Basic Concepts: Partial Differential Equations (Pde)

Download as pdf or txt
Download as pdf or txt
You are on page 1of 19

PARTIAL DIFFERENTIAL EQUATIONS (PDE)

Basic Concepts

An equation involving one or more partial


derivatives of an unknown function of two or
more independent variables is called a PDE.
The order of the highest derivative is called the
order of the equation
PDE is linear if it is first degree in the dependent
variable and its derivative.
A solution of a PDE is a function that satisfies the
equation.

Example:

u = x2 y2, u = ex cos y, u = In(x2 + y2 )

are functions that are different from each other but


are solutions to

2u 2u
0
x 2
y 2

Focus will be given in solving second order and a


linear PDE.
A general first order and linear PDE with three
variables (one dependent and two independents)
is

1
u u
a x, y
b x, y f x, y u g x, y 0
x y
A general second order and linear PDE with
three variables (one dependent and two
independents) is

2u 2u 2u u u
a x, y 2 b x, y c x, y 2 d x, y e x, y
x xy y x y

f x, y u g x, y 0

Important linear PDE of the second order

1. Wave equation:

2u 2u
c 2
t 2
x 2

2. Heat equation:

u 2 u
2
c
t x 2

3. Laplace equation:

2u 2u
0
x 2
y 2

2
4. Poisson equation

2u 2u
f x, y
x 2
y 2

5. 3-dimensional Laplace equation

2u 2u 2u
0
x 2
y 2
z 2

In the above equations, c is a constant, t is time and x,


y, and z are cartesian coordinates. For equation (4)
where f 0 are said to be homogenous are other
equations are non- homogenous.

WAVE EQUATION

MODELLING: VIBRATING STRING

Objective: to derive the equation governing small


transverse vibrations of an elastic string violin string.
Stretch the string to length L and fix it at the ends.
Then distort at some instant say t = 0, release it and
allow it to vibrate.

3
The problem is to determine the vibrations of the
string, that is to find the deflection u(x, t ) at any point
x and at any time t > 0.

Physical assumptions.

1. The mass of the string per unit length is


constant. The string is perfectly elastic and does
not offer any resistance to bending.
2. The tension caused by stretching the string
before fixing at the ends is so large that the
action of gravitational force on the string can
be neglected.
3. The string performs small transverse motions in
a vertical plane; that is every particle of the
string moves strictly vertically and so that the
deflection and slope at every point of the string
always remain small.

Differential Equation from Forces

To obtain the differential equation we consider the


forces acting on a small portion of the string. The
tension is tangential to the curve of the string at each
point.

Let T1 and T2 be the tension at the endpoints of P and


Q of that portion. Since there is no motion in
horizontal direction, horizontal components of the
tension must be constant. Therefore

4
T1 cos T2 cos T constant (1)

In vertical direction
2u
T2 sin T1 sin x 2
t

Using (1) we can divide this by T1 cos T2 cos T


obtaining

T2 sin T1 sin x 2u
tan tan (2)
T2 sin T1 cos T t 2

Now tan and tan are the slopes of the string at x


and x + x:

u u
tan and tan
x x x x x

Dividing (2) by x we therefore have

1 u u u
2

x x x x x x T t 2

We let x approaches zero, we obtain the linear PDE

2u 2 u
2
T
c , c2
t 2 x 2

5
Separation of Variables and Use of Fourier Series

The vibration of an elastic string are governed by the


one dimensional wave equation

2u 2u
c 2
(1)
t 2
x 2

where u(x, t) is the deflection of the string. To find


how the string moves, we solve this equation.
That is to determine the solution u that satisfies the
conditions imposed by the physical system.
Since the string is fixed at the ends x = 0 and x = L, we
have two boundary conditions

u(0, t) = 0, u(L, t) = 0 for all t (2)

The form of the motion of the string will depend on


the initial deflection (deflection at t = 0) and the initial
velocity (velocity at t = 0). Let the initial deflection be
f(x) and initial velocity be g(x). we therefore have two
initial values

u(x, 0) = f(x) (3)

u
g x (4)
t t 0

6
The problem is to find the solution to (1) that satisfies
the conditions (2) (4). Proceed step by step.

First Step:

By applying the method of separating variables or


product method, we shall obtain two ordinary
differential equations.

Second Step:

Determine the solutions of these two equations that


satisfy the boundary conditions.

Third Step:

Using Fourier Series, compose these solutions, in


order to get a solution of the wave equation (1) that
also satisfies the given initial conditions.

First Step: Two Ordinary Differential equations

In the method of separating variables, we determine


the solution of the wave equation (1) of the form

u(x, t) = F(x)G(t) (5)

7
which is a product of two functions, each depending
only on one of the variables x and t. By differentiating
(5) we obtain

2u 2u
FG and F "G (6)
t 2 x 2

where denote derivatives with respect to t and ''


derivatives with respect to x. By inserting (6) into
differential equation (1) we have

FG = c2F"G

Dividing by c2FG, we find

G F "
2

c G F

Expression on the left depends only on t, and that on


the right on x. So we claim that the equation equals a
constant k. If they were variable, then changing t or x
would affect only the left or right side, respectively.
Thus

G F"
2
k
c G F

This yields two ODE namely

F"- kF = 0 (7)

8
and G c2kG = 0 (8)

Second Step: Satisfying the Boundary Conditions

Now determine solution F and G for (7) and (8) so


that u = FG satisfies the boundary conditions (2), that
is

u(0, t) = F(0)G(t) = 0, for all t

u(L ,t) = F(L)G(t) = 0, for all t

If G = 0, then u = 0 which is of no interest. Thus G 0


and then

F(0) = 0 and F(L) = 0 (9)

In order to obtain a solution, take into consideration


three different values of k,

Case (i), k zero (k = 0)

The general solution to (7) is F = ax + b, and from (9)


we obtain a = b = 0. Hence F = 0, which is of no
interest because then u will be zero.

Case (ii), k positive (k = p2)

The general solution to (7) is

9
F Ae px Be px

And from (9) we obtain F = 0, as for case (i). Hence


we are left with the possibility of choosing k negative.

Case (iii), k negative (k = -p2)

With the value of k negative, equation (7) becomes

F" + p2F = 0

Its general solution is

F(x) = A cos px + B sin px (10)

From (10) and (9) we have

F(0) = A = 0 and then

F(L) = B sin pL = 0

We must take B 0, since otherwise F = 0. Hence

sin pL = 0

or pL = n

so that p = (n)/L, (11)

10
di mana n integer positif.

Setting B = 1, we thus obtain infinitely many solutions,


F(x) = Fn(x), where

n
Fn x sin x where n = 1, 2, 3, (12)
L

The solution satisfies (9). {for negative integer n we


obtain the same solutions, except for a minus sign,
because sin (-) = -sin }.

k is restricted to k = -p2 = -(n/L)2. For these k,


equation (8) becomes

G + n2G = 0, where n = (cn)/L.

A general solution is

Gn(t) = Bn cosnt + Bn* sin nt

Hence the function un(x,t) = Fn(x) Gn(t) can be written


as


un x, t Bn cos nt Bn* sin nt sin
n
L
x (13)

where n = 1, 2, 3,

11
Equation (13) is a solution to (1), that satisfying
boundary conditions (2).

Third Step: Solution of the Entire Problem

A single solution un(x, t) will in general not satisfy the


initial conditions (3) and (4). Since equation (1) is
linear and homogeneous, it follows from
Fundamental Theorem that the sum of finitely many
solutions un , is a solution to (1). To obtain a solution
that satisfies (3) and (4), we consider the infinite series



u x, t un x, t Bn cos nt Bn* sin nt sin
n x
n 1 n 1 L

(14)

where n = cn/L, same as before. From (3), we


obtain

n
u x,0 B n sin x f x (15)
n1 L

Hence for (14) to satisfy (3), coefficient Bn must be


choosen so that u(x,0) become the Fourier sine series
of f(x), thus

2 L n x

L 0
Bn f x sin dx (16)
L

12
n = 1, 2, 3,

By differentiating (14) with respect to t and using (4),


we obtain

u



t t 0 n1
B
n n sin nt B *

n n cos n t
sin
n x
L

n x
= Bn* n sin g x
n1 L

Hence for (14) to satisfy (4), coefficient Bn* must be


choosen so that for t = 0, derivative u/t becomes
the Fourier sine series for g(x); thus

2 L n x

L 0
Bn* n g x sin dx
L

or

2 L n x
Bn* g x sin dx (17)
cn 0 L

As a result, from (14) with coefficients (16) and (17),


u(x,t) is a solution to (1) that satisfies the boundary
conditions (2) (4), provided the series (14)
converges and so do the series obtained by

13
differentiating (14) twice with respect to x and t, and
have sums u2/x2 and 2u/t2 respectively.

HEAT EQUATION

Heat flow in a body of homogenous material is


governed by the heat equation,
u 2 2 k
c u c2
t

where u(x,y,z,t) is the temperature in a body. Here c2 is


the thermal diffusivity k the thermal conductivity, is
the specific heat, and is the density of the material of
the body. 2u is the Laplacian of u, and with respect
to Cartesian coordinates x, y, z,

2u 2u 2u
u
2

x 2
y 2
z 2

Important application:

Consider temperature in a long thin bar or wire


of constant cross section and homogenous
material
Oriented along the x-axis
Perfectly insulated laterally-heat flows in x
direction only

14
Then u depends on x and time t, and the heat
equation becomes the one-dimensional heat equation

u 2 u
2
c (1)
t x 2

Solve (1) for some types of boundary and initial


conditions.
Begin with the case in which the ends x = 0 and x = L
of the bar kept at temperature zero.
So boundary conditions are

u(0, t) = 0, u(L, t) = 0 for all t (2)

and the initial temperature in the bar is f(x). So the


initial condition is

u(x, 0) = f(x) (3)

Need to determine a solution u(x,t) for (1) satisfying


(2) and (3).

First Step: Two ODEs

Use the method of separating the variables. First


determine the solution to (1) that satisfies the
boundary conditions (2). Assume that the solution is

u(x, t) = F(x)G(t) (4)

15
Substitute this expression into (1), we will obtain
FG = c2F"G
where denote differentiation with respect to t and "
differentiation with respect to x.

To separate the variables, divide the equation by


c 2 FG, that is
G F "
2
(5)
c G F

Expression on the left depends on t, whereas on the


right depends on x. Therefore both sides must be
equal to a constant , k.

For k 0, the solution u = FG that satisfies (2) is u = 0.


For negative k that is k = -p2, obtain from (5)

G F "
p 2
c 2G F

Two ODE are obtained

F" + p2F = 0 (6)

G + c2p2G = 0 (7)

Second Step: Satisfying The Boundary Conditions

From (6), a general solution is

16
F(x) = A cos px + B sin px (8)

From the boundary conditions (2),

u(0,t) = F(0)G(t) = 0 and

u(L,t) = F(L)G(t) = 0

Since G(t) = 0 would give u = 0, we require F(0) = 0


and F(L) = 0. From (8), F(0) = A. we have A = 0 and
therefore

F(L) = B sin pL

B 0, or else F = 0. Therefore the condition F(L) = 0


will result in

n
sin pL = 0, hence p = n = 1, 2,
L

Setting B = 1, obtain following solution of (6)


satisfying (2):

nx
Fn x sin n 1, 2, 3,
L

Now solve equation (7). For p = n/L

cn
G n 2G 0 where n
L

17
It has the general solution

Gn t Bn e 2nt
n = 1, 2, 3,

where Bn is a constant. Hence the functions

nx 2nt
un x, t Fn x Gn t Bn sin e (9)
L

are solutions of the heat equation (1) satisfying (2).

To obtain the overall solution we have to use infinite


series.

Third Step: Solution of the Entire Problem

To obtain a solution also satisfying the boundary


conditions (3), consider a series


nx n 2t
u x, t un x, t Bn sin e (10)
n 1 n 1 L

cn
n
L

From this series and (3), we have

18

nx
u x,0 Bn sin f x
n 1 L

Hence for (10) to satisfy (3), the Bns must be


coefficient of Fourier sine series , thus

2 L nx

L 0
Bn f x sin dx (11)
L

19

You might also like