BAREKENG: Jurnal Ilmu Matematika dan Terapan

March 2022 Volume xx Issue xx Page xxx–xxx

P-ISSN: 1978-7227 E-ISSN: 2615-3017
https://doi.org/10.30598/barekengxxxxxxxxxxxxx

SIMPLIFIED FORMULAS FOR SOME BESSEL FUNCTIONS

AND THEIR APPLICATIONS IN EXTENDED SURFACE
HEAT TRANSFER

Irvan, Irvan1*, Zahedi, Zahedi2*, Anjar, Agus3, Sarmin, Suparni4, Harahap, Amin5

Mathematics Education, Graduate program, University of Muhammadiyah Sumatera Utara

1

Jl. Denai No.217, Tegal Sari Mandala II, Kec. Medan Denai, Kota Medan, 20371, Indonesia
2
Universitas Sumatera Utara, Jln. Dr. T. Mansyur No. 9, Medan, Indonesia.
3
FKIP Universitas Labuhan Batu, Jln. SM. Raja No. 126, Aek Tapa, Rantau Prapat, Kab. Labuhan Batu, Sumatera
Utara, Indonesia
4
IAIN Padangsidimpuan, Jln. T. Rizal Nurdin km 4.5 Sihitang, Padangsidimpuan, Indonesia
5
FKIP Universitas Labuhan Batu, Jln. SM. Raja No. 126, Aek Tapa, Rantau Prapat, Kab. Labuhan Batu, Sumatera
Utara, Indonesia

Corresponding author e-mail: ¹* irvan@umsu.ac.id, 2*zahedi@usu.ac.id

Abstract. Bessel functions find many applications in Physics and Engineering fields. Some of these applications are
in the analysis of extended surface heat transfer where the cross-sections vary. Tables of various kinds of Bessel
functions are available in most handbooks of mathematics. However, the use of tables is not always convenient
particularly for applications where many values must be computed. In the applications of Bessel functions in
extended surface heat transfer, graphs are also available to provide quick evaluations of the values needed.
However, reading these graphs always needs interpolation; this will be cumbersome and time consuming if there
are many readings to be taken. Mathematical formulas for Bessel functions are available but they are usually
complicated. Software to calculate values of Bessel functions is also available. Excel, Maple and Mathematica can
also be used to compute the values of Bessel functions. A user can write a program for an application that involves
Bessel functions. However, the use of Bessel functions in Excel is limited while Maple and Mathematica are
commercial software which are expensive. In this paper, formulas for Bessel functions of I 0 ( x) and I 1( x) are
simplified with adequate accuracy that can be used to easily compute values needed in the extended surface heat
transfer analysis. It is found that errors for I 0 ( x) and I 1( x) are relatively small (maximum errors are 0.004%
and 0.003%, respectively) in the range of 0.05 to 3.75 while maximum error for I 2( x) is 3.678% for the same
range. However, maximum error for I 2( x) is reduced to 0.166 if the range is from 0.25 to 3.75.

Keywords: Bessel functions, approximation functions, extended surfaces, simplified formula

Article info:
Submitted: date, month, year Accepted: date, month, year

How to cite this article:

First author, second author, etc, “TITLE OF ARTICLE”, BAREKENG: J. Il. Mat. & Ter., vol. xx, iss. xx, pp. xxx-xxx, Month, Year.

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.
Copyright © 2021 First author, Second author, etc.

1. INTRODUCTION

Bessel functions are special functions which have many applications in engineering and science. The
applications include problem solving in acoustics, hydrodynamics, radio physics, oscillations of plates,
49
https://ojs3.unpatti.ac.id/index.php/barekeng/ barekeng.math@yahoo.com
50 Irvan, et. al. Simplified Formulas for Some Bessel Functions and Their Applications in Extended Surface Heat Transfer

stress concentration on cracks and numerous others [8]. In heat transfer, applications of Bessel functions
include conduction problems and extended surface heat transfer with variable cross-sections; see [9] and
[16] for complete examples. A rather short but useful exploration of the use of Bessel functions in the fin
analysis is given by [3]. Petrova mentions various applications of Bessel functions in chemical engineering
[14] while Ribeiro Jr. and Andrade use Bessel functions in their simulation of the drying of air heating
system in a powdered milk plant [15]. Mokheimer analyzes heat transfer from extended surfaces subject to
locally variable heat transfer coefficient and concludes that the assumption of constant heat transfer
coefficient along the fin in such cases leads to a significant underestimation of the fin efficiency [11].
Mansoor introduces the boundary integral equation method to the field of extended surface heat transfer
[10]. Various authors have analyzed heat conduction using Bessel functions; see [4], [5], [7], [13] and [17],
for example. Analysis of some fluid problems using which involves Bessel functions is given by [19].

2. RESEARCH METHODS

Bessel functions come from the solution of second order differential equation
2
d y dy
x2 + x + ( x 2−n2 ) y=0 (1)
dx
2
dx
Eq. (1) is called Bessel equation of order n . If n is not an integer, the solution is given by
y=C 1 J n ( x)+C 2 J−n (x ) (2)
and when n is an integer,
y=C 1 J n ( x)+C 2 Y n ( x ) (3)
C 1 and C 2 are constants while J n (x ) is the Bessel function of the first kind, of order n and argument x , and
Y n (x ) is the Bessel function of the second kind, of order n and argument x .
A modified Bessel equation which resembles Eq. (1) is given by
d2 y dy
+ x −( x +n ) y=0
2 2 2
x (4)
dx 2
dx
If n is not an integer, the solution is given by
y=C 1 I n ( x )+C 2 I −n ( x ) (5)
and when n is an integer,
y=C 1 I n ( x )+C 2 K n (x) (6)
Here, I n ( x) the modified Bessel function of the first kind, of order n and argument x , and K n ( x ) is the
modified Bessel function of the second kind, of order n and argument x .
A general formula of Bessel’s equation is given by
d2 y dy
+ x + ( λ x −n ) y=0
2 2 2 2
x 2 (7)
dx dx
which has the general solution
y=C 1 J n ( λ x)+C 2 Y n ( λ x ) (8)
Similarly, a general formula of modified Bessel’s equation is given by
d2 y dy
x2 + x −( λ 2 x 2 +n2 ) y=0 (9)
dx
2
dx
which has the general solution
y=C 1 I n (λ x )+ C2 K n (λ x) (10)
Eqs. (7) and (9) arise in many applications; examples will be given in the next section. Complete
treatments of Bessel functions are given by [2] and [18]. Various relationships involving Bessel functions
are given in [1] and [2]. Readable analysis of Bessel functions as applied in extended surface heat transfer
are presented in [3] and [9].
Solutions of Bessel functions are given in infinite series, which may not be convenient for applications
which need quick solutions. Tables of Bessel functions which list values of arguments are therefore
preferable. Very complete tables of Bessel functions for a large range of arguments are given by [1].
We can also use software to compute Bessel functions for arbitrary order; see [12] which has also
many mathematical functions apart from Bessel functions. Excel can also compute Bessel functions but the
orders are limited. Maple and Mathematica have excellent libraries to compute Bessel functions.
BAREKENG: J. Il. Mat. & Ter., vol. xx(xx), pp. xxx - xxx, month, year. 51

3. RESULTS AND DISCUSSION

3.1. Applications of Bessel Functions in Extended Surface Heat Transfer

Extended surfaces are commonly utilized to increase the heat removal between a structure and a
surrounding ambient fluid. Here, they are attached to the primary surface [3]. An example of extended
surfaces can be seen in a motor cycle. The hot engine is cooled by air using an array of fins.

Figure 1 An air-cooled engine

(from https://faculty.virginia.edu/ribando/modules/ExtendedSurface/).

Since there are almost unlimited configurations of extended surfaces (or often called fins), we will
only take two examples which can be analyzed using Bessel functions; other examples will be presented in
another paper. All formulas are given without derivation. Interested readers can see the derivations in [3] or
[9]. Figure 2 shows a straight fin with triangular cross section or profile. Here,
b=¿ height of the fin
δ b=¿ thickness of the fin
L=¿ width of the fin

Figure 2 A straight fin with triangular profile

We will seek temperature of the fin at a distance x from the tip, T ( x); look at the coordinate of the
system. Let T s be the surrounding temperature. We will then work with the temperature excess defined by
θ ( x )=T ( x )−T s . Without derivation, the governing differential equation for the temperature excess is
given by
d 2 θ dθ
x + −m2 bθ=0 (11)
d x2 dx
where
1 /2
m=( 2 h/ kδ ) (12)
Here, h is heat-transfer coefficient (W/m .°C) and k is conductivity of the fin (W/m.°C)
2
52 Irvan, et. al. Simplified Formulas for Some Bessel Functions and Their Applications in Extended Surface Heat Transfer

The general solution of Eq. (11) is given by

θ ( x )=C 1 I 0 ( 2 m √ bx ) +C 2 K 0 ( 2 m √ bx ) (13)
Since temperature excess at the tip (x=0) is finite, C 2 must be zero since K 0 (0) is unbounded. So, we
have
θ ( x )=C 1 I 0 ( 2 m √bx ) (14)
Initial condition: at x=b ,θ=θb . So, C 1 is found and substituting C 1 back to Eq. (14) yields
θ I ( 2 m √ bx )
θ ( x )= b 0 (15)
I 0 (2 mb)
The heat dissipated by the fin is given by

q b=
kAdT
dx | x=b
=
2hL θb I 1 ( 2 mb )
m I 0 (2 mb)
(16)

while efficiency of the fin is given by

I 1 (2 mb)
η= (17)
mbI 0 (2 mb)
As the second example, consider a conical spine as shown in Figure 3.

Figure 3 A conical spine

Without derivation, the governing differential equation for the temperature excess, θ ( x )=T ( x )−T s , is
given by
d2 θ
2 dθ 2
x +2 x −M xθ=0 (18)
dx 2
dx
where
1 /2
M =( 2 m b )
2
(19)
and m is given by Eq. (12).
θ ( x )=x
−1 /2
[ C1 I 1 ( 2 M √ x ) +C 2 K 1 ( 2 M √ x ) ] (20)
Since temperature excess at the tip (x=0) is finite, C 2 must be zero since K 1 ( 2 M √ x ) / √ x is unbounded
at x=0 . Initial condition: at x=b ,θ=θb . So, C 1 is found and substituting C 1 back to Eq. (14) yields
I 1(2 M √ x )
( ) I ( 2 M √b )
1
b
θ ( x )=θb 2
(21)
x 1
The heat flow through the base is given by
πk δ b θb M I 2 ( 2 M √b )
2
q b= (22)
4 √b I 1 ( 2 M √b )
while efficiency of the fin is given by
qb √ 2 I 2 ( 2 √2 mb )
η= =
qid (mb) I 1 ( 2 √ 2 mb) (23)

All variable writing in italic with the numeric index in Century Schoolbook, for example, X 1 , … , X n.
Y i ,i=1, … , n . Use the sign ":" to explain the meaning of the variable and "=" if the variable is worth a
BAREKENG: J. Il. Mat. & Ter., vol. xx(xx), pp. xxx - xxx, month, year. 53

number or equation. N: amount of data, X = 1. Writing decimal numbers separated by a comma "," and if
more than one use a semicolon, ";", as a separator for example (0,234; 123,135; 4562,234). All parentheses
are written with the letters Century Schoolbook, (X 1 , X 2 ), [i].

3.2 Approximations of Bessel Functions

Values of Bessel functions J n ( x ) , Y n ( x ) , I n (x ) and K n ( x ) for n={0,1 } can be computed using

approximated polynomials given by [1]. Bessel functions for other orders can be computed using
recurrence relations. We will only list approximation functions for I 0 ( x ) and I 1 ( x ) which are relevant to
extended surfaces described in the previous section.
Let u be x /3.75 and v be 1/u (or 3.75/ x ). We then have the following approximation functions for
I 0 ( x ) and I 1 ( x )
6
I 0 ( x )=∑ ai u 2i + ε
i=0
where −3.75 ≤ x ≤3.75 and │ε│ < 1.6×10−7.
8
x ½ e− x I 0 ( x )=∑ b i ui + ε
i=0
where 3.75 ≤ x < ∞ and │ε│ < 1.9×10−7.
6
x−1 I 1 ( x )=∑ c i u2 i +ε
i=0
where −3.75 ≤ x ≤3.75 and │ε│ < 8×10−9.
8
x e ½ −x
I 1 ( x )=∑ d i ui +ε
i=0
where 3.75 ≤ x < ∞ and │ε│ < 2.2×10−7.
Values for all constants (a i , bi , c i and d i ) are given in [1]. [9] also quotes [1] for those constants but
with different notations.

3.3 Simplified Formulas for Bessel Functions

Various authors have devised simpler formulas for Bessel functions with less accuracy compared to
that of Abramowitz and Stegun [1]. We have also developed simplified formulas which is less accurate but
still acceptable for engineering purposes particularly for the extended surface heat transfer problems.
Using I 0 ( x ) values from 0 to 3.75, I 0 ( x ) has been regressed into
3
a4
I 0 ( x )=∑ ai u 2i +
i=0 2.1061−u
where u=
x 2
3.75 ( )
. Constants a 1 to a 4 are given as follow:

a0 a1 a2 a3 a4
-4.41575916E+01 -2.14395508E+01 -6.70459742E+00 -4.56248175E+00 9.51063985E+01

Relative errors for that equation are as follow:

Maximum = 0.004 %, Average = 0.002%, Standard deviation = 0.001%.
At first, it was planned to regressed I 0 ( x) to a quartic. However, the errors were bigger.
Using I 1( x) values from 0 to 3.75, I 1( x) has been regressed into
4
b5
I 1 ( x ) = ∑ b i u2 i +
i=0 2.3613−u

( 3.75x ) . Constants b to b are given as follow:

2
where u= 1 5
54 Irvan, et. al. Simplified Formulas for Some Bessel Functions and Their Applications in Extended Surface Heat Transfer

b0 b1 b2 b3 b4 b5
-1.01670E+02 -4.11823E+01 -1.82262E+01 -4.49396E+00 -3.00366E+00 2.40074E+02

Relative errors for that equation are as follow:

Maximum = 0.003 %, Average = 0.000%, Standard deviation = 0.001%.
Formulas for a conical spine need the value of I 2( x). However, it is not necessary to develop a
separate function for I 2( x) since for n>1, I n ( x) can be computed from the recurrence relation (see [9] or
any book on Bessel functions)
2n
I ( x )=I n−1 ( x )−I n+1 (x )
x n
2 I1 ( x )
For n=1, I 2 ( x )=I 0 ( x )− . Knowing I 0 ( x) and I 1( x), I 2( x) can then be easily computed. Relative
x
errors for I 2( x) are found to be
Maximum = 3.678 %, Average = 0.132%, Standard deviation = 0.535%.
Admittedly, the errors are relatively big. However, big errors are found at low values of x . If we limit
the argument from 0.25 to 3.75, the relative errors are much smaller as follow
Maximum = 0.166 %, Average = 0.024%, Standard deviation = 0.040%.
Now, we will compute errors for efficiency of a triangular fin. From Eq. (17) we see that the efficiency
is proportional to I 1(2mb )/ I 0 ( 2mb). Rather than just computing one particular value of the argument
(2 mb) , we will take it from 0.05 to 3.75 with an interval of 0.05, we find that relative errors are as follow:
Maximum = 0.262 %, Average = 0.043%, Standard deviation = 0.058%.
These errors are relatively small and acceptable for most engineering applications! Errors for the heat
dissipated by the fin are the same for those of the efficiency of the fin because both of them are proportional
to I 1(2mb )/ I 0 ( 2mb).
For the spine, the efficiency is proportional to I 2(2 √ 2 mb)/ I 1 (2 √ 2 mb) . So, we will compute the
ratio from 0.25 to 3.75 because we limit the lowest argument to be 0.25. we find that relative errors are as
follow:
Maximum = 0.167 %, Average = 0.024%, Standard deviation = 0.040%.
These errors are relatively small and acceptable for most engineering applications! Errors for the heat
dissipated by the fin are the same for those of the efficiency of the fin because both of them are proportional
to I 1(2 √ 2 mb)/ I 0 (2 √ 2 mb).
Approximation functions for I 0 (x) and I 1( x) for x >3.75 will be treated in another paper, together
with approximation functions for K 0 ( x ) and K 1 (x) . The latter two Bessel functions will be needed for
radial fins.

4. CONCLUSIONS

Bessel functions find numerous applications in various engineering and physics fields. One of practical
application is in extended surface heat transfer which needs evaluation of modified Bessel functions.
Simplified functions have been developed to compute those Bessel functions which are less accurate than
given in [1] but still acceptable for most engineering applications. For I 0 ( x) in the range of 0.05 to 3.75,
relative errors are as follow: Maximum = 0.004 %, Average = 0.002% and Standard deviation = 0.001%.
For I 1( x) in the same range, relative errors are as follow: Maximum = 0.003 %, Average = 0.000% and
Standard deviation = 0.001%. Values for I 2( x) are computed by using recurrence relations between I 0 ( x)
and I 1( x); the errors incurred are much bigger. However, when the range for I 2( x) is from 0.25 to 3.75,
the errors become smaller (Maximum = 0.166 %, Average = 0.024% and Standard deviation = 0.040%).
When applied to the computation of fin efficiencies, their errors are still relatively small (maximum errors
are less than 0.3% for a triangular fin and less than 0.2% for a spine). It is therefore concluded that
simplified formulas developed in this paper are acceptable for most engineering applications, at least for
extended surface heat transfer discussed here.
BAREKENG: J. Il. Mat. & Ter., vol. xx(xx), pp. xxx - xxx, month, year. 55

REFERENCES

[1] Abramowitz, M. and Stegun, I.A. Handbook of Mathematical Functions with Formulas, Graphs, and
Mathematical Tables. Dover Publications, Inc. New York: 1972.
[2] Andrews, L. Special Functions of Mathematics for Engineers 2 nd Ed. Oxford University Press. New
York: 1998.
[3] Chapman, A.J. Heat Transfer 2nd Ed. The MacMillan Company. New York: 1972.
[4] Cole, K.D., Beck, J.V., Haji-Sheikh, A. and Litkouhi, B. Heat Conduction Using Green's Functions
2nd Ed. CRC Press. Boca Raton: 2011.
[5] Hahn, D.W. and Ozisik, M.N. Heat Conduction 3rd Ed. John Wiley & Sons, Inc. Hoboken: 2012.
[6] Jeffrey, A. and Dai, H. Handbook of mathematical formulas and integrals 4 th Ed. Elsevier.
Burlington: 2008.
[7] Kakac, S., Yener, Y. and Naveira-Cotta, C.P. Heat Conduction 5 th Ed. CRC Press. Boca Raton: 2018.
[8] Korenev, B.G. Bessel functions and their applications. Taylor and Francis: London, 2002.
[9] Kraus, A.D., Aziz, A. and Welty, J. Extended surface heat transfer. John Wiley & Sons, Inc. New
York: 2001.
[10] Mansoor, M. Heat Flow Through Extended Surface Heat Exchangers. Springer-Verlag. Berlin-1983.
[11] Mokheimer, E.M.A. Heat transfer from extended surfaces subject to variable heat transfer
coefficient. Heat and Mass Transfer 39 pp. 131-138, 2003.
[12] Oldham, K., Myland, J.C. and Spanier, J. An Atlas of Functions with Equator, the Atlas Function
Calculator 2nd Ed. Springer: New York, 2009.
[13] Ozisik, M.N. Boundary Value Problems of Heat Conduction. Dover Publications, Inc. New York:
1968.
[14] Petrova, T.S. Application of Bessel’s functions in the modelling of chemical engineering processes.
Bulgarian Chemical Communication Vol. 41(4) pp. 343-354, 2009.
[15] Ribiero Jr., C.P. and Andrade, M.H.C. Analysis and Simulation of the Drying Air Heating System of
a Brazilian Powdered Milk Plant. Brazilian Journal of Chemical Engineering Vol. 24(02) pp. 35-355,
2004.
[16] Taler, J. and Duda, P. Solving Direct and Inverse Heat Conduction Problems. Springer-Verlag.
Berlin-2006.
[17] Wang, L., Zhou, X. and Wei, X. Heat Conduction: Mathematical Models and Analytical Solutions.
Springer-Verlag. Berlin-2018.
[18] Watson, A. Treatise on The Theory of Bessel Functions. Cambridge University Press. Cambride:
1922.
[19] Weigand, B. Analytical Methods for Heat Transfer and Fluid Flow Problems 2 nd Ed. Springer-Verlag.
Berlin-2015.