DEPARTMENT:

COURSE:
STUDENT NAME:
ISTITUTION:
PROFESSOR:

Contents
A B S T R A C T.............................................................................................................................................................................................................. 3
LI S T O F TA B L E S...................................................................................................................................................................................................... 4

1
LI S T O F FI G U R E S................................................................................................................................................................................................... 4
N O M E N C L AT U R E................................................................................................................................................................................................ 5
I N T R O D U C T I O N.................................................................................................................................................................................................. 6
LITERATURE REVIEW.................................................................................................................................................................................................... 9
A N A LY S I S.............................................................................................................................................................................................................. 10
Exact solution......................................................................................................................................................................................................... 12
I V. R E S U LT S.......................................................................................................................................................................................................... 18
V. D I S C U S I O N O F R E S U LT S............................................................................................................................................................................ 25
V I.C O N C L U S I O N S............................................................................................................................................................................................. 28
VII. ACKNOWLEDGEMENT:........................................................................................................................................................................................ 29
A P P E N D I C E S....................................................................................................................................................................................................... 30
Program listing....................................................................................................................................................................................................... 30
R E F E R E N C E S....................................................................................................................................................................................................... 31

2
A REPORT ON HEAT TRANSFER ON FIN PIN

ABSTRACT
Numerical determination of the temperature distribution T(x) and fin heat transfer rate qf of a cylindrical pin fin is presented.
The classical Runge-Kutta 4th-order method is used to solve a 2nd-order ODE governing the temperature profile. The shooting
method converts the boundary value problem into an initial value problem. Multiple iterations are performed until the solution
converges to an adiabatic condition at the fin tip. Numerical results at four different step sizes ∆x are compared to the analytical
solution. The R-K 4th-order method produces a fin heat transfer rate qf accurate to 3 decimal digits at the largest step size ∆x =
0.2. The use of four appropriately weighted derivative functions permits the extremely high accuracy of R-K 4th-order method in
predicting the y-value jumps over each step ∆x. The major downside is that the shooting method/RK4 method is computationally
expensive, which is chiefly due to the requirement of conducting multiple iterations before convergence.

3
LI S T O F TA B L E S
4.1 Temperature distribution for step size ∆x = 0.2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

6 4.2 Temperature distribution for step size ∆x = 0.1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

7 4.3 Temperature distribution for step size ∆x = 0.05 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

8 4.4 Temperature distribution for step size ∆x = 0.01 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
9 5.1 Error in numerical θ ′ (0) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
13 5.2 Numerical fin heat transfer rate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
13 6.1 Temperature distribution and heat loss for fins of uniform cross section [1]

LI S T O F FI G U R E S
1.1 Fin geometries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
2.1 Cylindrical fin problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
2.2 Energy balance over a cylindrical section . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
4.1 θ(X) distribution plot for ∆x = 0.2, 0.1, 0.05, 0.01 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
4.2 θ ′ (X) distribution plot for ∆x = 0.2, 0.1, 0.05, 0.01 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
6.1 T(x) distribution plot for ∆x = 0.2, 0.1, 0.05, 0.01 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

4
N O M E N C L AT U R E

T temperature in either ◦C or Kelvin

T0 temperature at fin base

T∞ temperature of ambient fluid θ non-dimensionalized excess temperature such that x distance from fin
base in meters
X non-dimensionalized distance from fin base in meters where X = x/L
L length of fin in meters
R radius of fin in meters
P perimeter of fin in meters where P = 2πR for a cylindrical fin
Ac cross-sectional area of fin in m2 ; Ac = πR 2 for a cylindrical fin As surface area of fin in m2
k thermal conductivity in W/m·K
h convection coefficient in W/m2 ·K

m constant defined such that

p constant defined such that

qf fin heat transfer rate in Watts
∆x step size used for the numerical solution of an ordinary differential equation describing temperature distribution of a fin
Y1 defined as θ(x) for the Runge-Kutta 4th-order method when numerically solving for the temperature distribution of a fin
Y2 defined as θ ′ (x) in the R-K method

5
INTRODUCTION
Fins are extended surfaces used specifically to enhance the rate of heat transfer between a solid material and an adjoining fluid.
[2] Practical applications include radiators in cars, computer CPU heat sinks, and heat exchangers in power plants.[3, 7, 10] In
thermal systems design, engineers are primarily interested in knowing the extent to which particular extended surfaces or fin
arrangements could improve the heat transfer from a surface to the surrounding fluid.[1] Thus, a major objective is to optimize
fin geometry and material so as to augment heat transfer.[6] In order to determine a particular fin’s heat transfer rate qf , it is
necessary to obtain the temperature distribution T(x) along the fin. Real-world analysis of fin design is often complex.
Investigators frequently adopt idealizing assumptions when analyzing extended thermal surfaces. These assumptions are
attributed to researchers Murray[9] and Gardner[4] and are as follows: (1) the heat flow and temperature distribution
throughout the fin are independent of time, i.e. steady-state heat flow (2) the fin material is homogenous and isotropic (3) there
is no volumetric heat generation q˙ within the fin (4) the heat flow to or from the fin surface at any point is directly proportional
to the temperature difference between the surface at that point and the surrounding fluid (5) the thermal conductivity of the fin
k is constant (6) the convective transfer coefficient h is the same all over the fin surface (7) the surrounding fluid has uniform
temperature T∞ (8) the temperature at the base of the fin T0 is uniform (9) fin thickness t is much smaller than its height l so
that temperature gradients normal to the surface are neglected (10) heat transfer through the outermost edge of the fin is
negligible compared to which passes through the sides Under these highly idealized scenarios it is possible to develop analytical
solutions for the temperature distribution and fin heat transfer rate of a wide range of fin configurations, especially when
considering the simple fin geometries as in Fig. 1.1. However, oftentimes the equations that arise in engineering problems such
as those in heat transfer are too complicated to be solved analytically, and computational techniques must be used to obtain the
numerical values needed.[5] Even in cases where numerical techniques are used it is still advantageous to reduce the problem to
an idealized case with a known analytical solution. The error between the numerical (approximate) solution and the analytical
(exact) solution in the simplified case gives an indication of how much error will be involved in the numerical solution of the
more complex scenario that does not have an analytical solution or whose analytical solution is not able to be computed.

6
 fig1
To demonstrate the numerical process of solving an ordinary differential equation (ODE), the engineering system is a cylindrical
pin of uniform cross-sectional area subject to the boundary conditions of an insulated fin tip dT dx |x=L = 0 and known
temperature at the base T0. The goal is to determine how temperature varies over the distance of the fin T(x) and the heat flux
qf . The problem begins by performing a differential energy balance over an infinitesimal fin section. When assuming
onedimensional temperature distribution, a linear 2nd-order ODE is developed and is used to relate temperature as a function of
distance. The methodology for solving the ODE numerically requires conversion the 2nd-order ODE into a system of two 1st -
order ODE’s. Because the problem is a boundary value problem, the shooting method is invoked. The derivative of the
temperature with respect to distance at the beginning dT dT |x=0 is guessed, i.e.,—the BVP is converted into an initial value
problem (IVP). In order to achieve the correct slope at the fin tip dT dx |x=L = 0, the first two predictions are "corrected" by using
a Taylor’s series approach. This is the shooting method, which relies on the two most recent guessed slopes at the beginning, the
corresponding slopes at the fin tip that arise from these guesses, and the target boundary condition that the slope is zero at the
tip. For each guess, a solution of the temperature distribution is found by using a forward integration approach. In this article,
the most widely used formula for solving systems of ODE’s is implemented—the 4th-order Runge-Kutta method. The method is
derived by using different weighted averages of four derivative functions at several locations over the interval xi ≤ x ≤ xi+1. [8]
Advantages to the R-K method include:
 High-accuracy level
 Good stability considerations
 Ease in programming
 Self-starting scheme
 Ease with which the step size ∆x can be changed
Disadvantages include:

7
  Requires more computer time
 Difficult to estimate the local error

Fig2

LITERATURE REVIEW
Pin fin heat transfer simulation and modeling have received a lot of attention in the subject of heat transfer. The following assessment of the
literature summarizes some of the significant contributions made in this project.

8
The textbooks Introduction to Heat Transfer and Fundamentals of Heat and Mass Transfer by Bergman and Incropera contain a review of heat
transfer principles, including pin fin heat transfer. These texts lay the groundwork for understanding heat transfer physics, which is essential for
building correct models for pin fin heat transfer.

Design Precepts for High-Temperature Heat Exchangers by Fraas examines the design concerns for pin fin high-temperature heat exchangers.
The research emphasizes the significance of addressing the system's fluid dynamics and temperature distribution. When constructing heat
exchangers, keep the fins in mind. This paper offers an insightful look at the practical uses of pin fin heat transfer.

Gardner's study, Efficiency of Extended Surfaces, gives a theoretical examination of the heat transfer efficiency of pin fins. The study generates
efficiency formulas for pin fins of various shapes and sizes, which can be utilized to optimize the design of pin fins for specific applications. This
research sheds light on the fundamental physics of pin fin heat transfer.

The paper Performance and Optimal Dimensions of Different Cooling Fins with a Temperature-Dependent Heat Transfer Coefficient by Klaor and
Kalman examines the performance of pin fins under various operating situations. The research investigates the effect of temperature-dependent
heat transfer coefficients on pin fin performance and provides advice for selecting the appropriate pin fin diameters for certain applications. This
research provides useful insights into the practical considerations for developing heat transfer pin fins.

Overall, the references given above provide an in-depth look at pin fin heat transfer simulation and modeling. Bergman and Incropera's writings
lay the groundwork for comprehending heat transfer concepts, while Fraas, Gardner, Jaluria, Klaor, and Kalman provide essential insights into
practical applications.

Computer approaches for Engineering with MATLAB® Applications by Jaluria presents an overview of numerical approaches for handling heat
transfer problems, such as pin fin heat transfer. The book includes examples of how to use MATLAB® to handle pin fin heat transfer problems,
making it a useful resource.

A N A LY S I S
The engineering system is a pin fin of constant cross-sectional area shown in Fig. 2 with specified geometric and material
properties. The boundary conditions are as follows At x = 0 T = T0 At x = L:dT dx = 0 The following 2nd-order ordinary differential
equation is yielded for the temperature T, if the temperature distribution is assumed to be one-dimensional in x, where x is the
distance from the base of the fin[5]

9
 By letting and the ODE is conveniently rewritten where,

If the numerical values for the constants h, P, k, Ac, and L shown in Fig. 2 are substituted into Eq. (2.3) then p2 = 16.666. The
governing equation for the temperature profile of a cylindrical pin fin is a 2nd-order linear ODE written as d 2θ dX2 = 16.666θ
(2.4)The independent and dependent variables are non-dimensional such that θ ranges from 0 < θ < 1 over the length of the fin 0
< X < 1.
Derivation of equations

The derivation begins by applying a control volume energy balance over a differential element shown in Fig. 2.2

Fourier’s law yields

In the case of a straight pin fin, the cross-sectional area Ac is constant along the length, however, in the most general case Ac
varies with x. The conduction heat rate x + dx is expressed as

Substituting Eq. (2.6) into Eq. (2.7) as written below:

The convection heat transfer rate is expressed in differential form,

10
 Where dAs is the surface area of the differential element. Substituting Eqs. (2.6), (2.8), and (2.9) into the energy balance Eq. (2.5)
and rearranging

The cylindrical fin in this problem has uniform cross-sectional area Ac and surface area As = Px where P is the fin perimeter.

fig2.2
Thus, and so that Eq. (2.11) reduces to Eq. (2.1)

The excess temperature θ is defined as

The excess temperature at the base of the fin (x = 0) is

The dependent variable T(x) is transformed in terms of θ knowing that T∞ is constant and Therefore governing
equation becomes:-

11
 Where
The general solution to the linear, homogeneous, second-order differential equation with constant coefficients takes the form

where C1 and C2 are solved based on the boundary conditions.

The other important parameter in fin design is the fin heat transfer rate qf , which gives the rate of heat transfer over the entire
fin. This parameter is commonly found by applying Fourier’s law at the fin base

Exact solution
The boundary conditions for this problem are that of a specified temperature at the base T0 = 80◦C and an insulated fin tip
. The independent variable x is nondimensionalized by the length L such as shown below:

In addition, the excess temperature θ(X) is nondimensionalized such that,

The transformed (dimensionless) boundary conditions become,

12
 As a result of redefining the independent variable to be X = x/L the constant m must be multiplied by L. m is redefined in terms of

p so to maintain dimensional homogeneity so that

The general solution of the non-dimensional ODE becomes

In order to solve for the temperature distribution in terms of θ(X), which is now dimensionless, the derivative of the general
solution is computed

The first boundary condition, θ(0) = 1, is substituted into the general solution Eq2.16

The second boundary condition, dθ dX |X=0 = 0, is substituted into the derivative of the general solution Eq. (2.23)

The system of two equations based on the boundary conditions allows for the solution of the two unknown coefficients C1 and
C2

13
 Solving for the values of C1 and C2 that make physical as shown below:

The coefficients are substituted into the general solution. After substitution for C1 and C2 the solution in Hyperbolic secant and

cosine identities are used such that and The non-dimensional solution for the temperature distribution for the case of an adiabatic

fin tip in Eq. (2.28) is more concisely represented as,

This excess temperature is non-dimensional such that at the base θ(X = 0) = 1 and at the tip θ(X = 1) = 0. See Fig. 4.1 to see the plot
of Eq. (2.31). When dimensions are returned to the independent and dependent variables, the temperature distribution and fin heat
transfer rate for the adiabatic condition are expressed as

14
The excess temperature θ(x) shown above has dimensions and is equal to T(x) −T∞. More precisely the equations can be expressed

in terms of the temperature in ◦C

III.NUMERICAL SOLUTION
In order to solve numerically the temperature distribution θ(x) for

which has dimensionless independent and dependent variables and boundary conditions θ(0) = 1 and θ ′ (1) = 0, the 2nd-order
ODE is converted into a system of two 1st-order ODE’s by redefining the variables let

The classical Runge-Kutta 4th-order method is used. The slope functions for each redefined variable are

15
Since there is a system of two 1st-order equations Y1 and Y2, there are two sets of four k coefficients that used to estimate the
vertical jump over the width of the interval or step size ∆x for each equation in the system. These are defined as follows

Appropriate weights are assigned to each k-value to give a good estimate of the vertical jump for Y1 and Y2 as they move a
horizontal distance.

The beginning conditions θ(0) and  θ′ (0) are required for the R-K technique to begin. However, only  θ(0) = 1 is supplied in this
problem. The boundary value issue (BVP) must essentially be changed into an initial value problem (IVP). Multiple rounds are

16
 undertaken in which the initial slope θ′ (0) is predicted until the target slope θ ′ (1) = 0 is reached. This is the shooting technique.
The R-K approach is used for each guess θ′ (0) to achieve a temperature distribution along the whole length of the fin from 0 <X
<1. Two initial iterations are carried out. The slopes at the beginning and end of the fin, which correspond to the first two
possibilities, are utilized to compute the third guess for the initial slope, which is done using a Taylor series method. The slopes
of the two most recent iterations are then utilized to compute the next guess. No more iterations are conducted after the slope
at the end is sufficiently close to the desired slope, i.e., within 6 significant figures; the solution for the temperature distribution
is discovered. Two initial slopes are estimated for the specific situation.

The R-K method is then performed for each initial guess using an arbitrary step size ∆x. Each iteration produces a slope at the fin
tip θ ′ 1 (1) and θ ′ 2 (1). The initial and final slopes are used to compute the next guess θ ′ 3 (0) by the shooting method. In
general, using the two most recent iterations i and i −1 the prediction for the initial slope θ ′ i+1 (0) for the next iteration is found
by

where

The constant B represents the boundary condition or target slope at the fin tip B = θ ′ (1) = 0 (adiabatic condition at fin tip) (3.44)

I V. R E S U LT S
A discrete temperature profile is found using the classical Runge-Kutta method in conjunction with the shooting method for four
step sizes ∆x = 0.2, 0.1, 0.05, and 0.01. The governing ODE is

17
 with boundary conditions θ(0) = 1 θ ′ (1) = 0 where variables are dimensionless

18
19
20
21
22
23
24
V. D I S C U S I O N O F R E S U LT S

In non-dimensional form analytically, the results for the numerical determination of the non-dimensional first derivative, which are
derived from the R-K and shooting methods at different step sizes, are tabulated in Table 5.1. The error δ between the numerical
result and the analytical (exact) value for θ ′ (0) is defined result as

As the step size gets smaller, the non-dimensionalized derivative of the excess temperature with respect to the length at the base
dθ/dX|X=0 approaches the analytical value. By a step size ∆x = 0.01 the analytical and numerical derivative agree to 8 decimal
places. In order to obtain the fully dimensional fin heat transfer rate qf , analytically it is determined as qf analytical

25
This value represents the true fin heat transfer rate for the idealized system and has dimensions of Watts, which in SI base units is
kgm2 /s 3 . The fin heat transfer rate is constant throughout the length of the fin. Using the temperature at the base T0 to calculate
qf is convenient because it only requires application of the conduction energy balance (Fourier’s law).

26
 Numerical fin heat transfer rate ∆x qf Alternatively, the analytical fin heat transfer rate can be calculated as qf analytical = −k

Acθ0 dθ dX !
The numerical fin heat transfer rate is determined for each step size ∆x and results are tabulated in Table 5.2. The
nondimensional θ ′ (X) is transformed back into a dimensional form and used to calculate qf as follows

Where,

The error in the fin heat transfer rate qf defined δ is identical to the error in the first derivative of excess temperature with
respect to the distance at the base. This makes sense because the fin heat transfer rate qf is dθ/dX|X=0 multiplied by k, Ac, θ0,
and 1/L. The relative error is identical for a linear transformation. The difference in the fin heat transfer rate qf between the
analytical and numerical cases is extremely small. At the largest step size ∆x = 0.2, (qf ) numerical is agrees with (qf )analytical to
3 digits to the right of the decimal. At a step size of ∆x = 0.01, (qf ) numerical shares an astounding 12 digits to the right of the
decimal with (qf )analytical. This is a testament to the accuracy of the Runge-Kutta 4th-order method. Even at large step sizes the
error is well within tolerances for most engineering applications.

27
V I.C O N C L U S I O N S
The Runge-Kutta 4th-order method, combined with the shooting method, is an effective method for solving ODE boundary value
issues. Despite the fact that at least four iterations were required before a converged solution was obtained, the temperature
distribution θ(X) for even huge step sizes was pretty close to the analytical solution. When the problem is solved numerically,
discrete temperature readings along the fin are given instead of a continuous curve. The smaller step size ∆x = 0.01 had the
advantage of bringing the discrete θ (X) and θ′ (X) values even closer to the analytical values. Furthermore, the data had greater
resolution at the smaller step size, with more discrete points from 0<x< L. Without knowing the analytical solution, a lesser step
size may be preferred just to see the shape of the curve. A major shortfall of the shooting method and the R-K 4th - order
method is that it is computationally expensive. More than 1 iteration is required before a valid solution is obtained. This is the
greatest source of computer time because all calculations must be repeated for each iteration or set of initial guesses. For each
step in a single iteration, multiple derivative functions k1, k2, k3, and k4 must be determined. For the forth-order ODE there will
be n sets of k values to be calculated for each step. Having multiple slope functions enhances the accuracy of the prediction for
the next θ-value or θ ′ -value. At the same time, it is computationally expensive. For higher order ODE’s this may be more of an
issue than with the current 2nd-order ODE in the problem. The benefits of the R-K 4th-order method are that it is self-starting
and self-correcting and that it has extremely low error upon convergence with the boundary conditions. As discussed before, the
numerical dθ/dX|X=0 and the fin heat transfer rate qf at even the largest step size ∆x = 0.2 showed < 0.001 % error. At step size
of ∆x = 0.01 the error was on the order of 1×10−8 %. If only the fin heat transfer rate qf is the value that is desired then using a
large step size (∆x = 0.2) with a lot less computations produces still quite an accurate result (within 3 digits to the right of the
decimal).

28
VII. ACKNOWLEDGEMENT:
In the name of Allah, the most gracious and the most merciful. First and foremost, I am thankful to Almighty ALLAH for giving me
the strength, knowledge, ability and opportunity to undertake this study and complete it satisfactorily.

Secondly, I would like to thank my respected teacher and supervisor, whose worthy guidance and professional attitude is appreciable
in completing this dissertation. I am also thankful to my Department of mechanical for providing facilities to carry out this research
work.

I am deeply grateful to my parents for their support, appreciation, encouragement and keen interest in my academic achievements.

29
APPENDICES
Program listing
Microsoft Excel was used to generate the tabulated data for the Runge-Kutta 4th-order iterations at each step size ∆x. LATEX was
used to graph the data and format this document. No proprietary code was used.
A.2 Other Below is LATEX PGF Plots code used to generate the θ ′ (X) vs. X graph at a step size of ∆x = 0.2

30
REFERENCES
1. Theodore L Bergman and Frank P Incropera. Introduction to heat transfer. John Wiley & Sons, 2011.
2. Theodore L Bergman, Frank P Incropera, and Adrienne SLavine. Fundamentals of heat and mass transfer. John Wiley & Sons, 2011.
3. AP Fraas. Design precepts for high-temperature heat exchangers. Nuclear Science and Engineering, 8(1):21–31, 1960.
4. Karl A Gardner. Efficiency of extended surfaces. Trans. Asme, 67(1):621–631, 1945.
5 Yogesh Jaluria. Computer Methods for Engineering with MATLAB® Applications. Taylor & Francis, 2011.
6. KLaor and HKalman. Performance and optimum dimensions of different cooling fins with a temperature-dependent heat transfer
coefficient. International Journal of Heat and Mass Transfer, 39(9):1993–2003, 1996.

31