Professional Documents
Culture Documents
ISSN 2229-5518
Abstract—There are many equations in mathematics which are used in our practical life. Burger’s equation is one of them which is a
good simplification of Navier-Stokes equation where the velocity is one spatial dimension and the external force is neglected in absence of
pressure gradient. This equation is used to analyze traffic congestion and acoustics. It occurs in various areas of applied mathematics,
such as modeling of various problems in fluid dynamics and traffic flow etc. Due to the complexity of the analytical solutio n, one needs to
use numerical methods to solve this equation. For this we investigate finite difference method for Burger’s equation and present an explicit
central difference scheme. We implement the numerical by computer programming for artificial initial and boundary data and verify the
qualitative behavior of the numerical solution of burger’s equation.
Index Terms— Burger's equation, Navier-Stokes equation, Cauchy problem, Inviscid fluid, Viscous fluid,Finite difference schemes ,
Analytical solution, Numerical solution.
—————————— ——————————
1 INTRODUCTION
IJSER
problem of the interaction between nonlinear and dissipative of the viscosity term.
phenomena. Even though it is a simplest case study’ which in In order to understand the non-linear phenomenon of the Na-
many setting is not realistic, it has been important in wide vier-Stokes equation, one needs to study Burger's equation
range of mathematical problems, from hydrodynamics to ge- analytically and numerically as well. Many works has been
ometry. It is now realized that Burger's equation was used by a appeared in the last several years e.g. [12], [13]etc.
number of scientists before its re-introduction by Burgers, for In this paper, we present the analytical solution of one-
example see H. Bateman [3] and A.R. Forsyth[4]. dimensional Burger's equation as an initial value problem in
It is now known that it was first introduced by Bateman [3] in infinite spatial domain and some numerical methods for solu-
1915 who found its steady solutions, descriptive of certain vis- tion of Burger's equation as an initial boundary value problem.
cous flows. It was later proposed by Burgers [4] as one of a
class of equations describing mathematical models of turbu- 2 BURGER’S EQUATION
lence and due to the extensive work of Burger it is now known
as Burger's equation. It involves series solution that converges Burger’s equation is a fundamental partial differential equa-
very slowly for small values of the viscosity constant [5]. Many tion from fluid mechanics. It occurs in various areas of applied
authors Cole, J.D [6], Mittal R.C and Singhal P [7], Caldwell, J., mathematics, such as modeling of gas dynamics and traffic
P. Wanless and A.E. Cook [8] have discussed the numerical flow. It is named for Johannes Martinus Burgers (1895-1981).
solution of Burger’s equation using Finite Difference Methods The Burger’s equation was known to Forsyth (1906) and had
and Finite Element Methods. been discussed by Bateman (1915). Due to extensive works of
The applications of Burger equation are demonstrated in the Burgers (1948) it is known as Burger’s equation. It is a nonlin-
modeling of water in unsaturated soil, dynamics of soil water, ear equation for which exact solutions are known and is there-
statistics of flow problems, mixing and turbulent diffusion, fore important as a benchmark problem for numerical meth-
cosmology and seismology [9, 10, 11]. ods. Burger’s equation is a good simplification of Navier-
In the content of gas dynamics, it was discussed by Hopf and Stokes equation where the velocity is in one spatial dimension
Cole. They also illustrated independently that the Burger's and the external force is neglected and without any pressure
equation can be solved exactly for an arbitrary initial condi- gradient. The Burger’s equation has been used to test and in-
tion. Benton and Platzman have surveyed the analytical solu- vestigate the numerical method for Navier-Stokes equation.
This equation is used to analyze traffic congestion and acous-
———————————————— tics.
α,β,,∆:Lecturer, IUBAT-International
University of Business Agriculture and
µ
Technology. :Professor,Jahangirnagar University
ΩAssistant Engineer, DESCO-Dhaka Electric Supply Companny Limited.
α β
shafiqul.islam@iubat.edu, saiduzzaman.ju@gmail.com,
Ω ∆ µ
sobuj_kumar_ray@yahoo.com, shifatahmed63@yahoo.com, saz-
zad67@hotmail.com
IJSER © 2014
http://www.ijser.org
International Journal of Scientific & Engineering Research, Volume 5, Issue 4, April-2014 492
ISSN 2229-5518
2.1 Derivation of Burger’s equation quantity.
The Navier-Stokes equation is given by The characteristics satisfy x΄(t) = u(x(t),t) and each characteris-
q 1 tic u is constant, since
(q.)q F p vq q
t u( x(t ), t ) ) = u( x(t ), t ) + u( x(t ), t ) x(t ) )
The 1D form of this equation is written as t t t
u u 1 p 2u = ut uu x 0
u Fx v 2 (1) u is constant on each characteristic, the slope x΄(t) is
t x x x Moreover
constant and so the characteristics are straight lines, deter-
Now in absence of external force and no pressure gradient the mined by the initial data(figure)
above equation takes the form If the initial data is smooth then this can be used to determine
u u 2u the solution u(x,t) for small enough t that characteristics do not
u v 2
t x x cross. For each (x,t) we can solve the equation
x = ξ +u(ξ,0)t
u 1 2 2u
( u ) v For ξ and then
t x 2 x 2 U(x,t) = u(ξ,0)
There is an implicit relation that determines the solution of the
This equation is known as burger’s equation. It was named for inviscid burger’s equation provided characteristics do not in-
Johannes Martinus Burger’s (1895 – 1981). This is non-linear tersect. If the characteristics do intersect, then a classical solu-
2nd order partial differential equation. tion to the PDE does not exist.
Burger’s equation is a fundamental partial differential equa- The viscous Burger’s equation can be linearized by the Cole-
IJSER
tion from fluid mechanics. It occurs in various areas of applied Holf substitution
mathematics, such as modeling of traffic flow and gas dynam- 1 2
U 2v v 2 ,
ics etc. x t x
When ν→0, Burger’s equation becomes the inviscid burger’s
which turns into the diffusion equation
equation and written as
2
v 2
u u t x
u 0
t x That allows one to solve an initial value problem:
This is 1st order Quasi-linear partial differential equation.This U(x,t)= 2v
1 x
is prototype for equation for which the solution can develop ( x x) 1
discontinuous (shock wave). The previous equation is the con-
ln{(4vt)
2
exp[
4vt
u ( x,0)dx]dx}
2v 0
vection form of the Burger’s equation.
3 Analytical Solution
In this chapter, we solve 1d viscid burger's equation for initial
2.2 Solution of inviscid Burger’s equation condition in infinite space analytically by transforming to heat
equation
3.1 Burger's equation as an IV problem
We need to solve the following Initial value problem
u 1 2 2u
( u )v 2
t x 2 x (1)
With I.C. u ( x,0) u 0 ( x ) , for x (2)
3.2 The Cauchy Problem
Fig. 1. characteristics and solution for Burger’s equation (small t)
The Cauchy problem for the Heat Equation is
The inviscid Burger’s equation is first order partial differential
2
equation. Its solution can be constructed by the method of 2 (3)
characteristics. t x
Consider the inviscid equation in the above figure with 1
x
e e
pending on the similarity variable , or on . However,
t t 1 4 t
dx 4 t
dx
t t 0
there is a property of the heat equation we would like to pre-
serve in Our similarity solution, that of conservation of energy x 2 , we have,
Making the variable transformation y
. Suppose is a solution of the heat equation with the proper- 4t
2 t
y 1
2A
e
4 ty
1 dy A 4 e y y 2 dy
ty that (x,0)dx and x ( x , t ) 0 as
IJSER
t 0 0
IJSER
x y
0
3.6 Numerical evaluation of Analytical solution The function under integration sign in the numerator of is an
Now we are interested how our analytical solution behaves odd function, so we must have
when we try to implement it numerically. In order to perform y2 1
numerical estimation, we have to consider a function u 0 for ( y ) exp 4t 2 cos y dy 0
which the two integrations appeared in the numerator and
Which implies that u (0, t ) 0 .
denominator of (13) converge. We know that any bounded
function does the trick. We consider the bounded periodic Now, for x 2 , we have,
function u 0 ( x ) sin x as initial condition and find the solu- ( 2 y) 2 1
tion over the bounded spatial domain [0,2] at different time ( 2 y ) exp 4t 2 cos y dy
u (2, t )
steps. For the above initial condition we get the following ana- ( 2 y) 2
1
lytical solution of Burger's equation, t exp cos y dy
4t 2
Making the variable change z 2 y , we have,
IJSER © 2014
http://www.ijser.org
International Journal of Scientific & Engineering Research, Volume 5, Issue 4, April-2014 495
ISSN 2229-5518
z2 1 uin1 uin (uin ) 2 (uin1 ) 2
4t 2 cos ydy
z exp
t
2x
0
u (2, t )
z2 1 n1 t
t exp or, ui ui [(uin ) 2 (uin1 ) 2 ]
n
cos y dy
4t 2 2x
The function under integral sign in the numerical of is which is the explicit upwind difference scheme of inviscid
an odd function, so we have, Burger’s equation.
z2 1 4.2 Explicit Central Difference Scheme of Burger’s
4t 2 cos ydy 0
z exp equation for inviscid fluid
Consider the inviscid Burger’s equation as an initial boundary
which implies that u (2, t ) 0 . value problem
So we have the boundary values u 1 2
u(0, t ) u(2, t ) 0 ( u ) 0; t>0, x (a,b) (18)
t x 2
I.C u (0, x) u0 ( x)
4 Numerical Solution
In this chapter, we present some numerical methods to solve B.C u(t , a) ua (t ) , u(t , b) ub (t )
the Burger's equationas an initial boundary value problem.
We use explicit and implicit finite difference schemes to solve Now we take the forward difference formula for time deriva-
Burger's equationand then try to pro proceed in a different tive
way using Cole-Hopf transformation. We solve our C-H trans- u uin1 uin
IJSER
formed heat equation with Neumann boundary conditions
usingboth of explicit and implicit finite difference schemes for
t t
heat equation. And the central difference formula for spatial derivative we
4.1 Explicit Upwind Difference Scheme of Burger’s know
equation for inviscid fluid w win1 win1
Consider the inviscid Burger’s equation as a initial boundary
x 2x
value problem
For equation (1) w=1/2 u2 then
u 1 2
( u ) 0 ; t>0, x (a,b) (16) 1 2 1 2 (u n ) 2 (uin1 ) 2
t x 2 ( u ) (u ) i1
x 2 2 x 4x
I.C u (0, x) u0 ( x)
Then from (1) we can write
B.C u(t , a) ua (t ) , u(t , b) ub (t )
uin1 uin (uin1 ) 2 (uin1 ) 2
+
From (1) we have t 4x
u u t
u 0 (17) uin1 uin [(uin1 ) 2 (uin1 ) 2 ]
t x Or,
4x
Now we get the explicit upwind difference scheme for this
uin1 uin1
initial boundary value problem using forward difference for Consider as uin
time derivative 2
u uin1 uin Then we can write,
0 uin1 uin1 t
t t uin1 [(uin1 ) 2 (uin1 ) 2 ]
and the backward difference for spatial derivative 2 4x
w w w n n which is the Lux-Friedrich scheme.
i i 1
0 4.3 Explicit Central Difference Scheme of Burger’s
x x equation for viscous fluid
1 2 1 2 (uin ) 2 (uin1 ) 2 Consider the Burger’s equation as an initial boundary value
For w= u we find (u ) 0
2 2 x 2x problem
Then from (8) we get
IJSER © 2014
http://www.ijser.org
International Journal of Scientific & Engineering Research, Volume 5, Issue 4, April-2014 496
ISSN 2229-5518
u 1 2 2u x i a ih , i 0,1,............., M (22)
( u )v 2 ; x (a,b), t (0,T) (18)
t x 2 x and t n nk, n 0,1,..............N (23)
I.C. u (0, x) u0 ( x)
Where,
B.C. u(t , a) ua (t ) , u(t , b) ub (t ) ba T
M and N
For equi-distant grid, with temporal step-size Δt and spatial h k
step-size Δx Now we are interested to solve the above IBVP numerically.
u
The discretization of is obtained by forward difference 4.5 Explicit finite difference scheme
t To obtain an explicit finite difference scheme, we discretize
formula
u u 2u
n1 , and at any discrete point ( x i , t n ) as follows:-
u u u n
t x x 2
i i
t t u uin1 uin
(First order Forward difference formula)
1 2 t k
The discretization of ( u ) is obtained by backward dif-
x 2 u uin11 uin1
ference formula (First order Central difference formula)
x 2h
1 2 1 2 (u n ) 2 (uin1 ) 2 2u uin1 2uin uin1
( u ) (u ) i (Second order Centered difference
x 2 2 x 2x 2 x h2
IJSER
2u formula)
The discretization of is obtained by centeral difference Inserting the above formulas, the discrete version of the viscid
2 x Burger's equation formulates the second order finite difference
approximation
scheme of the form
2u uin1 2uin uin1 uin1 uin uin n u n 2uin uin1
(ui1 uin1 ) = v × i1
2 x x 2 k 2h h2
Using the above approximations in (1) we have Or,
0.3
=0.3
=0.5
=1
.02;∆t = .02 0.2
Clearly, for v = .01, shock occurs as time grows in the middle of 0.1
u
the graph. -0.1
-0.2
-0.4
-0.2
scheme for the viscous Burger’s equation, we discuss our ex- -0.6
IJSER
-0.8
0 1 2 3 4 5 6 7
We perform the numerical experiment for the equation Fig. 3. Solution of viscous Burger’s equation using explicit finite difference
u 1 2 2 scheme at different viscosity ν with ∆x=0.1, ∆t = 0.0016.
( u )v 2. We use the initial value which
t x 2 x After calculating stability condition we consider ν = 3.3, 3.6,
satisfies stability condition.
3.9, 4.2, 4.5, 6.0 for h = 0.1 and k = 0.0008196, we get the follow-
Now we implement the numerical finite central difference ing fig. 4.
scheme of viscous Burger’s equation by computer program- 12
x 10
-3
v= 3.3
-4
0 1 2 3 4 5 6 7
Fig. 4. Solution of viscous Burger’s equation using explicit finite differ-
.002
max u i0 v ence scheme at different viscosity ν with ∆x=0.1, ∆t = 0.0008196.
2
But since the initial condition is u0 ( x) sin x , so we must After calculating stability condition we consider ν= 6.3, 6.6, 6.9,
have, 7.2, 7.5, 9.0 for h = 0.1 and k = 0.000555, we get the following
max ui0 1
fig. 5.
3.5
x 10
-4
v=
v=
6.6
6.9
.002 2.5
v 2
2
u
1.5
i.e. 0.001 v 1
2
0.002 0.5
2*k v
3.5
v=
v=
12.5
13.5
tions, Wiley-Interscience, New York, 1994.
v= 15.0
2 [13] C.I. Byrnes, D.S. Gilliam and V.I. Shubov. Boundary control, feed-
1.5 back stabilizationand the existence of attractors for a viscous Burgers'
1
Equation, (1994), Preprint.
0.5
0
0 1 2 3 4 5 6 7
v
6 Conclusions
Burger’s equation is one of the interesting and implemented
equations in our practical life for both viscous and inviscid
fluid. In this paper, we have considered Burger’s equation is
IJSER
fundamental partial differential equation from fluid mechan-
ics. First we have shown derivation of Navier-Stokes equation,
Burger’s equation and numerical methods of Burger’s equa-
tion. At last we have shown numerical result based on the ex-
plicit central difference scheme agrees with basic qualitative
behabior of viscuss Burger’s equation. In future, we try to de-
velop better numerical ways to solve Burger's equation with
initial condition and two-sided boundary conditions infinite
space.
REFERENCES
[1] Soliton and Numerical Solutions ofthe Burgers’ Equation and com-
paring them- Int. Journal of Math. Analysis, Vol. 4, 2010, no. 52, 2547
- 2564
[2] Burgers, J. M. A Mathematical Model Illustrating the Theory of Tur-
bulence. Advances in Applied Mechanics, Academic Press, New York
1, 171-199(1948).
[3] H.Bateman, Some recent researches on the motion of fluids, Monthly
Weather Rev. 43(1915) 163-170.
[4] J.M.Burgers, A mathematical model illustrating the theory of turbu-
lence, Adv. Appl. Mech. 1(1948) 171-199.
[5] E.L.Miller, Predictor-corrector studies of Burgers' model of turbulent
flow, M.S. thesis, University of Delaware, Newark, DE, (1966). 50 S.
Dhawan
[6] Cole, J.D., on a quasi-linear parabolic equation occurring in aerody-
namics, Quart. Applied Math., 1951, 225-236.
[7] Mittal R.C., Singhal P., Numerical solution of the Burgers’ equation,
Commun. Numer. Meth. Eng., 1993, 9, 397-406.
[8] Hassan N. A. Ismail, Kamal Raslal, Aziza A. Abd Rabboh, “Adomain
decomposition method for Burger’s Huxley anBurger’s Fisher equa-
tions”, Applied mathematics and computation, 159, 291-301, 2004.
[9] Ninghu Su, Jim P.C. Watt, Keith W. Vincent, Murray E. Close, Ren-
zhao Mao, Analysis of turbulent flow patterns of soil water under
field conditions using Burgers’equation and porous suction-cup
samplers. Aust. J. soil Res. (42). (1). (2004), pp. 9-16.
IJSER © 2014
http://www.ijser.org