Numerical Simulation of the Diffusion Phenomenon in

Blood Flow

Thesis Submitted in Agreement with the Requirements of the University of

Peshawar for the Degree of BS in Mathematics

By
Muhammad Rehan and Eman Sohail

Supervised by
Dr. Tahir Saeed Khan

DEPARTMENT OF MATHEMATICS
UNIVERSITY OF PESHAWAR
June, 2024
 Author’s Declaration
I hereby certify that this thesis represents my ideas in my own words, I have
made this thesis independently and without the help of others. And where
ideas or words of others have been taken, I have adequately referenced and
cited all materials and results that are not inventive to this work. I also
declare that, this thesis has not been submitted in the same form to any
inspection authority and has not been published.

Muhammad Rehan and Eman Sohail

Department of Mathematics,
University of Peshawar,
Pakistan.

i
 Certificate
This thesis entitled Numerical Simulation of the Diffusion Phenomenon
in Human Blood Flow by Mr. Ali Khan is approved for the degree of BS
in Mathematics, under the supervision of Dr. Tahir Saeed Khan at Depart-
ment of Mathematics, University of Peshawar.

External Examiner :

Supervisor :
Dr. Tahir Saeed Khan

Chairman :
Dr. Imran Aziz

ii
Dedicated

To my
Parents

iii
Table of Contents

Table of contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii

Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii

1 Introduction 2
1.1 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2 Historical Background . . . . . . . . . . . . . . . . . . . . . . 2
1.3 Aim and Objectives . . . . . . . . . . . . . . . . . . . . . . . . 3
1.4 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.5 Work Plan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.6 Thesis Structure . . . . . . . . . . . . . . . . . . . . . . . . . . 5

2 Mathematical Formulation of the Problem 6

2.1 Assumptions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.2 Developing the Model in terms of position . . . . . . . . . . . 8
2.3 Developing the Model in terms of time . . . . . . . . . . . . . 9
2.4 Solutions of the Models . . . . . . . . . . . . . . . . . . . . . . 9
2.4.1 The model for the position effect is given as . . . . . . 10
2.4.2 The model for the position effect is given as . . . . . . 11

3 Result Discussions and Conclusion 12

3.1 Results Discussions . . . . . . . . . . . . . . . . . . . . . . . . 12
3.2 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

iv
References 18

v
List of Figures

2.1 Sketch of the cross section of a blood vessel . . . . . . . . . . . 7

3.1 Graph of Concentration of solute Cx against distance, x . . . . 13

3.2 Graph of Concentration of solute Ct against distance, t . . . . 14
3.3 Graph of Concentration of solute Cx against distance x with-
out D, V, R. . . . . . . . . . . . . . . . . . . . . . . . . . . 14
3.4 Graph of Concentration of solute in blood against time t with-
out D, V, R . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

vi
Abstract
Models describing the variation of concentration of solute in the bloodstream
over distance, x and time, t as blood solution moves at constant velocity
through the blood vessel were formulated from first principle. The models
are solved both analytically and numerically. The models were used to sim-
ulate the diffusion process in human bloodstream, determine the parameters
that affect the flux density and how the system responds to unit change
in these parameters. It was discovered that the concentration gradient be-
tween the bloodstream and its surrounding fluid decreases exponentially as
the blood solution flows through the vessel at a constant velocity, v. Hence
concentration of solute in the bloodstream approaches that of the solute in
its surrounding fluid as the distance, x traveled becomes very large. The con-
centration of the solute in the bloodstream also decreases with an increase
in time, t and approaches that in the surrounding fluid as time spent gets
large. The flux density (the time rate at which solute diffuses per unit area)
decreases with distance, x since the concentration gradient decreases with
distance, x and approach zero (no diffusion) as the concentration gradient
approach zero. The flux density becomes zero (an equilibrium state) when
the concentration of solute in bloodstream is equal to the concentration in
the surrounding fluid. It was also noticed that the flux density decreases
with an increase in the velocity of the solution in the blood vessel. Hence the
flux density depends mainly on the solute concentration gradient between the
bloodstream and its surrounding fluid, the distance, x traveled by the blood
solution in the vessel, time, t spent and the velocity of the blood solution.
The usefulness of this work have been identified to include but not limited to
nutrient uptake from the blood, infections by pathogenic secretions, dialysis,
drug action, gaseous exchange etc.

vii
Acknowledgements
All praises to Allah and His blessing for the completion of this thesis. I thank
to Allah for providing me the opportunities and strength to complete writing
the thesis.
Undertaking this M.Phil has been a truly life-changing experience for me
and it would not have been possible to do without the support and guidance
that I received from many people.
I would like to first say a very big thank you to my supervisor Dr. Tahir
Saeed Khan, for all the support and encouragement he gave me, during both
the long months I spent undertaking my research work in Peshawar and also
the time I spent at The University of Peshawar. Without his guidance and
constant feedback this M.Phil. would not have been achievable.
Many thanks goes to Dr. Imran Aziz, Chairman, Department of Mathe-
matics, for giving us with an environment to carry out our research without
any hurdles.
I am also very much thankful to all of my worthy teachers in the depart-
ment, especially to Dr. Imran Khan, Dr. Tahir Hussain, Dr. Muhammad
Adil, Dr. Muhammad Farooq, Dr. Rohul Amin, and Dr. Akbar Zada.
Special thanks to all of my friends especially Dr. Saima Rani, Dr. M.
Ijaz khan, Aqleem Hayat khan, Shahida Parveen, Maria Rehman, Bilal Ali,
Muhammad Shahkar Khan, Javaria Fayaz, Kulsum. I owe thanks to a very
special person, my husband, Noor Alam, for his continued and unfailing love,
support and understanding during my pursuit of M.phil degree that made the
completion of thesis possible. There is so much to thanks to my best friend
and cousin Dr. Shamim Shahi for her invaluable guidance, support, encour-
agement which helped me immensely during my research work. I would like
to show my deepest gratitude to my brothers-in-law, Mr. Zar Alam, Mr. Sher
Alam, Mr. Mirza Wali, Mr. Sher Aziz for their kind support and prayers
and thanks to all my nieces,nephews and cousins, Riaz Ahmad Khan, Rubina

viii
Aslam, Mirza Aslam Baig, Mirza Hussain Baig, Zahid Ali Baig, Azhar Ud
Din, Masroor Alam. Finally, I would also like to say a heartfelt thank you
to my dearest father Mr. Sharaf Ud Din, dearest mother Mrs. Barish Sharaf
Ud Din and siblings, Hasina Parveen, Azhar Ud Din, Rukhshanda Parveen,
Sabir Ud Din and Sahil Ud Din for always believing in me and encouraging
me to follow my dreams.

Muhammad Rehan and Eman Sohail

June, 2024

1
Chapter 1

Introduction

1.1 Motivation
The dependence of life processes on diffusion mechanisms could not be more
prevalent. Diffusion occurs throughout the human body, and without it, cells
and body tissue could not get important nutrients for survival, the eyes would
dry out, and many medicines could not be absorbed into the body. The lack
of nutrients inside the cell, and between the cell and the blood vessel, creates
a concentration gradient between the blood vessel and the cell. Due to the
lower concentration in the cell, the nutrient diffuses through the blood vessel
wall into the cell.

1.2 Historical Background

The model of oxygen flow in the microcirculation by Pierce [1] and Waniewski
et al [2] highlighted the inefficiency of diffusion as a means of oxygen trans-
port in the blood and hence the use of microcirculation by the body in the
form of oxygen bounded to the hemoglobin. Ibrahim, et al [3] was able to
present his model as a simple first order differential equation which shows
that the amount of drug in the blood stream approaches a constant value
under steady state conditions at a particular infusion rate. Burns and Don-

2
ald [4] developed a deterministic model for predicting drug molecule diffusion
across the Blood-Brain Barrier (BBB). Kool [5], developed a model for an-
tibiotic Distribution and Eradication of Bacteria causing endocartis based on
predicting the duration of treatment of a patient suffering from endocartis fo-
cusing on the eradication or killing of the bacteria population. For medicines
taken orally as pills, the medicine must somehow find its way into the blood-
stream. If the pills capsule is a time release mechanism, the medicine must
first diffuse out of the capsule into the stomach and then into the blood-
stream, all by diffusion process Chaplya and Chernukha [6], Pierce [1] and
Crank [7]. Another area of pharmacokinetic where diffusion plays an impor-
tant role is in the release of drugs from a swellable collagen matrix. Works
in this area have been investigated by Mario [8], and Bause, et al [9] among
others.
From the foregoing, it is obvious that diffusion is the basic way in which
all nutrients, gases, wastes and neurotransmitters move in the body (unless
active transport is required) and involves the blood stream in one way or the
other [10]. Rao, P.T. and co, developed models for blood glucose level in a
type-2 Diabetes Mellitus patient, and also determined the optimal control
policies of glucose regulatory system based on certain assumptions, Rao,
et al [11]. The importance of the knowledge of the operation of biological
processes is made available by first building a model to represent such process
from which the effects of the various factors that affects the processes are
then determined, Vakalis [12], VanLbensels et al [13], Bender [14], Ottesen
et al [15] and Rice and Do [16].

1.3 Aim and Objectives

Following objectives are set for the present research work:

• To model a diffusion process in the human blood stream that will ac-
count for such phenomena as drug absorption, exchange of gases, nu-

3
 trient uptake, etc.

• To simulate the model to provide a better understanding of the influ-

ence of the factors affecting diffusion.

1.4 Methodology
A mathematical model for the two-dimensional incompressible fluid flow in
a channel having irregular stenosis will be formulated from the Boltzmann
equation. As the problem and its solutions are completely based on the
channel dimensions and on the size and nature of stenoses in it, its geo-
metrical setup will be made in accordance with the literature. After that,
suitable microscopic and macroscopic initial and boundary conditions will be
set. To resolve the stability issues in the numerical scheme, non-equilibrium
entropy limiters will be constructed. Now once the numerical scheme for the
problem is completely constructed, its computer code will be constructed in
FORTRAN. The data obtained from running this code will be analyzed in
MATLAB in form of tables and graphs.

1.5 Work Plan

The research work is planned in the following sequence:

• In the first phase, literature review related to the proposed problem

will be carried out

• Next, the mathematical formulation and numerical scheme to solve the

problem will be formulated

• In the third phase, a computer programm for the numerical scheme will
be written

4
 • Analysis of the obtained results will be carried out in fourth stage

• In the last phase, thesis writing will be done

1.6 Thesis Structure

Introduction, background, aims and objectives and methodology of research
are presented in Chapter 1 of this thesis. In Chapter 2, historical develop-
ment of LBM, its related physical and mathematical concepts are discussed.
Also, the two dimensional LBM numerical scheme is derived in this chap-
ter. Different types of microscopic initial and boundary conditions of the
model are described here. At the end of this chapter, Ehrenfests’ limiters are
introduced to cure the numerical stability issues of LBM.
The solution of the proposed problem has been found in Chapter 3. The
computational domain of the problem and the computational setup for the
simulation is described in this chapter. The results of the simulation are pre-
sented in form of snapshots of stream function and vorticity. The conclusion
and future recommendations are presented after Chapter 3.

5
Chapter 2

Mathematical Formulation of
the Problem

2.1 Assumptions
Within a particular blood vessel, the speed of blood flow is taken as constant.
The radius of the blood vessels is also constant. There exists a mechanism
that instantaneously removes the diffused substrate from the tissue fluid into
the cells thereby maintaining Cs constant. Diffusion alone is able to account
for the differences in the concentration of the substrate in the entrance and
exit of the section under consideration. The blood vessel is assumed to be of
negligible thickness and its exterior is immersed in tissue and cell fluid. The
blood vessel will be considered as a long cylindrical pipe of length L in which
there is a lateral diffusion as the blood flows with constant velocity through
it.
For the derivation, we assume that the molecules of the solute move along
the x-axis and denote the concentration at time t by Ct . If
Cx is the concentration of solute of interest at a position in the blood vessel;
Cs is the solute concentration in the surrounding tissue fluid;
V is the speed of blood flow;

6
 Figure 2.1: Sketch of the cross section of a blood vessel

D is the diffusivity constant of the solute in the blood;

r is the radius of the blood vessel then

V = πr2 4x. (2.1)

Amountatx = Concentration × V olumeatx = Cx × πr2 4x. (2.2)

Amountatx + 4x = Concentration × V olumeatx + 4x = Cx+4x × πr2 4x.

(2.3)
Amount of materials that have diffused over time is equal to Rate of
diffusion x Change in time

Rateof Dif f usion = D × 2πr4x[Cx+θ4x − Cs ], (2.4)

where θ is such that 0 ≤ θ ≤ 1 is a typical point between x and x + 4x at

which to assume an average value of the concentration.

7
∴ Amountof materialsthathavedif f usedovertime = 2Dπr4x[Cx+θ4x −Cs ]4t.
(2.5)
Taking a material balance across the section shown:

Accumulation = Input − Output + Generation − Consumption (2.6)

This is given as

[Cx − Cx+4x ]πr2 4x = 2Dπr4x[Cx+θ4x − Cs ]4t. (2.7)

2.2 Developing the Model in terms of posi-

tion
Since
4x = V 4t (2.8)
Substitute into the left hand side of equation (7) to get

[Cx − Cx+4x ]πr2 V 4t = 2Dπr4x[Cx+θ4x − Cs ]4t. (2.9)

Dividing through by 4x4t and rearranging, we have

Cx − Cx+4x 2D
= [Cx+θ4x − Cs ].
4x rV

Cx+4x − Cx 2D
=− [Cx+θ4x − Cs ].
4x rV
Taking limits as 4x → 0

Cx+4x − Cx 2D
lim =− lim [Cx+θ4x − Cs ].
4x→0 4x rV 4x→0

8
 dCx 2D
=− [Cx − Cs ]. (2.10)
dx rV
Equation (9) gives the required model in terms of the position under
consideration.

2.3 Developing the Model in terms of time

To develop the model with respect to time, we proceed as follows:
Taking a mass balance across the section shown:

Accumulation = Input − Output + Generation − Consumption (2.11)

This is given as

[Ct − Ct+4t ]πr2 4x = 2Dπr4x[Ct+θ4t − Cs ]4t. (2.12)

the mass balance is taken with respect to time and divide equation (10)
through by 4t and taking limits as 4t → 0, we have

dCt 2D
=− [Ct − Cs ]. (2.13)
dt r
Equations (9) and (11) are the two models which describes the system
under consideration independently with respect to position and time respec-
tively.

2.4 Solutions of the Models

The models were solved both analytically and numerically and their solutions
compared.

9
2.4.1 The model for the position effect is given as

dCx 2D
=− [Cx − Cs ].
dx rV
Rearranging this equation

dCx 2D
=− dx.
[Cx − Cs ] rV
Integrating, we have
Z Cx Z x
dCx 2D
=− dx.
C0 [Cx − Cs ] rV 0

2D x
ln[Cx − Cs ]|C
C0 = −
x
x| .
rV 0

2Dx
ln[Cx − Cs ] − ln[C0 − Cs ] = − .
rV

C x − Cs 2Dx
ln[ ]=− . (2.14)
C0 − Cs rV

Cx − Cs 2Dx
= e− rV . (2.15)
C0 − Cs

2Dx
Cx = Cs + [C0 − Cs ]e− rV . (2.16)

If we know only Cs , C0 , CL ; by substituting x = L and C = CL into

equation (12), we have

CL − Cs 2DL
ln[ ]=− . (2.17)
C0 − Cs rV
By dividing equation (12) by equation (14) yields

−Cs
ln[ CCx0 −C s
] x
−Cs
= .
ln[ CCL0 −C s
] L

10
 Cx − Cs x CL − Cs
ln[ ] = ln[ ]. (2.18)
C0 − Cs L C0 − Cs

Cx − Cs CL − Cs x
ln[ ] = ln[ ]L . (2.19)
C0 − Cs C0 − Cs
And finally,

CL − Cs x
Cx = Cs + [ ]L . (2.20)
C0 − Cs
Therefore, in the absence of some of the data needed in equation (13b),
our model can still be discussed to know how the concentration of our solute
of interest varies along the blood vessel.

2.4.2 The model for the position effect is given as

dCt 2D
=− [Ct − Cs ].
dt r
where Ct is the concentration of our desired solute at time in a particular
position in the blood vessel. Other parameters remain as defined previously.
Similarly, rearranging equation (10) and solving, we have

2Dt
Ct = Cs + [C0 − Cs ]e− r . (2.21)

The same transformation were performed when we know only Cs , C0 , Cτ

and τ to finally get

Cτ − Cs t
Ct = Cs + [ ]τ . (2.22)
C0 − Cs

11
Chapter 3

Result Discussions and

Conclusion

3.1 Results Discussions

The initial data used in the numerical solution and the simulation of the
system is as follows: Cs = 0.005; C0 = 0.55; CL = 0.005; L = 60; D = 2.35 ×
10−5 ; V = 1.0; r = 5 × 10−5; xO = 0.00;
A MATLAB program to draw the graph of Solute Concentration Cx
against distance x is given by Cs = 0.005; C0 = 0.55; CL = 0.005;
L = 60; D = 0.0000235; V = 1.0;
r = 0.00005; x0 = 0.00;
x = linspace(0, 100, 101);
Cx = Cs + (C0 − Cs ). ∗ exp(−2 ∗ D ∗ x/r/V );
plot(x, Cx )
xlabel(”Distance x travelled by bloodstream, cm”)
ylabel(”Concentration Cx of solute in bloodstream, mol/cm3 ”)

12
 A MATLAB program to draw the graph of Solute Concentration Ct
against time t is given by Cs = 0.005; C0 = 0.55; CL = 0.005;
L = 60; D = 0.0000235; V = 1.0;
r = 0.00005; x0 = 0.00;
t = linspace(0, 10, 100);
Ct = Cs + (C0 − Cs ). ∗ exp(−2 ∗ D ∗ /r);
plot(t, Ct )
xlabel(”T ime t spent by bloodstream, s”)
ylabel(”Concentration Ct of solute in bloodstream, mol/cm3 ”)

Figure 3.1: Graph of Concentration of solute Cx against distance, x

Equations (13b) and (18) shows that the concentration profile of a diffus-
ing substance in the bloodstream is a function of the diameter (or radius) of
the blood vessel, speed of blood flow, diffusivity of the substance in the blood,

13
 Figure 3.2: Graph of Concentration of solute Ct against distance, t

Figure 3.3: Graph of Concentration of solute Cx against distance x without

D, V, R.

14
Figure 3.4: Graph of Concentration of solute in blood against time t without

D, V, R

concentration gradient between the solute of the substance in the blood and
that in the surrounding tissue, the time spent and position of the blood fluid
in the body. We can therefore determine the exact concentration of the so-
lute at any position in the blood vessel and hence monitor its movement in
the body.
It was noticed from the curves generated that the concentration Cx , de-
creases exponentially with an increase in distance travelled by the blood-
stream as well as time spent in the blood vessel, Figure (3.1) and (3.2). As
the distance travelled and time spent tends to infinity, the concentration
gradient decreases and the concentration of the solute in the bloodstream
approach the concentration of the solute in the surrounding tissue Cs . At
this point equilibrium is said to have been reached and the net diffusion is
zero. In the absence of diffusivity constant D, blood speed V and radius R
of blood vessel values, we were able to study the solute concentration profile
with respect to distance and time, knowing the initial concentration of solute
in blood, concentration of solute in the tissue, length of the blood vessel and

15
the concentration at the end of the blood vessel using equations (17) and
(19) respectively.
It was also seen that in the absence of diffusivity constant D, blood
speed V and the radius R of blood vessel, the concentration Cx , decreases
exponentially with an increase in distance travelled by the bloodstream as
well as time spent in the blood vessel, Figures (3.3) and (3.4).
An increase in diffusivity constant D enhances diffusion and reduces the
concentration of the solute in the bloodstream much faster hence resulted to
the concentration of the solute in the bloodstream reaching the steady state
value at both lesser distance and time Figure (??).
Also an increase in radius of the blood vessel R, the velocity of the blood-
stream V , and the concentration of the solute in the surrounding tissue Cs ,
reduces the concentration gradient and the rate of diffusion and as such the
concentration of the solute in the bloodstream Cx , may not even get to the
concentration of solute in the surrounding tissue, Cs at the end of blood
vessel L Figures (??), (??) and (??). It was again noticed that the concen-
tration gradient decays exponentially with distance travelled as well as the
time spent by the blood fluids Figures (??) and (??).
It was also noticed that the flux density decreases with distance travelled
and the time spent by the blood fluid Figures (??) and (??). It was observed
from the equations and the curves that large blood vessel beyond (510 m in
diameter) makes the exponential function to approach zero implying that rel-
atively no diffusion has taken place at such diameter or radius. The computer
codes, MATLAB programs and numerical techniques of solving the models
using MATLAB Eulers method used in solving the models and generating
the graphs for the models were presented in the appendices.

16
3.2 Conclusion
The solution to the models reveals that the concentration gradient decays
exponentially with distance and time until an equilibrium is reached. The
flux density decreases with an increase in distance travelled by bloodstream
x, time spent by bloodstream t, and reduction in concentration gradient
(CX-CS), and approach zero as concentration gradient approach zero, that
is concentration of solute in bloodstream becomes equal to that in the sur-
rounding tissue. Hence the flux density depends majorly on the concentration
gradient between the solute in the bloodstream and that in the surrounding
fluid, the distance travelled by the blood solution in the blood vessel, time
spent and the velocity of the blood solution. The usefulness of this work
have been identified to include but not limited to nutrient uptake from the
blood, infections by pathogenic secretions, dialysis, drug action, gaseous ex-
change etc. These models however were limited due to the assumption that
the solute concentration in the tissue is constant by assuming a mechanism
that quickly removes the diffused solute from the tissue fluid, which does not
always apply.
Finally, within the limits of acceptability, the models can account for
many diffusion processes in the human blood stream.

17
References

[1] C. Pierce, Mathematical models of oxygen flow in the microcirculation,

Project: Department of Physiology, University of Arizona, USA (2006)
1–6.

[2] J. Waniewski, Effects of Blood Perfusion on Diffusive Transport in Per-

itorial Dialysis, Kidney International article, 1999.

[3] M. O. Ibrahim, Mathematical model of pharmacokinetics, Department

of Mathematics, Usmanu Danfodiyo University, Sokoto State (2006) 1–4.

[4] J. Burns, F. W. Donald, a mathematical model for the prediction of drug

molecule diffusion across the blood-brain barrier, neurological sciences
31 (2004) 520–527.

[5] M. Kool, Mathematical modeling for antibiotic distribution and erad-

ication of bacteria causing endocarditis, PhD Thesis, Department of
Applied Mathematics, University of Twente, Netherland (2006) 25–28.

[6] Y. Chaplya, O. Chernukha, Mathematical modeling of diffusion of de-

caying particles in regular structures, Revised Advanced Material Sci-
ence 23 (2010) 21–31.

[7] J. Crank, The mathematics of diffusion, Clarendon, Oxford, UK.

18
 [8] G. Mario, Mathematical modeling and controlled drug delivery: Matrix
systems, Current Drug Delivery. Bentham Science Publishers Ltd (2005)
97–116.

[9] M. Bause, Model describing drug release from collagen matrices, Pro-
ceedings of the Fifth Workshop on Mathematical Modeling of Environ-
mental and Life Sciences Problems, Constanta, Romania (2006) 27–34.

[10] M. J. Bech, Stochastic models and methods to characterize the glu-

cose/insulin system, M. Sc Thesis: Informatics and Mathematical Mod-
eling, Technical University of Denmark, Kongens, Lyngby (2008) 1–16.

[11] T. P. Rao, Stochastic Modeling of Blood Glucose level in type-2 Diabetes

mellitus, Asian Journal of Mathematics and Statistics (4).

[12] I. Vakalis, Diffusion in biology: A mathematical modelling approach,

, Ph. D Thesis, Department of Mathematics and Computer Science,
Capital University (2002) 2–5.

[13] E. M. VanLbensels, Modeling of diffusion limitation of gas exchange in

lungs containing per fluorocarbon, Applied Physiology 86 (1999) 273–
284.

[14] E. A. Bender, Introduction to Mathematical Modeling, 1978.

[15] J. T. Ottesen, Applied mathematical models in human physiology,

BioMath-Group, Department of Mathematics and Physics, Roskilde
University, Denmark (2006) 09–19.

[16] R. G. Rice, D. D. Do, Applied mathematics and modeling for chemical

engineers, John Wiley and Sons, New York, USA (1995) 39–50.

19