Advanced Engineering Analysis - Notes

Ramesh Kadambi
November 14, 2009
2
Contents
1 Ordinary Dierential Equations 5
1.1 ODE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.2 Principle of Superposition (Linearity Test) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.3 General Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.4 Generalized Linear ODE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.5 Dierence Linear ODE and Non Linear DE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.6 Linear Dependence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.7 Establishing Linear Dependence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.8 Linear Equation Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.8.1 Homogeneous Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.9 First Order ODE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.10 LDE of n
th
order with constant coecients . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.11 Case of Repeated Roots . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.12 Complex Roots . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.13 Method of Undetermined Coecients MOUC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.14 Method of Variation Of Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.15 Euler Equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
3
4 CONTENTS
Chapter 1
Ordinary Dierential Equations
1.1 ODE
denition: An ODE is a statement of functional dependence.
1.2 Principle of Superposition (Lin-
earity Test)
f(ax
1
+ bx
2
) = af(x
1
) + bf(x
2
)
1.3 General Solution
The general solution to an ODE is the set of all functional re-
lations that satisfy the ODE.
ex:
d
2
y
dx
2
+ y = 0 solution y = c
1
sin(x) + c
2
cos(x)
1.4 Generalized Linear ODE
a
0
(x)
d
n
y
dx
n
+ a
1
(x)
d
n1
y
dx
n1
+ .... + a
n
y = f(x)
1.5 Dierence Linear ODE and Non
Linear DE
A nonlinear DE has nonlinear function of the depended vari-
able.
ex:
y
d
2
y
dx
2
+ e
x
y
2
dy
dx
= y
1.6 Linear Dependence
A linear combination of functions are linearly independent over
an interval x [a, b] if over the interval no nontrivial linear
combination of the functions is identically 0. This basically
means that none of the functions of the set can be expressed in
terms of the other functions in the set. In mathematical terms,
c
i
such that
n

i=1
c
i
u
i
(x) = 0, unless c
i
= 0, x [a, b]
1.7 Establishing Linear Dependence
Consider u
i
(x) dened in the interval [a, b] such that u
i
(x)
have n nite derivatives in the interval [a, b]. If c
i
such that
n

i=1
c
i
u
i
(x) = 0, then we just dierentiate n times and since each
of the equations so obtained must also sum to zero, we have the
following,
n

i=1
c
i
u
i
(x) = 0
n

i=1
c
i
u
1
i
(x) = 0
n

i=1
c
i
u
2
i
(x) = 0


n

i=1
c
i
u
n1
i
= 0

u
1
u
2
u
3
u
n
u
1
1
u
1
2
u
1
3
u
1
n
u
2
1
u
2
2
u
2
3
u
2
n

u
n1
1
u
n1
2
u
n1
3
u
n1
n

c
1
c
2
c
3

c
n

= 0
For a non trivial solution we have,
W(u
1
, ..., u
n
)(x) =

u
1
u
2
u
3
u
n
u
1
1
u
1
2
u
1
3
u
1
n
u
2
1
u
2
2
u
2
3
u
2
n

u
n1
1
u
n1
2
u
n1
3
u
n1
n

= 0
for some x [a, b]
5
6 CHAPTER 1. ORDINARY DIFFERENTIAL EQUATIONS
1.8 Linear Equation Solution
The standard from ODE is given as
d
n
f(x)
dx
n
+ a
1
(x)
d
n1
f(x)
dx
n1
+ + a
n1
dy
dx
+ a
n
(x)f(x) = h(x)
. The expression
d
n
dx
n
+ a
1
(x)
d
n1
dx
n1
+ + a
n1
(x)
d
dx
+ a
n
(x)
is called the linear dierential operator. So the equation can be
rewritten as
Lf(x) = h(x)
. One way to solve the ODE is to integrate things through n
times. Each integration step will leave you with a constant c
i
,
leaving you with n constants. Every possible solution can be
obtained by specifying values to these constants. Another point
to note is that the solutions dier from each other only by a
constant.
1.8.1 Homogeneous Solution
The solution of the ODE is broken into two parts. The rst sol-
tuion is called the homogeneous solution f
h
(x) solves the DE
Lf
h
(x) = 0. The second part is called the particular solution
which will solve the DE Lf
p
(x) = h(x). The complete solution
is then given by
f(x) = f
h
(x) + f
p
(x)
.
1.9 First Order ODE
The rst order ODE is obtained when n = 1. So the general
FODE is given by,
df(x)
dx
+ a
1
(x)f(x) = h(x) (1.9.1)
. The solution to this is obtained by solving two parts.
df
h
(x)
dx
+ a
1
(x)f
h
(x) = 0
df
h
(x)
dx
=a
1
(x)f
h
(x)
1
f
h
(x)
df
h
(x) =a
1
(x)dx
integrating both sides we get
log(f
h
(x)) =

a
1
(x)dx + C
f
h
(x) = e

R
a1(x)dx+C
= e

R
a1(x)dx
+ C
In order to obtain the particular solution we do the following,
we are looking for a function P(x) such that,
dP(x)f
p
(x)
dx
= P(x)h(x) (1.9.2)
. In order to do that we rst multiply our original ODE by
P(x) We have,
P(x)
df(x)
dx
+ P(x)a
1
(x)f(x) = P(x)h(x) (1.9.3)
This method is called the method of integrating factor. An
integrating factor is a function that is chosen to facilitate the
solving of a given dierential. The integrating factor converts
an inexact dierential to an exact dierential. This nds appli-
cations in multivariate calculus. The interesting aspect of this
method is that it can be used to solve some types of non-linear
dierential equations
1
.
We now obtain the dierential of
d
dx
P(x)f
h
(x)
dP(x)f
p
(x)
dx
= P(x)
df
p
(x)
dx
+ f
p
(x)
dP(x)
dx
(1.9.4)
comparing the original ODE (1.9.3) with (1.9.4) we see that,
dP(x)
dx
= P(x)a
1
(x)
P(x) = e
R
a1(x)dx+C
= e
R
a1(x)dx
+ C
Now integrating (1.9.2) we obtain,
P(x)f
p
(x) =

P(x)h(x)dx + C
f
p
(x) =
1
P(x)

P(x)h(x)dx + C

(1.9.5)
Example:
1. x
dy
dx
ky = x
2
writing in cannonical form
dy
dx

k
x
y = x
we haveP(x) = e
R
a(x)dx
where a(x) =
k
x
P(x) = e

R
k
x
dx+C
= C
1
e
k lnx
= C
1
x
k
y
p
(x) =
1
C
1
x
k

C
1
x
k
xdx +
C
2
C
1
x
k
= x
k
x
k+2
k + 2
+ Cx
k
if k != 2
=
x
2
2 k
+ Cx
k
k != 2
if k = 2 then we go back to our original problem and solve for
that specic case.
for k = 2 we have
y
p
(x) = x
2

1
x
dx + Cx
2
y
p
(x) = x
2
lnx + Cx
2
1
Wikipedia url:http://en.wikipedia.org/wiki/Integrating factor
1.10. LDE OF N
TH
ORDER WITH CONSTANT COEFFICIENTS 7
1.10 LDE of n
th
order with constant
coecients
Consider the dierential operator L given by,
L =
d
n
dx
n
+ a
1
d
n1
dx
n1
+ + a
n1
d
dx
+ a
n
In this section we will look into solving ODE of the form,
Ly = h(x)
, Let us assume a solution of the form y = e
rx
where r is a
constant. We would like to solve the homogeneous equation
Ly = 0. Dierentiating the sulution of the form y = e
rx
and
substituting we have,
(r
n
+ a
1
r
n1
+ + a
n1
r + a
n
)e
rx
= 0
since e
rx
= 0 we have the following polynomial in terms of r
equal to zero,
r
n
+ a
1
r
n1
+ + a
n1
r + a
n
= 0
Let r
1
, r
2
, , r
n
be the roots of the polynomial. The solution
y
h
is now of the form,
y
h
= c
1
e
r1x
+ c
2
e
r2x
+ + c
n
e
rnx
=
n

i=1
c
i
e
rix
1.11 Case of Repeated Roots
Suppose some of the roots are repeated, i.e. r = r
1
= r
2
are
roots then we have,
Le
rx
= r
n
+ a
1
r
n1
+ a
2
r
n2
+ a
n
e
rx
= (r r
1
)(r r
1
)(r r
3
) (r r
n
)e
rx
From the above equations we can see that

r
[Le
rx
]
r=r1
= 0
We no switch the dierential operator (I dont know the tech-
nical restrictions to do so but will do it anyway).
L[
e
rx
r
] = 0
L[xe
rx
]
r=r1
= 0
This implies that xe
r1x
is a sulution to the ODE as well. It
follows from the above argument that if there a root repeats n
times then xe
r1x
, x
2
e
r1x
, ...., x
n
e
r1x
are all solutions.
1.12 Complex Roots
Since the coecients of the Characteristic Equation (CE) are
all real, the complex roots will be in complex conjugate pairs.
The solution is given by Ae
r1x
+ Be
r2x
where r
1
= a + ib and
r
2
= a ib. The solution can then be written as
y
H
= e
ax
[(A + B) cos bx + (A + B) sinbx]
.
Example 1:
d
4
y
dx
4
k
4
y = 0
r
4
k
4
= 0
The roots are r
1
= k, r
2
= k, r
3
= ik, r
4
= ik. The solution
y = c
1
e
kx
+ c
2
e
kx
+ c
3
cos kx + c
4
sinkx
Example 2: y

+y = 0 The CE r
3
r
2
r +1 = 0
The CE simplies to (r 1)(r
2
1) = 0 We there fore have
(r 1)
2
(r + 1) = 0. The solution is therefore give by:
y = c
1
e
x
+ c
2
xe
x
c
3
e
x
1.13 Method of Undetermined Coef-
cients MOUC
The objective is to nd the particular solution. This method
works only on systems with constant coecients. The RHS
has have certain kinds of functions in order for this method to
work. The idea is to guess the solution to be of the same form
as the right hand of the inhomogeneous equation.
Homogenious Soln Solution Form
kx
m
, m = 0, 1, 2, k
n
x
n
+ + k
n1
x
n1
ke
ax
Ce
ax
k cos ax or k sinax Acos ax + B sinax
ke
ax
sinbx or ke
ax
cos bx e
ax
[Acos ax + B cos bx]
(

n
1
k
i
x
i
) cos ax or (

n
1
k
i
x
i
) sinax e
ax
[cos ax

n
1
A
i
x
i
+sinax

n
1
B
i
x
i
]
8 CHAPTER 1. ORDINARY DIFFERENTIAL EQUATIONS
Example 1: y

+ y = sinx
y = y
p
+ y
h
we rst nd the homogeneous solution
y

+ y = 0
d
2
+ d = 0
d = i
y
h
= Acos x + B sinx
using the method of undetermined coecients and the fact that
the homogeneous solution is not linearly independent of the RHS.
y
p
= x[K
1
cos x + K
2
sinx]
substituting the particular solution into the DE
x[K
1
cos x K
2
sinx] + x[K
1
cos x + K
2
sinx]
+ 2K
2
cos x 2K
1
sinx = sinx
2K
2
cos x 2K
1
sinx = sinx
K
1
=
1
2
The complete solution
y = Acos x + B sinx
x
2
cos x
1.14 Method of Variation Of Param-
eters
1.15 Euler Equation
The Euler Equidimensional Linear D.E is given below., Ly =
[x
n d
n
dx
n
+ b
1
x
n1
+ b
n
]y = h(x) where, b
i
are constants. We
will solve this problem by using a change of variables. We will
use the substitution x = e
z
, i.ez = log x. We now see that the
Euler equation gets converted to a constant coecient ODE.
dy
dx
=
dy
dz
dz
dx
=
1
x
dy
dz
x
dy
dx
=
dy
dz
d
2
y
dx
2
=
1
x
d
dz
dy
dx
=
1
x
d
dz

1
x
dy
dz

=
1
x
2
dy
dz
+
1
x
2
d
2
y
dz
2
x
2
d
2
y
dx
2
=
d
dz

d
dz
1