Beruflich Dokumente
Kultur Dokumente
Lecture Notes
ii
Contents
1 Lecture 01
1.1 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.2 Basic Definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1
1
2
2 Lecture 02
2.1 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.2 Initial conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.3 Boundary conditions . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.4 Solution of a Boundary Value Problem . . . . . . . . . . . . . . . . .
2.5 Linear Boundary Conditions . . . . . . . . . . . . . . . . . . . . . . .
2.6 Classification of Boundary Conditions . . . . . . . . . . . . . . . . .
2.6.1 Dirichlet Conditions . . . . . . . . . . . . . . . . . . . . . . .
2.6.2 Neumann Conditions . . . . . . . . . . . . . . . . . . . . . . .
2.6.3 Mixed Boundary Conditions or Robins Boundary Conditions
2.7 Superposition Principle . . . . . . . . . . . . . . . . . . . . . . . . .
2.8 Formation of Partial Differential Equation . . . . . . . . . . . . . . .
9
9
9
10
10
10
11
11
11
11
13
14
3 Lecture 03
3.1 First Order Partial Differential Equations . . . . . . . . . . . . . . .
3.2 First Order Partial Differential Equation . . . . . . . . . . . . . . . .
3.3 Solution of First Order Partial Differential Equation . . . . . . . . .
17
17
18
19
4 Lecture 04
4.1 Solution of First Order Partial Differential Equations
4.1.1 Method of Characteristics . . . . . . . . . . .
4.1.2 Examples . . . . . . . . . . . . . . . . . . . .
4.2 Integral Surfaces Passing Through a Given Curve . .
4.2.1 Parametric Equations . . . . . . . . . . . . .
4.2.2 Methodology . . . . . . . . . . . . . . . . . .
4.2.3 Examples . . . . . . . . . . . . . . . . . . . .
21
21
21
21
23
23
23
24
iii
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
iv
5 Lecture 05
5.1 Compatible System of First Order Differential
Equations . . . . . . . . . . . . . . . . . . . .
5.1.1 Example . . . . . . . . . . . . . . . . .
5.2 Non-linear Partial Differential Equations . . .
5.2.1 Charpits Method . . . . . . . . . . .
5.2.2 Examples . . . . . . . . . . . . . . . .
CONTENTS
25
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
6 Lecture 06
6.1 Non-linear Partial Differential Equations . . . . . . . . . . .
6.1.1 Charpits Method . . . . . . . . . . . . . . . . . . .
6.1.2 Examples . . . . . . . . . . . . . . . . . . . . . . . .
6.2 Special Types of First Order Non-Linear Partial Differential
6.2.1 Equations involving p and q only . . . . . . . . . . .
6.2.2 Separable Equations . . . . . . . . . . . . . . . . . .
6.2.3 Clairauts Form . . . . . . . . . . . . . . . . . . . . .
7 Lecture 07
7.1 Second Order Partial Differential Equation . . .
7.1.1 Definition . . . . . . . . . . . . . . . . . .
7.2 Classification of Second Order Partial Differential
7.2.1 Examples . . . . . . . . . . . . . . . . . .
7.3 Reduction of Second Order Partial Differential
Equation into Canonical Form . . . . . . . . . . .
8 Lecture 08
8.1 Reduction of Second Order Partial Differential
Equations in Canonical Form . . . . . . . . . .
8.1.1 Result . . . . . . . . . . . . . . . . . . .
8.2 Canonical Form for Hyperbolic Equation . . . .
8.2.1 Examples . . . . . . . . . . . . . . . . .
9 Lecture 09
9.1 Reduction of Second Order Partial Differential
Equations in Canonical Form . . . . . . . . . .
9.1.1 Result . . . . . . . . . . . . . . . . . . .
9.2 Canonical Form for Parabolic Equation . . . .
9.2.1 Examples . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
25
26
26
26
27
. . . . .
. . . . .
. . . . .
Equations
. . . . .
. . . . .
. . . . .
29
29
29
30
30
30
31
32
.
.
.
.
.
.
.
.
33
33
33
33
34
. . . . . . . . . . .
35
. . . . . .
. . . . . .
Equations
. . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
37
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
37
38
38
40
41
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
41
42
42
43
10 Lecture 10
10.1 Reduction of Second Order Partial Differential
Equations in Canonical Form . . . . . . . . . . . . . . . . . . . . . .
10.1.1 Result . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
45
45
46
CONTENTS
46
47
11 Lecture 11
11.1 Reduction of Second Order Partial Differential
Equations in Canonical Form . . . . . . . . . .
11.1.1 Result . . . . . . . . . . . . . . . . . . .
11.2 Mixed Type of Partial Differential Equation . .
11.2.1 Example . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
49
50
50
50
12 Lecture 12
12.1 Mathematical Modeling . . . . . . . . . .
12.1.1 Objective . . . . . . . . . . . . . .
12.1.2 Background . . . . . . . . . . . . .
12.1.3 Approximations and Idealizations .
12.1.4 Modeling . . . . . . . . . . . . . .
12.1.5 Model Validation . . . . . . . . . .
12.1.6 Compounding . . . . . . . . . . . .
12.2 Review of Physics Laws . . . . . . . . . .
12.3 The Continuity Equation . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
53
53
54
54
54
54
55
56
56
57
.
.
.
.
.
.
59
59
63
63
64
64
65
13 Lecture 13
13.1 The Heat (Or Diffusion) Equation
13.2 Initial Conditions . . . . . . . . . .
13.3 Boundary Conditions . . . . . . . .
13.4 Initial Boundary Value Problem . .
13.4.1 Examples . . . . . . . . . .
13.4.2 Solution . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
49
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
vi
CONTENTS
Lecture 01
1.1
Motivation
Chapter 1 Lecture 01
Course Contents
Introduction of partial differential equations
Partial Differential equations of first order
Linear and non-linear partial differential equations
Applications of first order partial differential equations
Partial Differential equations of second order
Mathematical Modeling of heat, Laplace and wave equations
Classification of second order partial differential equations
Boundary and Initial value problems
Reduction to canonical form and the solution of second order partial
differential equations
Sturm-Liouville system
Technique of Separation of Variable for the solution of partial differential
equations
Laplace, Fourier and Hankel transforms for the solution of partial differential
equations and their application to boundary value problems.
Recommended Books
Richard Haberman, Elementary partial differential equations, Prentice-Hall,
INC., Englewood Cliffs, New Jersy
K. Sankara Rao, Introduction to partial differential equations, Prentice-Hall
of India New Delhi.
M. Humi, W. B. Miller, Boundary value problems and partial differential equations, PWS-KENT publishing company, Boston.
T. Myint-U, L. Debnath, Linear partial differential equations for scientists and
engineers, Fourth Edition, Birkhauser, Berlin
Grading Scheme
1.2
Basic Definitions
Differential Equations
Definition 1.1. A differential equation is an equation that relates the derivatives
of a function depending on one or more variables.
For example
d2 u du
+
= cos x
dx2 dx
is a differential equation involving an unknown function u(x) depending on one
variable and
2u 2u
u
+ 2 =
2
x
y
t
is a differential equation involving an unknown function u(t, x, y) depending on three
variables.
Partial Differential Equation (PDE)
Definition 1.2. A partial differential equation (PDE) is an equation that contains,
in addition to the dependent and independent variables, one or more partial derivatives of the dependent variable.
Suppose that our unknown function is u and it depends on the two independent
variables then the general form of the PDE is
F (x, y, , u, ux , uy , uxx , uxy , uyy , ) = 0
Here subscripts denotes the partial derivatives, for example
ux =
u
u
2u
2u
2u
, uy =
, uxx =
,
u
=
,
u
=
xy
yy
x
y
x2
xy
y 2
Chapter 1 Lecture 01
(1.2.1) eq:01
for any two functions u1 and u2 , where c1 and c2 are arbitrary constants. /t and
2 /x2 are the examples of linear operators since these two satisfy equation (4.2.1):
u1
u2
(c1 u1 + c2 u2 ) = c1
+ c2
t
t
t
2
2
u1
2 u2
(c
u
+
c
u
)
=
c
+
c
1
1
2
2
1
2
x2
x2
x2
Note: Any linear combination of linear operators is a linear operator.
However (/t)2 is not a linear operator since it does not satisfy equation (4.2.1).
To prove the non-linearity of the operator (/t)2 , we do the following calculations.
If our L = (/t)2 then Lu = (u/t)2
u
derivative
| {z }
u
t
Square
| {z }
u
t
2
Now
(c1 u1 + c2 u2 )
t
2
=
=
=
6=
u1
u2 2
c1
+ c2
t
t
2
u1
u2 2
u1 u2
c1
+ c2
+ 2c1 c2
t
t
t t
2
2
u2
u1 u2
u1
+ c22
+ 2c1 c2
c21
t
t
t t
2
2
u2
u2
c1
+ c2
t
t
Chapter 1 Lecture 01
(1.2.2) ?eq:2?
If u1 and u2 are two solutions of equations (4.1.1) then they must satisfy equation
(4.1.1), that is,
1 u1
2 u1
1 u2
2 u2
=
,
=
.
k t
x2
k t
x2
According to the principle of superposition, c1 u1 + c2 u2 is again the solution of
equation (4.1.1).
1
1 u1
1 u2
(c1 u1 + c2 u2 ) = c1
+ c2
k t
k t
k t
2 u1
2 u2
2
= c1
+ c2
=
(c1 u1 + c2 u2 ) = R. H. S.
x2
x2
x2
L.H.S. =
Note: Here the superposition principle is stated for two solutions only, it is true
for any finite linear combinations of solutions.
Furthermore, under proper restrictions, it is also true for infinite number of
solutions. If ui , i = 1, 2, are solutions of a homogeneous linear partial differential
equations, then
n
X
w=
ci ui
i=1
is also a solution.
Chapter 1 Lecture 01
n
X
bi (u)xi + = f (x)
i=1
n
X
i=1
au2 +
n
X
i=1
n
X
i=1
bi (u1 u2 )xi + = 0
Lecture 02
2.1
Motivation
2.2
Initial conditions
10
Chapter 2 Lecture 02
Alternatively, if the conditions relates one value of the independent variable such
as u(x0 ) = A and ux (x0 ) = B. Then these conditions are called Initial Conditions
and x0 is called the initial point.
Example:1 As a simple example, we suppose that our unknown function u is
dependent on one variable x. Then the following problem is known as initial value
problem
uxx + ux 2u = 0,
u(0) = 3,
ux (0) = 7.
Example:2 Now we suppose that our unknown function u is dependent on two
variable t, x. Then the following problem is known as initial value problem
uxx + ut 2u = 0,
2.3
u(0, x) = 3x,
ut (0, x) = sin x.
Boundary conditions
Definition 2.2. The set of conditions specified for the behavior of the solution to
a set of differential equations or partial differential equations at the boundary of its
domain are known as Boundary Conditions. A system with boundary conditions is
known as the boundary value problem.
Alternatively, the problem of finding the solution of a differential equation such
that all the associated conditions relate to two different values of the independent
variable is called a boundary value problem.
Example: If u(x, t) is the displacement of a vibrating string and its ends are
fixed at x = a and x = b, then the conditions
u(a, t) = 0,
and
u(b, t) = 0
2.4
2.5
Definition 2.3. The boundary conditions are linear if they express a linear relationship between u and its partial derivatives (up to appropriate order) on D. (In
other words, a boundary condition is linear if it is expressed as a linear equation
between u and its derivatives on D. )
11
2.6
2.6.1
Definition 2.4. The boundary conditions specify the values of the unknown function
u on the boundary. This type of boundary condition is called the Dirichlet condition.
2.6.2
Neumann Conditions
Definition 2.5. The boundary conditions specify the derivatives of the unknown
function u in the direction normal to the boundary, which is written as u/n. This
type of boundary condition is called the Neumann condition.
Remark: The normal derivative on the boundary u/n is defined as
u
u
u
= grad u n =
, ,
n,
n
x1
xn
where n is the outward normal to D.
2.6.3
for t > 0.
12
Chapter 2 Lecture 02
for t > 0.
for t > 0.
p(t)
,
K
and ux (L, t) =
q(t)
,
K
for t > 0.
K > 0 is the thermal conductivity. These also are Neumann, or free, boundary
conditions.
3) Each end-point is exposed and radiates heat into the surrounding medium
which has a temperature of T (t):
ux (0, t) = [u(0, t) T (t)], and ux (L, t) = [ux (L, t) T (t)], for t > 0,
which simplify to:
u(0, t) ux (0, t) = T (t), and u(L, t) + ux (L, t) = T (t), for t > 0,
where , , are positive constants. These are Robin, or mixed, boundary conditions.
If the surrounding medium has a temperature of 0 C (i.e., T (t) = 0), and u(x, t)
is measured in C, then we have
u(0, t) ux (0, t) = 0, and u(L, t) + ux (L, t) = 0, for t > 0.
13
2.7
Superposition Principle
u(0, y) = g1 (y)
u(x, b) = f2 (x),
u(a, y) = g2 (y)
2 u2 2 u2
+
=0
x2
y 2
u1 (x, 0) = f1 (x)
u2 (x, 0) = 0
u1 (x, b) = f2 (x)
u2 (x, b) = 0
u1 (0, y) = 0
u2 (0, y) = g1 (y)
u1 (a, y) = 0
u2 (a, y) = g2 (y)
14
Chapter 2 Lecture 02
u
(x, b) = f2 (x)
y
u
= g2 (x).
x
2.8
or
15
function of x, y, we obtain
F u u
F v
v
+
p +
+
p = 0
u x z
v x z
F u u
F v v
+
p +
+
p = 0
u y
z
v y z
Eliminating F/u and F/v from above equations, we obtain
P p + Qq = R,
where
P =
(u, v)
,
(y, z)
Q=
(u, v)
,
(z, x)
z
,
x
q=
and
p=
R=
(u, v)
,
(x, y)
z
.
y
Problem Sheet
Form the partial differential equation by eliminating the arbitrary function from
2z
x2
2z
t2
= 0.
2. (z y)p + (x z)q = y x
3. yp xq = y 2 x2
4. px = qy
5. z = px + qy + pq
16
Chapter 2 Lecture 02
Lecture 03
3.1
z
x ,
q=
z
y ;
x and y are
Example Form the partial differential equation from the following equation
x2 + y 2 + (z c)2 = a2
with a and c are arbitrary constants.
Note: The equation in above example represents the set of all spheres whose
center lie along the z-axis.
Example Form the partial differential equation from the following equation
x2 + y 2 = (z c)2 tan2
with c and are arbitrary constants.
Note: The equation in above example represents the set of all right circular
cones whose axis coincides with the line OZ.
Example
Form the partial differential equation from the following equation
x2 + y 2 + (z c)2 = a2
with a and c are arbitrary constants.
Solution: The obtained differential equation is
yp xq = 0.
17
18
Chapter 3 Lecture 03
Example Form the partial differential equation from the following equation
x2 + y 2 = (z c)2 tan2
with c and are arbitrary constants.
Solution: The obtained differential equation is
yp xq = 0.
Surfaces of Revolution
Now a question arises that why these two different geometrical entities have
the same partial differential equation.
The spheres and cones are basically surfaces of revolution which have the line
OZ as axes of symmetry.
Definition 3.2. A surface of revolution is a surface in Euclidean space created by
rotating a curve around a straight line in its plane (the axis).
All surfaces of revolution with the property that they have the line OZ as axes
of symmetry are characterized by an equation of the form
z = f (x2 + y 2 )
where the function f is arbitrary. The corresponding differential equation is
yp xq = 0.
3.2
3.3
19
20
Chapter 3 Lecture 03
Lecture 04
Review
Formation of First Order Partial Differential Equations
Solution of First Order Partial Differential Equations
4.1
4.1.1
(4.1.1) eq:02
4.1.2
Examples
22
Chapter 4 Lecture 04
z
z
z
+ P2
+ + Pn
= R.
x1
x2
xn
23
4.2
4.2.1
Parametric Equations
Definition 4.2. Parametric equations are a set of equations that express a set of
quantities as explicit functions of a number of independent variables, known as parameters.
Examples:
The equation of a circle in Cartesian coordinates is x2 + y 2 = r2 , while its one
set of parametric equations are
x = r cos
y = r sin
x = t,
while the second set is
x=
4.2.2
t,
y = t.
Methodology
v(x, y, z) = C2 .
(4.2.1) eq:01
Then the solution of the partial differential equation can be written in the form
F (u, v) = 0.
24
Chapter 4 Lecture 04
Now we wish to determine an integral surface, containing a given curve described by the parametric equation of the form
x = x(t),
y = y(t),
z = z(t),
Eliminating the parameter t from above equations to obtain a relation of the type
F (C1 , C2 ) = 0. Substituting the values of C1 and C2 from equation (4.2.1) leads to
the integral surface passing through the given curve .
4.2.3
Examples
Lecture 05
Review
Solution of First Order Partial Differential Equations
Method of Characteristics
Examples
Integral Surfaces Passing Through a Given Curve
Parametric Equations
Methodology
Examples
5.1
Definition 5.1. Two first order partial differential equations are said to be compatible if they have a common solution.
Necessary and Sufficient Condition for Compatible Differential Equations
Consider two partial differential equations
f (x, y, z, p, q) = 0,
g(x, y, z, p, q) = 0.
Then the necessary and sufficient condition for compatible differential equations
is to satisfy the following equation
(f, g)
(f, g) (f, g)
(f, g)
+p
+
+q
= 0.
(x, p)
(z, p) (y, q)
(z, q)
25
26
Chapter 5 Lecture 05
5.1.1
Example
Example Show that the following partial differential equations are compatible
xp yq = x,
x2 p + q = xz
and
Home Work
5.2
5.2.1
27
g
g
g
g
g
+ fq
+ (pfp + qfq )
(fx + pfz )
(fy + qfz )
= 0.
x
y
z
p
q
from which we can determine g. The auxiliary equations of above quasi-linear equation are
dy
dz
dp
dq
dx
=
=
=
=
.
fp
fq
pfp + qfq
(fx + pfz )
(fy + qfz )
These equations are called Charpits equations.
Note:
Any Integral of the above Charpits equations which involves p and q can be
taken as the second equation
g(x, y, z, , p, q) = 0.
It may also be noted that all Charpits equations need not to be used, but it
is enough to choose the simplest.
5.2.2
Examples
28
Chapter 5 Lecture 05
Lecture 06
Review
Compatible System of First Order Differential Equations
Examples
Non-linear Partial Differential Equations
Charpits Method
Examples
6.1
6.1.1
30
Chapter 6 Lecture 06
6.1.2
Examples
6.2
z2
.
a
6.2.1
z
z
= a, q =
= b.
x
y
31
p + q = 1.
Solution: The given partial differential equation of the form f (p, q) = 0. Therefore we assume that its solution in the form
z = ax + by + c,
where
a+
b=1
or
b = (1
a)2 .
z = ax + (1 a)2 y + c.
6.2.2
Separable Equations
Definition 6.1. An equation in which z is absent and the terms containing x and
p can be separated from those containing y and q is called a separable equation.
Example Find the complete integral of the following partial differential equation
p2 y(1 + x2 ) = qx2 .
Solution: The given equation can be written as
p2 (1 + x2 )
q
= .
2
x
y
Assume that this equation is equal to a, an arbitrary constant, such that
p2 (1 + x2 )
q
= = a.
x2
y
Then,
p2 (1 + x2 )
q
= a,
= a.
x2
y
Or
ax
p=
, q = ay.
1 + x2
Substituting these values of p and q in
dz = pdx + qdy,
we obtain
ax
dz =
dx + aydy
1 + x2
On Integration we obtain the solution of the given equation
p
a
z = a 1 + x2 + y 2 + b.
2
32
Chapter 6 Lecture 06
6.2.3
Clairauts Form
p
1 + a2 + b2 .
Lecture 07
Review
Non-linear partial differential equations
Charpits Method
Examples
Special types of non-linear partial differential equations are
Type 1: Equations involving p and q only
Type 2: Separable equations
Type 3: Clairauts form.
7.1
7.1.1
Definition 7.1. The general form of a second order partial differential equation is
Auxx + 2Buxy + Cuyy + Dux + Euy + F u + G = 0,
where A, B, C, D, E, F, G are the functions of x and y only or in a simple case these
coefficients are constant.
7.2
34
Chapter 7 Lecture 07
The discriminant of the above equation is B 2 AC, then the above equation is
Elliptic
if
B 2 AC < 0
Parabolic
if
B 2 AC = 0
Hyperbolic
if
B 2 AC > 0.
Generalizing this concept for the second order partial differential equations
Auxx + 2Buxy + Cuyy + Dux + Euy + F u + G = 0.
Then we classify the above above partial differential equation as
7.2.1
Elliptic
if
B 2 AC < 0
Parabolic
if
B 2 AC = 0
Hyperbolic
if
B 2 AC > 0.
Examples
Or
uxx ut = 0
1
utt ,
c2
Or
uxx
1
utt = 0.
c2
Here A = 1, B = 0 and C = c12 , and now we calculate the discriminant such that
B 2 AC = 0 (1)(
Therefore it is a hyperbolic equation.
1
1
) = 2 > 0.
2
c
c
7.3
35
(, ) x
J=
= x
6= 0
(x, y) y y
In the new coordinates (, ) the second order partial differential equation
2u
2u
2u
u u
A 2 + 2B
+ C 2 + F x, y,
,
=0
x
xy
y
x y
can be written as
2u
2u
2u
A 2 + 2B
+C 2 +F
u u
x, y,
,
=0
where
2
2
A = A
+ 2B
+C
,
x
x y
y
B = A
+
+
+C
,
x x
x y y x
y y
2
2
+ 2B
.
C = A
+C
x
x y
y
36
Chapter 7 Lecture 07
Lecture 08
Review
Second Order Partial Differential Equations
Definition
Classification of Second Order Partial Differential Equation
Examples
Reduction of Second Order Partial Differential Equations in Canonical Form
8.1
u u
,
=0
x, y,
x y
u u
x, y,
,
=0
where
2
2
A = A
+ 2B
+C
,
x
x y
y
B = A
+B
+
+C
,
x x
x y y x
y y
2
2
C = A
+ 2B
+C
.
x
x y
y
37
38
Chapter 8 Lecture 08
8.1.1
Result
We know that
2
B A C = (B 2 AC)
x y y x
2
,
where A, B, C are not zero at the same time. Thus the transformation of independent
variables does not change the type of equation. Since,
2
2 x
x
= > 0.
y y
x y y x
Therefore the sign of the discriminant does not change, either we are in (x, y) coordinates or in (, ) coordinates.
8.2
B B 2 AC
x
B B 2 AC
x
=
,
=
.
y
A
y
A
Note
The condition B 2 AC > 0 ensures that the slopes of the curves (x, y) = c1
and (x, y) = c2 are real.
This means that at any point (x, y) there exist two real directions given by
the two roots along which the given partial differential equation reduces to the
canonical form.
39
There are two solutions for each quadratic however we will consider only one
solution for each otherwise we will end up with the same two coordinates.
Thus the chosen solution are of the form
x
B + B 2 AC
=
,
y
A
B B 2 AC
x
=
.
y
A
These are called characteristic equations.
Along the curve (x, y) = c1 , we have
d = x dx + y dy = 0
or
y dy = x dx
This implies
dy
x
=
dx
y
Similarly, along the curve (x, y) = c2 , we have
d = x dx + y dy = 0
or
y dy = x dx
Hence
x
dy
=
dx
y
Substituting the values of (x /y ) and x /y in the following equations such that
!
B + B 2 AC
dy
x
=
=
,
dx
y
A
!
dy
x
B B 2 AC
=
=
.
dx
y
A
which reduces to
2u
u u
= (, , u,
,
).
40
8.2.1
Chapter 8 Lecture 08
Examples
Lecture 09
Review
Reduction of Second Order Partial Differential Equations in Canonical Form
Result
Canonical Form for Hyperbolic Equation
Examples
9.1
u u
x, y,
,
=0
x y
u u
x, y,
,
=0
where
2
2
A = A
+ 2B
+C
,
x
x y
y
B = A
+B
+
+C
,
x x
x y y x
y y
2
2
C = A
+ 2B
+C
.
x
x y
y
41
42
Chapter 9 Lecture 09
9.1.1
Result
We know that
2
B A C = (B AC)
x y y x
2
,
where A, B, C are not zero at the same time. Thus the transformation of independent
variables does not change the type of equation. Since,
2
2 x
= x
> 0.
y
x y y x
y
Therefore the sign of the discriminant does not change, either we are in (x, y) coordinates or in (, ) coordinates.
9.2
B B 2 AC
x
=
.
y
A
Since B 2 AC = 0, therefore
x
B
= .
y
A
This is the characteristic equation. Also this is a slope along the curve
(x, y) = c1 .
The total differential is
d = x dx + y dy = 0
y dy = x dx
dy
x
= .
dx
y
Or
dy
=
dx
B
A
43
=
B
.
A
which reduces either of the forms
2u
2
2u
2
9.2.1
= (, , u, u , u ),
= (, , u, u , u ).
Examples
44
Chapter 9 Lecture 09
10
Lecture 10
Review
Reduction of Second Order Partial Differential Equations in Canonical Form
Result
Canonical Form for Parabolic Equation
Examples
10.1
u u
x, y,
,
=0
x y
u u
x, y,
,
=0
where
2
2
A = A
+ 2B
+C
,
x
x y
y
B = A
+B
+
+C
,
x x
x y y x
y y
2
2
C = A
+ 2B
+C
.
x
x y
y
45
46
Chapter 10 Lecture 10
10.1.1
Result
B A C = (B AC)
x y y x
2
,
where A, B, C are not zero at the same time. Thus the transformation of independent
variables does not change the type of equation. Since,
2
2 x
= x
> 0.
y y
x y y x
Therefore the sign of the discriminant does not change, either we are in (x, y) coordinates or in (, ) coordinates.
10.2
B B 2 AC
dy
=
.
dx
A
Suppose (x, y) = c is the
solution of the characteristic curve which is not real,
2
because B AC < 0 B 2 AC is imaginary. Therefore (x, y) can be written
in this form
1 (x, y) + i2 (x, y) = c.
Assume that
1 (x, y) = (x, y),
2 (x, y) = (x, y).
Thus
(x, y) = (x, y) + i(x, y).
Since (x, y) is the solution of equation
dy
B B 2 AC
=
,
dx
A
which basically comes by putting A = 0. In other words (x, y) is the solution of
2
2
A
+ 2B
+C
= 0.
x
x y
y
47
which reduces to
10.2.1
2u 2u
u u
+ 2 = (, , u,
,
)
2
Example
Example
Example Reduce the following partial differential equations
2
2u
2 u
+
x
=0
x2
y 2
to a canonical form.
Example Reduce the following partial differential equations
(1 + x2 )
to a canonical form.
2u
2u
u
u
+ (1 + y 2 ) 2 + x
+y
=0
2
x
y
x
y
48
Chapter 10 Lecture 10
11
Lecture 11
Review
Reduction of Second Order Partial Differential Equations in Canonical Form
Result
Canonical Form for Elliptic Equation
Examples
11.1
u u
x, y,
,
=0
x y
u u
x, y,
,
=0
where
2
2
A = A
+ 2B
+C
,
x
x y
y
B = A
+B
+
+C
,
x x
x y y x
y y
2
2
C = A
+ 2B
+C
.
x
x y
y
49
50
Chapter 11 Lecture 11
11.1.1
Result
B A C = (B AC)
x y y x
2
,
where A, B, C are not zero at the same time. Thus the transformation of independent
variables does not change the type of equation. Since,
x y y x
2
2
> 0.
Therefore the sign of the discriminant does not change, either we are in (x, y) coordinates or in (, ) coordinates.
11.2
11.2.1
Example
x 6= 0,
Problem Sheet
Exercises
51
2z
2z
2z
+
(x
+
y)
+
x
=0
x2
xy
y 2
2z
2z
2 z y 2 z
x2 z
y 2 2 2xy
+ x2 2
= 0,
x
xy
y
x x
y y
2
1 z
2z
2 z
+
x
=0
2
2
x
y
x x
2z
2z
x2 2 = 0
2
x
y
2
z
2z
2z
+
2
+
= 0,
x2
xy y 2
2z
2z
2z
+ y 2 2 = 4x2
x2 2 + 2xy
x
xy
y
2
2
2
u
u
u
2 u
2
sin
x
cos
x
cos x
=0
2
2
x
xy
y
y
Exercises Find the region in the xy-plane in which the following equations are
hyperbolic and elliptic in nature
1)
2)
2u 2u
2 =0
x2
y
2
2
u
u
+ xy 2 = 0.
2
x
y
(1 x2 )
52
Chapter 11 Lecture 11
12
Lecture 12
Review
Reduction of Second Order Partial Differential Equations in Canonical Form
Canonical form for Hyperbolic Equations
Examples
Canonical form for Parabolic Equations
Examples
Canonical form for Elliptic Equations
Examples
Mixed Type Equations
Examples
12.1
Mathematical Modeling
54
Chapter 12 Lecture 12
12.1.1
Objective
In this step the real-life system to be modeled is defined. At this point it is often
convenient to strip the system to its essential features so that a prototype1 model
for it can be constructed.
12.1.2
Background
In this step the pertinent laws and data about the system must be examined. In
particular, if no data are available we must carry out proper experiments to obtain
this information. As a result we should be able to identify the important variables
that influence the evolution of the system and their relations.
12.1.3
12.1.4
Modeling
At this stage the mathematical relations that govern the behavior of the system are
derived.
Mathematical models can be classified in the following ways:
Linear versus Non-Linear
Discrete versus Continuous
Deterministic versus Probabilistic (Stochastic)
1
A prototype is a model which is not based on strict planning, but is an early approximation of
the final product. A prototype acts as a sample to test the process. From this sample we learn and
try to build a better final product.
55
12.1.5
Model Validation
Methods must be devised to solve the model equations and compare their predictions
with the actual data about the system. If large unaccountable deviations between
the model predictions and data are detected, the model must be reexamined and
modified accordingly.
56
12.1.6
Chapter 12 Lecture 12
Compounding
At this stage the prototype model is modified to take into account some aspects of
the system that were neglected earlier in order to simplify the modeling process.
12.2
57
Kirchoff s Second Law The algebraic sum of the changes in potential around
any closed path of a circuit must be zero.
12.3
Objective: Derive a model equation for the traffic flow on a highway without
exists and with one entrance and one lane.
Discussion: One possible approach to modeling the traffic flow is to describe
each care as a finite element on the highway and then write a discrete model,
which describes the motion of each such car. However, if there are many cars
on the highway, this approach is not practical and it is better to construct a
continuous model, which treats those cars as smeared out quantities.
Approximations and Idealizations:
We assume that the highway is infinite
We define the car density (x, t) as
(x, t)
=
58
Chapter 12 Lecture 12
Z
a
(x, t)
dx
t
The rate of change must equals the flux of cars entering at a less the flux of
car leaving at b. Therefore
Z a
dn
q(x, t)
= q(a, t) q(b, t) =
dx
dt
x
b
The we infer from above equations that
Z b
q(x, t)
+
dx = 0.
t
x
a
And since a, b are arbitrary it follows that
q(x, t)
+
= 0.
t
x
Modeling:
Since we know that the flux q = u so the above equation becomes
( u)
+
=0
t
x
This equation is the equation of continuity in one dimension.
Note Continuity equation contains two unknown quantities and u. Therefore
to solve it we must either be able to express u = u() or find an additional equations
that relates these two quantities.
13
Lecture 13
Review
Mathematical Modeling
Objective
Background
Approximations and Idealizations
Modeling
Model Validation
Compounding
Review of Physics Laws
The Continuity Equation
13.1
60
Chapter 13 Lecture 13
lim
S0, t0
61
where u(x, y, z, t) is the temperature at (x, y, z), at time t and k is the thermal
conductivity of the material.
Note Remember that the gradient of a function increases more rapidly while in
the direction opposite to it the function decreases most rapidly. Thus, a restatement
of the Fourier law is that heat flows in the direction in which the temperature
decreases most rapidly. This is the reason of minus sign in the above equation
Principle of Energy Conservation The total amount of energy in an isolated
system remains constant over time.
Approximations and Idealizations:
We assume that rod is homogeneous. It follows that c, and are independent of the position x. Also, we further assume for a prototype model
that c, and are independent of the temperature u.
The length of the rod remains constant in spite of the changes in its
temperature.
We also assume that the rod is perfectly insulated along its lateral surface
(Idealization). Hence, heat can flow only in the horizontal direction, since
a vertical flow will lead to heat accumulation along the edges, which is
forbidden by the Fourier law of conduction. Therefore we infer that the
temperature on a vertical cross section of the rod must be the same.
Thus, the temperature u depends only on x and t; that is u = u(x, t)
We assume that the heat flows in the rod from left to right, which requires
the left side to be warmer than the right side.
Modeling: We consider an infinitesimal element of the rod between x and
x + x and write the equation for energy conservation in it.
V = volume of the element = A x
where A is the cross sectional area of the element.
m = mass of the element = A x = V.
Also the amount of the heat at time t is
Q(x, t, x) = c m u(x, t).
From above two equation, we obtain
Q(x, t, x) = c Ax u(x, t).
The rate of change in heat is
dQ
u(x, t)
= c Ax
.
dt
t
62
Chapter 13 Lecture 13
Modeling:
Heat flowing in = q(x, t) A
Heat flowing out = q(x + x, t) A
By Principle of Heat Conservation The rate of change must equal the rate at
which heat is flowing in less the rate at which it is flowing out. Hence
dQ
dt
= q(x, t) A q(x + x, t) A
= A[q(x, t) q(x + x, t)]
dQ
dt
= c Ax
u(x, t)
t
u(x, t)
c
t
c Ax
u(x,t)
t
u(x, t)
t
q(x, t) q(x + x, t)
x0
x
q
= .
x
=
lim
2 u(x, t)
x2
1 u(x, t)
k t
=
=
k=
Example Generalize the prototype model for the case when heat is generated
in the rod at a rate of r(x, t) per unit volume.
13.2
63
Initial Conditions
Since one dimensional heat equation is first order in t, it needs only one initial
condition, which is normally taken to be
u(x, 0) = f (x),
0 < x < L.
13.3
Boundary Conditions
The equation is of second order in with respect to the space variable, so we need
two boundary conditions. There are three main types of such conditions prescribed
at the end points x = 0 and x = L which has physical significance.
The temperature may be at one end point; for example,
u(0, t) = (t),
t > 0.
If the rod is insulated at an end point then the heat flux at that end must
be zero. This is equivalent to the derivative ux = u
x being equal to zero; for
example,
u(L, t)
= 0,
t > 0.
x
More generally, if the flow of heat at each end-point to be (t) then the above
condition becomes
u(L, t)
= (t),
t > 0.
x
When one of the end points is in contact with another medium, we use Newtons
Law of cooling, which states that
heat flux at the endpoint is proportional to the difference of the rod and the
temperature of the external medium,
For example
ux (0, t) = H[u(0, t) U (t)],
t > 0,
where U (t) is the known temperature of the external medium and H > 0 is the heat
transfer coefficient. Owing to the convention concerning the direction of heat flux,
at the other endpoint this type of condition becomes
ux (0, t) = H[u(0, t) U (t)],
Remark
t > 0.
64
Chapter 13 Lecture 13
,t > 0
ux (0, t) = ux (L, t)
13.4
, t > 0.
Definition 13.1. A partial differential equation associated with the initial and
boundary conditions is known as the initial boundary value problem.
If only initial conditions or boundary conditions are present, then we have an
initial value problem or a boundary value problem respectively.
13.4.1
Examples
Example The initial boundary value problem modeling heat conduction in a one
dimensional uniform rod with sources, insulated lateral surface, and temperature
prescribed at both end points is of the form
ut (x, t) = kuxx (x, t) + Q(x, t) ,
| {z }
0 < x < L,
Source Term
t > 0,
u(L, t) = (t),
t > 0,
0 < x < L.
65
0 < x < L,
t > 0,
t > 0,
0 < x < L,
13.4.2
Solution
66
Chapter 13 Lecture 13