kedir RIVIES on Manuscript Template

International Journal of Advanced Engineering and Nano Technology (IJAENT)
ISSN: 2347-6389, Volume-4, Issue-6, April 2021
Paper ID: D0460024521

Peaceman-Rachford Alternating Direct Implicitly Method for Linear Advection Diffusion
equation and Its Application
Kedir Aliyi Koroche

Department of Mathematics, College of Natural and Computational Sciences, Ambo University,
Ambo,paper_id//2021©BEIESP
Retrieval Number: Ethiopia Published By:
Blue Eyes Intelligence
DOI: 10.35940/ijaent.xxxxx.xxxxxx
Engineering and Sciences
1 Publication
Abstract: In this paper, Peaceman-Rachford significant applications of the
alternating direct implicitly methods for linear advection-diffusion equation lie in fluid
advection-diffusion equation. First, the domain was dynamics, heat transfer, and mass
discretized using the uniform mesh of step length transfer [1].
and time step. Secondly, by applying the Taylor
series methods, we obtain the central difference The diffusion-convection-reaction equation arises
equation. Then rearranging this central difference in several physical phenomena, such as the
equation; we write the two half scheme of the dispersion of chemicals in reactors, the shocks in
present method. From each half of these schemes, hydrodynamical flows, the mass transfer in the
we obtain tri-diagonal coefficient matrices for capillary membrane [2] bioreactor, and the tracer
system of difference equation. Lastly by applying the dispersion in the porous medium. Mathematically,
Thomas algorithm and writing MATLAB code for
the class of equations can be expressed as a set of
the scheme we obtain solution of the problem. To
partial differential equations (PDEs) with the
validate the applicability of the proposed method,
three model examples are considered and solved for
initial or boundary conditions. Most physical
different values of mesh sizes in both directions. The quantities are involved with the interaction of the
convergence has been shown in the sense of temporal and spatial domain [2]. This equation is
used in the momentum conservation part of the
maximum absolute error -norm and -norm,
Navier-Stokes equations [2], [3].
numerical error and experimental order of The nonlinear advection-diffusion type equation is
convergence. The stability and convergence of the one of the popular and important models
present numerical method are also guaranteed and describing many phenomena derived from various
the comparability of the stability of the present areas of mathematical physics and engineering
method is presented by using the graphical and fields such as gas dynamics, hydrodynamics,
tabular form. The numerical results presented in shock waves, heat conduction, and so on [4].
tables and graphs confirm that the approximate Also, the type of equation represents the Burgers
solution is in good agreement with the exact equation, the heat conduction equation, the
solution. nonlinear Schrödinger equation, the Navier–
Keywords: Linear advection-diffusion equation, Stokes equation [4].
Peaceman-Rachford alternating direct Hence, the development of efficient and accurate
implicitly method, Taylor series algorithms for solving the equations is of great
methods, tri-diagonal coefficient importance in the computational fluid dynamics
matrices, Thomas algorithm, stability community and has been widely studied by many
and convergence. researchers.
Advection-diffusion problems arise when the time
Manuscript received on March 9, 2021. evolution of chemical or biological species in a flowing
Revised Manuscript received on April 5, 2021. medium is described [15].
Manuscript published on April 30, 2021. The blending of particulate materials is a common
* Correspondence Author manufacturing unit operation in many industries,
Kedir Aliyi Koroche such as those that produce chemicals,
Department of Mathematics, College of Natural and pharmaceuticals, food products, and
Computational Sciences, Ambo University, Ambo, agrochemicals [5]. For instance, multi-scale
Ethiopia
modeling approach for predicting the blending of
Email address: kediraliyi39@gmail.com (Kedir A.)
particulate material in a rotating cylindrical drum
1. Introduction
is an advection-diffusion equation [5]. Solute
Advection–diffusion equation is a parabolic partial
dispersion under the combined effects of diffusion
differential equation which is derived on the
and advection in the heterogeneous porous
principle of conservation of mass using Fick’s medium is modeled by parabolic advection-
law. Particles, energy, or physical quantities are diffusion equations (ADE). Many
transferred inside a physical system due to two harmful effects on humans and
processes: advection and diffusion [11]. The the environment such as
advection-diffusion equation describes these two atmospheric pollutions,
processes for several substances [11].The contaminated flows in
Published By:
Retrieval Number: paper_id//2021©BEIESP Blue Eyes Intelligence Engineering
DOI: 10.35940/ijaent.xxxxx.xxxxxx and Sciences Publication
2
groundwater aquifers, chemicals and migration of through an open medium like air, rivers, lakes,
contaminations in the seawaters and river systems, and porous medium. River water pollution can be
tracer dispersion in a porous medium are modeled established by the one-dimensional advection-
by such types of equations [6], [11]. It also diffusion equation [11]. It has wide applications in
considers physical phenomena wherein the other disciplines too, like soil physics, petroleum
diffusion process particles are moving with a engineering, chemical engineering, and
certain velocity from higher concentration to biosciences [11]. The practitioners of
lower concentration [11]. The analytical and computational fluid dynamics have
numerical solutions along with an initial condition decomposed the analysis of the
and two boundary conditions aid to understand the complexity and stiffness of the Navier–
pollutant
Retrieval concentration distribution behavior
Number: paper_id//2021©BEIESP Stokes equations
Published By: into simpler problems like the
3 Publication
Stokes (linear) equations that embody the accurate and stable Peaceman-Rashford
difficulties of the space discretization of the alternating direct implicitly scheme that is capable
velocity and pressure fields and the advection- of solving linear advection-diffusion equation and
diffusion problem that is related to the transport approximate the exact solution. The convergence
character of the non-linear terms [7]. This last has been shown in the sense of maximum absolute
class of problems includes the non-linear Burgers error -norm and -norm, numerical error
equations and the linear advection-diffusion
equation. and so that the local behavior of the
This paper address the one-dimensional linear solution is captured exactly. The stability and
advection-diffusion equation with homogeneous convergences of the present methods are also
Dirichlet boundary conditions as this is a investigated.
meaningful test for established or novel discrete The present paper is organized as follows. Section
schemes. For high Reynolds number flows the two is Statement of the problem, section three is
advection is dominating diffusion but the presence Formulation of the numerical scheme, section
of the boundaries imposing no-slip wall four is Stability and convergence analysis for the
conditions complicates the solution of the approximate solution, section five is Numerical
problem. Boundary layers develop and in most results, and section six is Discussion and
cases influence deeply the flow dynamics. No-slip conclusion.
wall boundary conditions impede the general use 2. Statement of the problem
of periodic Fourier representation and spectral Consider the following linear Advection diffusion
calculation [7]. Even though the linear advection- Equation:
diffusion equation is linear it is difficult to find a ,
closed-form analytical solution in the literature.
Most of the efforts have been devoted to the
solution of linear advection-diffusion equation (1)
with an upstream boundary condition and a Robin With initial and boundary condition respectively
or Neumann downstream condition. The presence
of the gradient condition at the exit of the domain
eases the development of the analytical solution. , (2)
The paper by Pérez Guérrero et al. [10] uses a Where a, d are none zero constant and ,
change of variable to obtain a heat equation which
is then solved by a generalized integral transform are smooth function on .
technique proposed by Cotta [8]. In [9], van Now we define a mesh size h and k and the
Genuchten et al. can use a variable transformation constant grid point by drawing equidistance
that reduces the partial differential equation to an horizontal and vertical line of distance ‘h’ and ‘k’
ordinary differential equation the solution of respectively in ‘x’ and ‘t’ direction. These lines
which is expressed by the complementary error are called gridline and the point at which the
function [7]. interacting is known as the mesh point. The mesh
Moreover, in the literature review, there are point that lies at end of the domain is called the
different numerical methods applied by the author boundary point. The points that lie inside the
to solve linear advection-diffusion equations [16]. region are called interiors points. The goal is to
However, still, the accuracy and stability of the approximate the solution ‘u’ at the interior mesh
method need attention because the treatment of points. Hence we discretized the solution domain
the method used to solve the linear advection- as:
diffusion equation is not trivial distribution. Even ,
though the accuracy and stability of the
aforementioned methods need attention, they (3)
require large memory and long computational Where and
time. So the treatments of this method present
severe difficulties that have to be addressed to , ,
ensure the accuracy and stability of the solution. . M and N are the
To this end, this paper aims to develop the
Published By:
4
maximum numbers of grid points respectively in order derivatives. By using Taylor series
the x and t direction. expansion, we have
3. Formulation of the numerical scheme
Central difference method
Assuming that has continuous higher order
partial derivative on the region [a, b]x[0,T]. For
the sake of simplicity, we use ,
and for is
Retrieval Number: paper_id//2021©BEIESP Published By:
5 Publication
different method. The difference equation for
PDEs at interior mesh point lead linear system of
the equation whose coefficient matrices has at
most five none zero entries in each row [13]. For
many years splitting methods have proved
valuable in the numerical solution of time-
(4) dependent, multi (space)-dimensional partial
Adding the first equation to the second equation, differential equations (PDE's). The general idea of
subtract the second equation from the first splitting is to attack a multi-dimensional problem
equation, and again subtract the fourth equation in such a way that only one-dimensional
from the third equation of Eq. (4), respectively we computations are required. This idea has led to the
obtain: development of a great variety of so-called
alternating direction implicit (ADI) methods,
, locally one-dimensional (LOD) or fractional step
methods, and hopscotch type methods [13].
, Therefore in this paper is interest to use this idea
to solve the advection-diffusion equation.
(5) Now we choose starting points to apply this
method. Let is starting point lies at each
Where , and interior mesh point of the computational domain
and our interior-point contains two half.
are respectively their local Hence form Eq.(6) equation of these two half of
truncation error terms. Now substituting Eq.(5) Peaceman-Rachford ADI methods are:
into Eq.(1) we obtain the differential equation of First-half:
the form:
Second-half:
Multiplying both side of this equation by and

simplifying it we obtain
(7)
Where
the
(6)
algorithm in Eq.(7) contains the complete time
Where , and step of the first half from =
, and second half from to

are constant. into two sub-step respectively. For n=p=0, the
Peaceman-Rachford initial condition or starting points are preformed
Alternative Direct Implicitly Methods for each interior points . Hence by choosing
Alternative direct implicit methods were first
introduced by Peaceman, Douglas, and Rachford the starting point for each interior point
for obtaining the solution of parabolic and elliptic we update the scheme for
Partial differential equations in two of three both halves. For each interior
variables [12]. They develop alternative direct point, our starting points are
implicitly (ADI) methods to overcome the values of the initial condition at
difficulty to find the solution of PDEs by a each interior point. Note that in the
Published By:
6
single iteration process we have two half. For where j=0(1)M and p=n=0. In this equation, the
instance, in the first iteration process, we have same value of “u” is known value and they are
two. So for the first half of the first iteration boundary value that lies at j=0 and j=M. They are
process, we update the starting point ‘ ’ row not affected by the starting value. The remaining
values of ‘u’ at the interior point in the scheme
by row of first of Eq.(7). Hence for the row are remaining unchanged until the process
we have: is finished. Therefore for a fixed value of
‘n’ in Eq. (8) is applied to all interior
mesh points along with the row n. Thus
the obtained system of the equation
Retrieval Number: paper_id//2021©BEIESP (8) contains
Published the
By: M equation. So these equations
7 Publication
contain rows and the system contains NM Fourth iteration (p=3)
First-half:
total equation. The coefficient matrix is tri-
diagonal. The solution of this iteration expressed
in terms of . For the second half of the first

, j=1(1)M, nth row.
Second-half:
iteration, we update the value that obtained
from the first half of the first iteration. Hence the
second half of the first iteration generated by the
formula
, n=1(1)N, jth column.
This process is continuous until the tolerance of
the solution is obtained. For all-iteration, we
obtain the tri diagonal matrices and we use the
Thomas method to find the solution of each
(9) system of iteration. One of the termination
Where p=0, n=0(1)N. For fixed ‘j’, we apply Eq. conditions is assembling the value of ‘u’ at e each
(9) for all interior-point along the jth column and interior point’s into a matrix with the same
produced the N equation. Once again our starting configuration of the grid point. If the norm of the
difference between two successive such matrices
value is the value of that obtained first half is less than a prescribed tolerance, the iteration is
and the remaining value of “u” at interior point (j, terminated.
n) remain unchanged. Like the first half, the Thomas Method.
second half of the first iteration contains the NM The tri-diagonal system of equations arises after
equation. The coefficient matrix is a tri-diagonal in engineering application. It appears in the
matrix. The values of solutions for this half stored special form whose coefficients matrix contains
three none zero entries with main diagonal. Let us
in terms of .Generally, the successive iteration conceder the coefficient matrix of the first
of the PRADI-method are:
iteration given us and
Second-iteration(for-p=1): First-half:
respectively for both first and second half’s where
, j=1(1)M, nth row.

Second-half:
, n=1(1) N, jth column.

Third iteration (for p=2):
First half:
And the
, j=1(1)M, nth row.

Second-half:
corresponding column vectors are:
, n=1(1)N, jth column.
Published By:
8
iteration is the starting value for the second

iteration. The process is continuous until the
tolerance is a capture. To solve these types of
system of the equation the most recommended
numerical method is the Thomas method. This is
because the coefficient matrix contains
several zero entries. This method is to use
the Gauss elimination technique with the
(10) diagonal entries scaled to 1 in each step.
If the prescribed tolerance of iteration is not Thus we can write MATLAB code and find
Retrieval Number: paper_id//2021©BEIESP
captured at the first iteration, the solution of this thePublished
solution By:to Eq.(1).
9 Publication
4. Stability and Convergent Analysis.
Since the time step changed from to
+ and + to . Now we want

to find the global error bound condition for the
scheme. Let be the small constant who’s their
Nth sum is less than one for all point (j, n).
Hence form Eq.(7) the next succeeding says:
(15)
where and
, G is constant. Let j=1

and n=0. Then from Eq. (15), we obtain
(11)
Where be constant. Now let and implies that
two numerical solutions at two were succeeding .This shows that the global
points. Hence the error between these two
error is bounded. Generally from the above
numerical solutions is: for condition, we mean that for any ‘(j, n)’ we have a
n=1(1) N and j=0(1) M for is a round off different matrix coefficient of the error term
error. Subtracting Eq.(7) from Eq.(11), we obtain: and the error is by the norm of this matrix which
is given as:
(16)
where . This
shows that the proposed method (PR-ADI
(12) methods) in Eq.(7) is stable.
Now eliminating the error term and Definition 1: For linear difference operator
is stable if, for sufficiently small mesh
from Eq.(12) and simplifying it, we size h and k, there is a constant “G” independent
obtain: of size h and k such that
(13)
where tri-diagonal matrix and I is an (17)

identity matrix whose Eigen-value is less than where is any mesh function. By adding the
one. From Eq.(13) we obtain first and second half of Eq.(7) we obtain the
differential equation given as:
(1
4)
Now applying the properties of matrix norm and
using triangular inequality on Eq.(14) we obtain:
Published By:
10
and
E
(18)
where

11 Publication
equation in Eq.(18) is stable. This implies that the
solution to the system equation in Eq.(7) is
. uniformly bounded and independent of mesh-size
Theorem 1: The difference operator defined in h and k. Therefore the scheme is stable.
Eq.(18) is stable any for constant G. Corollary 1. Under the condition for the above
Proof: Let and are denoted the theorem, the error between the exact solution
difference operator that defined on the left and and the numerical solution given as
right-hand side of Eq.(18) respectively and is
with boundary and initial
any mesh function that satisfies
condition satisfy:
(19)
where . If the occur

where local truncation error in Eq.(6).
at or from
definition 1, the difference equation is stable. Let Proof: Let be the error and under the given
us consider exist at . condition its satisfies
from Eq.(19) we have:
where j=2(1)M-1,n=2(1)N-1 and

. By the above theorem stability of implies
that
Thus the error is bounded and the scheme is

stable. From this as the maximum
where is constant. Hence the difference
local truncation is tending to zero. It means
equation is stable. Therefore this condition is true
for . Now for any interior point
. Again the
we have:
scheme is consistent. Therefore the proposed
scheme is convergent.
4. Criteria for Investigating the Accuracy
of the Method
In this section, we investigate the accuracy of the
present method. There are two types of errors
Round-off errors and Truncation errors occur
when differential equations are solved
numerically. Rounding errors originate from the
fact that computers can only represent numbers
using a fixed and limited number of significant
figures. Thus, such numbers or
where provided that all cannot be represented exactly
in computer memory. The
matrix norms are different from zero. The discrepancy introduced by this
inequality in Eq.(17) is held. So the difference limitation is to call a Round-off
error. Truncation errors in
Published By:
12
numerical analysis arise when approximations

are used to estimate some quantity. The accuracy
of the solution will depend on how small we
make the step size, h, and time step k. To test the
performance of the proposed method, and
is used to measure the accuracy of the where is the maximum number of
method. This error norm is calculated as
step, is the exact solution and
followed in [14] is given by
,
13 Publication
approximation solution of advection-diffusion Khatun in [11] given as: ,

, and the exact solution is
equation in Eq.(1) at the grid point . and a=0.3m/s ,
5. Numerical Experiments d=.005m2/s.
To test the validity of the proposed method, we Example 3: Consider the advection-diffusion
have considered the following three model equation considered by Appadu in [1] given as:
problem. Numerical results and errors are
, , and
computed and the outcomes are represented
tabular and graphically. initial and boundary condition is given as
Example 1: Consider the advection-diffusion
equation considered by Sigrun in [15] given
as: ,
, and the exact solution is
and we obtain initial
and boundary conditions from the exact solution.
Example 2: Consider the advection-diffusion
equation considered by Abdullah and
Table 1 Comparison of -errors and experimental order of convergence (EOC) concerning the
classical diffusion fluxes for the exponentially growing solution with a=1 and d = 0.1. Computations
carried out until final time T = 10 with time step h=k.
NO of Sigrun in [15](BR2) Present Method
grid
M N EOC EOC
20 20 1.03E-01 __ 1.6288E-08 2.2297
40 40 2.36 E-02 2.13 6.2484E-08 2.6962
80 80 5.63E-03 2.07 2.2977E-09 3.1788
160 160 1.38E-03 2.02 8.2810E-10 3.6699
320 320 3.44E-04 2.01 8.2671E-10 3.9218
640 640 8.60E-05 2.00 3.1289E-11 4.1193
a) The surface of numerical solution b ) plotting graphs of numerical

solution M=9.
Fig 1. Graphs of numerical solution of example one for k=0.5.
Table 2 Expiration of -errors, -errors -errors and experimental order of
convergence (EOC) for the classical diffusion fluxes for the exponentially growing
Published By:
14
solution of Example 2 with a=0.03 and d = 0.005.

Step size Present methods
h k EOC
2.5 3 2.4273E-07 5.3400E-6 1.1385E-06 2.2297
1.25 1.5 1.2794E-06 5.3733E-05 8.2912T-06 2.6962
0.83 1 4.6734E-06 2.8975E-04 3.6799E-05 2.9771
0.625 0.75 8.6415E-06 7.0860E-04 7.8252E-05 3.1788
0.5 0.6 1.2005E-06 1.2245E-04 1.2124E-05 3.3362

15 Publication
Fig 2. Graphs of numerical solution of example two for k=3, h=2.5.

Table 3 Computations of -errors, -errors -errors and
experimental order of convergence (EOC)
for the classical diffusion fluxes for the exponentially growing solution of Example 3 with a=1 and d =
0.1.
Mesh size By present method
h k L1 L2 EOC
0.005 0.02 2.3352E-07 4.7350E-05 3.3189E_06 3.8291
0.01 0.02 4.6246E-07 4.7170E-05 4.6706E-06 3.3362
0.01 0.04 4.6245E-07 4.6175E-05 4.6705E-06 3.3362
0.02 0.04 9.0713e-07 4.7171e-05 6.5414e-06 2.8502
Appadu in [1]
0.005 0.02 1.8166E-04 5.8157E-04 1.6348E-04 0.25
0.01 0.02 7.3296E-04 0.0024 1.6348E-04 0.50
0.01 0.04 0.0021 0.0065 0.0011 0.25
0.02 0.04 1.2252E-04 3.7946E-04 3.7946E-04 0.50
a) The surface of numerical solution b ) plotting graphs of numerical versus exact solution
Fig 3. Graphs of numerical solution of example three for 02
5. Discussion and Conclusion diffusion equation. To

In this paper, Peaceman-Rachford alternating demonstrate the competence of
direct implicitly method for linear advection- the method, it is applied on
Published By:
16
three model examples by taking different values superior to the method developed in [1],[11], and
for mesh size, h, and k. Numerical results [15] and approximate the exact solution very well.
obtained by the present method has been Further, as shown in Figs. 1-3, the proposed
computed with numerical results obtained by the method approximates the exact solution very well
methods in [1], [11], and [15] and they are for which most of the current methods fail to give
summarized in Tables and graph. In this paper good results. To further verify the
both theoretical and numerical error bounds has applicability of the planned method,
been established for the Peaceman-Rachford graphs were plotted aimed at the above
alternating direct implicitly method .As can be examples for exact solutions versus the
seen from the numerical results and predicted in numerical solutions obtained with their
tablesNumber:
Retrieval and paper_id//2021©BEIESP
graphs above, the present method is errors. As By:
Published Figs. 1(b) indicates numerical solution
17 Publication
obtained by present method is good agreement of analytical solutions for contaminant transport in
the results with exact solutions, which proves the rivers 1. The equilibrium advection-dispersion
equation", Journal of Hydrology and
reliability of the method for example 1. Figs. 2 Hydromechanics, 61(2),146-160, 2013.
indicate the numerical solution has a good 10. Guerrero, JS Pérez, Luiz Cláudio Gomes
agreement of the results with the exact solution Pimentel, Todd H. Skaggs, and M. Th Van
for example 2. Figs. 3(b) indicate the numerical Genuchten. "Analytical solution of the advection –
diffusion transport equation using a change-of-variable
solution has a good agreement of the results with and integral transform technique." International
the exact solution for example 3. Further, the Journal of Heat and Mass Transfer 52(13) 3297-3304,
numerical results presented in this paper validate 2009.
11. L.S. Andallah, MR Khatun. “Numerical solution of
the improvement of the proposed method over advection-diffusion equation using finite difference
some of the existing methods described in the schemes”, Bangladesh Journal of Scientific and
literature.. Industrial Research, 55(1) 15-22, 2020.
Acknowledgements 12. W.H. Hundsdorfer, JG Verwer. “Stability and
Convergence of the Peaceman-Rachford alternative
The author wish to express his thanks to the direct implicitly method for Initial-Boundary Value
authors of literatures for the provision of initial Problems”, Journal of mathematics of computation, 53
idea for this work. (187), 81-101, 1987.
13. A.R. GOURLAY. “Splitting methods for time-
Last but not least, the author wish to express his dependent partial differential equations”, The State of
thanks to all his beloved friends; your the Art in Numerical Analysis (D. Jacobs, ed.),
kindness and helps will be a great memory for Academic Press, New York, pp.757-791, 1977.
14. Gurhan Gurarslan, Halil Karahan, Devrim Alkaya,
me. Murat Sari, Mutlu Yasar. “Numerical Solution of
Reference Advection-Diffusion Equation Using a Sixth-Order
1. Appadu, Appanah Rao. "Numerical solution of the 1D Compact Finite Difference Method”, Hindawi
advection-diffusion equation using standard and Publishing Corporation Mathematical Problems in
nonstandard finite difference schemes", Journal of Engineering.2013 (7), 2013.
Applied Mathematics, 2013 .
2. Zhong, Jiaqi, Cheng Zeng, Yupeng Yuan, Yuzhe 15. Sigrun Ortleb. “ -stability analysis of IMEX-
Zhang, Ye Zhang. "Numerical solution of the unsteady
(σ,μ)DG schemes for linear advection-diffusion
diffusion-convection-reaction equation based on
equations”, Applied Numerical Mathematics 147,43–
improved spectral Galerkin method." AIP
65, 2020.
Advances, 8(4),2018.
16. D.M. William, A. Jameson. “Energy stable
3. Evstigneev, M Nikolay. "Numerical analysis of Krylov
Flux Reconstruction Schemes for Advection–
multigrid methods for stationary advection-diffusion
Diffusion Problems on Tetrahedral", Journal
equation", In Journal of Physics: Conference Series,
of Scientific Computing, 2013.
1391(1), IOP Publishing, 2019.
4. Soyoon Bak, Philsu Kim, Xiangfan Piao, Sunyoung Bu. AUTHORS PROFILE
“Numerical solution of advection-diffusion type
equation by modified error correction scheme”, Open
Springer Journal of Advances in Difference Kedir Aliyi Koroche is Graduate
Equations,2018. B.sc in Mathematics from Ambo
5. Yu Liu, Marcial Gonzalez, and Carl Wassgren. University in 2016 G.C. and
Modeling Granular Material Blending in a Rotating
M. Sc. in
Drum Using a Finite Element Method and Advection-
Mathematics(specialized in
Diffusion Equation Multi-scale Model. American
Institute of Chemical Engineers, 64(9)3277-3292, 2018. Numerical Analysis) from Jimma
6. A.V.C.I. Derya, Aylin Yetim. "Analytical solutions to University in 2019 G.C. Now he works at Ambo University
the advection-diffusion equation with the as Lecturer and Researcher. He has two published research
Atangana-Baleanu derivative over a finite paper.
domain", Balıkesir Universities Fen Balmier
Enstitüsü Dergisi , 20(2) 382-395, 2018. .
7. Abdelkader Mojtabi, Michel Deville. “One-
dimensional linear advection-diffusion equation:
Analytical and finite element solutions”,
Computers and Fluids, Elsevier, 107,189-195,
2015.
8. R.M Cotta. Integral transforms in computational
heat and fluid flow. Boca Raton (FL): CRC Press;
1993
9. V.A.N Genuchten, Martinus TH, Feike J. Leij,
Todd H. Skaggs, Nobuo Toride, Scott A.
Bradford, Elizabeth M. Pontedeiro. "Exact
Published By:
18

kedir RIVIES on Manuscript Template

Uploaded by

Copyright:

Available Formats

kedir RIVIES on Manuscript Template

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

kedir RIVIES on Manuscript Template

Uploaded by

Copyright:

Available Formats

International Journal of Advanced Engineering and Nano Technology (IJAENT)

ISSN: 2347-6389, Volume-4, Issue-6, April 2021

Paper ID: D0460024521

Kedir Aliyi Koroche

Multiplying both side of this equation by and

, and second half from to

in terms of . For the second half of the first

, j=1(1)M, nth row.

, n=1(1) N, jth column.

, j=1(1)M, nth row.

, n=1(1)N, jth column.

iteration is the starting value for the second

+ and + to . Now we want

, G is constant. Let j=1

where tri-diagonal matrix and I is an (17)

Retrieval Number: paper_id//2021©BEIESP Published By:

where . If the occur

where j=2(1)M-1,n=2(1)N-1 and

Thus the error is bounded and the scheme is

numerical analysis arise when approximations

approximation solution of advection-diffusion Khatun in [11] given as: ,

a) The surface of numerical solution b ) plotting graphs of numerical

solution of Example 2 with a=0.03 and d = 0.005.

Retrieval Number: paper_id//2021©BEIESP Published By:

Fig 2. Graphs of numerical solution of example two for k=3, h=2.5.

5. Discussion and Conclusion diffusion equation. To

You might also like