1.30

Trapezium

82

Plane tube

80

1.20

Nuavg

Twx/Tinf

1.25

2

q = 3000 W/m , n = 4 . m = 0.0196 kg/s D1 = 70 mm, D2 = 50 mm

1.15

78 76

1.10

2

q = 3000 W/m , t = 5 mm . m = 0.0196 kg/s, n = 4 D1 = 70 mm, D2 = 50 mm

1.05

0.0

0.2

0.4

0.6

0.8

1.0

x/L

Fig. 10 Effect of fin shape on wall temperature

74 72 + - shape

Rectangular

Triangular

Trapezium

Fin shape

Fig. 11 Nusselt number as a function of fin shape

Fig. 10 shows the wall temperature distribution for different shape fin having same inlet mass flow rate and fin height. It can be seen from the plot the wall temperature is minimum in trapezoidal shaped fin compare to other

AIJRFANS 14-231; Š 2014, AIJRFANS All Rights Reserved

Page 68

Dipti Prasad Mishra et al., American International Journal of Research in Formal, Applied & Natural Sciences, 6(1), March-May 2014, pp.

63-69

shape which depicts the heat transfer is highest in trapezoidal shaped fin. The same result is also visualised from Fig. 11, which shows higher value of Nusselt number for trapezoidal shaped fin compare to other.

V. CONCLUSIONS The wall temperature, tube outlet temperature and friction factor for an internal annulus finned tube have been computed by solving the conservation equations for mass, momentum and energy. The average Nusselt number of the existing experimental observation matched well with the present CFD result. The heat transfer to the air is increases with fin height and fin number. The heat transfer is maximum in trapezoidal fin compare other shape.

REFERENCES [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13]

Braga C V M and Saboya F E M (1999) Turbulent heat transfer, pressure drop and fin efficiency in annular regions with continuous longitudinal rectangular fins. Experimental thermal and fluid science 20:55-65.

Yu B, Nie J H, Wang Q W, and Tao W Q (1999) Experimental study on the pressure drop and heat transfer characteristics of tubes with internal wave-like longitudinal fins Heat and mass transfer 35:65-73.

Nieckele A O and Mouraosaboya F E (2000) Turbulent heat transfer and pressure drop in pinned annular regions. Journal of Brazilian Society of Mechanical Science 22:119-132.

Li Z Y, Hung T C and Tao W Q (2001) Numerical simulation of fully developed turbulent flow and heat transfer in annularsector duct. Heat and mass transfer, 38:369-377.

Dirker J and Meyer JP (2005) Convective heat transfer coefficients in concentric annuli. Heat transfer engineering 26(2):38-44.

Qiuwang W, Mei L and Min Z (2008) Effect of blocked core-tube diameter on heat transfer performance of internally longitudinal finned tubes. Heat transfer Engineering 29(1):107-115.

Kuvvet K and Yavuz T (2011) Effect of fin pitches on heat transfer and fluid flow characteristics in the entrance region of a finned concentric passage. Journal of thermal science and technology 31(2):109-122.

Alijani M R and Hamidi A A (2013) Convection and radiation heat transfer in a tube with core rod and honeycomb network inserts at high temperature. Journal of mechanical science and technology 27(11):3487-3493.

Dash S K (1996) Heatline visualization in turbulent flow. Int. J. Numerical Methods for Heat and Fluid Flow 6:37-46.

Jha P K, Dash S K (2002) Effect of outlet positions and various turbulence models on mixing in a single and multi strand tundish.

Int. J. Num. Method for Heat and Fluid Flow 12:560-584.

Jha P K, Dash S K (2004) Employment of different turbulence models to the design of optimum steel flows in a tundish. Int. J. Numerical Methods for Heat and Fluid Flow 14:953-979.

Jha P K, Ranjan R, Mondal S S, Dash S K, (2003) Mixing in a tundish and a choice of turbulence model for its prediction. Int. J. Numerical Methods for Heat and Fluid Flow 13: 964-996.

Islam M A, Mozumder A K (2009) Forced convection heat transfer performance of an internally finned tube. Journal of Mechanical engineering ME 40:54-62.

AIJRFANS 14-231; Š 2014, AIJRFANS All Rights Reserved

Page 69