mod10lec46

IIT Roorkee July 201
61 followers
Dec 5, 2023
Hello friends welcome to my lecture on review of integral transforms. This is our first
lecture on this topic. Let us review the 2 most powerful methods to solve the differential
equations. The methods are Fourier transforms and Laplace transforms. These transforms allow
us to turn a complicated problem into a simpler one. The idea behind these transforms is very
simple. Suppose we want to solve a differential equation with the unknown function f.
Then we apply the transform to the differential equation to turn it into an equation which
we can solve easily. This equation is an algebraic equation for the transform of F of f.
We solve this equation of F and then apply the inverse transform to find f diagrammatically
this idea can be represented as
See the diagram; we are given this differential equation if you want to find the solution
of the differential equation directly, then most of the times its difficult, so we apply
this transform technique. We apply the transform to the differential equation for the function
F and obtain an algebraic equation. This algebraic equation is solved to get the function F and
the function F is then inverted to get the solution of the differential equation that
is the function F.
This is a simple idea. Let us see what is the use of Fourier and Laplace transforms
and where they are used.
The Fourier transforms is a tool to find the steady state solutions to differential equation.
These solutions may be thought of as the response of a system which has been driven for such
a long time that any transient solutions have died out. But in many systems, we also need
the transient solutions then the Laplace transforms is used. Further this transform takes into
account the initial conditions in a natural way. Let us review the Laplace transform.
Let us first define a Laplace transform let ft be a function defined for all t greater
or equal 0, then the Laplace transform of f is a function of s, say Fs defined as Fs
equal 0 to infinity e to the power minus st .ft dt.
Provided the integral on the right side of equation 1 exists. This might restrict the
values of s for which the transform exists. Existence theorem; let us see what are the
sufficient conditions on the function F so that its Laplace transform exist. Let Ft be
a function which is piecewise continuous on every finite interval in the range t greater
or equal 0 and satisfies mod of ft less or euqal M times e to the power alpha t for all
t gretater or equal 0 and for some constants alpha and M.
Then the Laplace transform of ft exists for all real s greater alpha. s could be a complex
number. So this is the existence theorem for Laplace transform. The conditions in the above
theorem are sufficient rather than necessary, and it is easy to check whether a given function
satisfies n in equality of the form 2. The function f is in exponential order or not.
For example, consider, cos hyperbolic t equal you can see cos ft equal cos hyperbolic t
equal e to the power t plus e to the power minus t by 2. We know that e to the power
minus t is less equal e to the power t if t is greater equal 0. If t equal 0 then Ft
equal 1 plus 1 by 2 that is 2 by 2 that is 1. Otherwise if t is gretater 0, then e to
the power minus t is strictly less e to the power t if t is greater 0. So, e to the power
t plus e to the power minus t by 2 is less e to the power t plus e to the power t by
e which means that we have, 2 e to the power t by 2 which is equal e to the power t.
So, cos hyperbolic t is less w e to the power t for all t better than 0. Similarly e to
the power t we know that it is sigma n equal 0 to infinity t to the power n by n factorial.
If t is positive, then t to the power n by n factorial will always be less e to the power
t. so cos hyperbolic t is less euqal t, t to power n is less n factorial e to power
t for all t greater their both functions ft equal cos hyperbolic t and ft equal t to the
power n satisfy the condition mod of ft is less equal m times e to the power alpha t.
If you take any bounded function, mod of ft less equal some constant k for all t greater
0, then it obviously satisfies the conditions mod of ft less equal m times e to the power
alpha t. For example, you can take ft equal sin t and cos t they are bounded by 1. So,
all bounded functions satisfy the condition of existence given in the existence theorem.
If ft equal e to the power t square then we see that the condition 2 is not satisfied.
Because however large M and Alpha are chosen, e to the power m square will be gretaer M
times e to the power Alpha t, for all t greater t0, where t0 is a sufficiently large number
and of course will be depending on M and Alpha.
Now note that the conditions in the existence theorem are sufficient rather than necessary.
The conditions given in the theorem are sufficient rather than necessary because we have functions
for which Laplace transform adjust even without their satisfying conditions of this theorem.
For example Ft equal 1 by root t √at t equal 0 it assumes the value infinity but Laplace
transform exists. Laplace transforms of t to the power minus 1 by 2.
We can find as 0 to infinity e to the power minus st and t to the power minus by dt. You
take st equal u. so sdt equal du. Then if s is positive the limits of integration will
remain t0 and infinity. And we will have e to the power minus u. t will be us to the
power minus 1 by 2 and dt will be dus. So, what we will get is 0 to infinity equal 1
by √roots 0 to infinity e to the power u and u to the power minus 1 by 2 du. This is
1/root s* gamma 1 by 2 equal root pi by root s where s is positive.
So Laplace transform of ft equal 1 by root t exist but it is not piece wise continuous
in a finite interval when t is greater equal 0. When we take finite interval say 0 to some
number a so from this interval it is not piece wise continuous. Now further if the Laplace
transform exist, it is unique conversely if 2 functions have the same Laplace transforms
then they cannot differ over an interval of positive length, although they may differ
at various isolated points.
Practically it is not significant. Hence we shall assume that the inverse of a given transform
is essentially unique.
The Laplace transform satisfy the linearity property if f and g are any 2 functions, where
Laplace transform exists then Laplace transform of aft plus bgt equal aLft plus bLgt where
a and b are constants. Laplace transforms of derivatives suppose f is a function whose
Laplace transform exist and f prime t is piece wise continuous and every finite interval
in the range t greater t equal 0 then Laplace transform of ft is given s/ Laplace transform
of the function ft minus f 0.
If we denote Laplace transform as ft/fs then we get sfs minus f0. And the result iterating
we can get the Laplace transform of f double dash t n higher order derivatives. Suppose
if we want to find out Laplace transform of f double dash t then we apply this formula
of f dash in place of f. So, then L f double dash twill be equal to s times L f dash t
minus dash 0 replacing f by f dash. Now f times Laplace transform of f dash t equal
SLfs Lft minus f0 minus f dash 0.
So, what we will have s square Lft minus sf0 minus f dash 0. So, we will get Laplace transform
of f double dash ts s square fs minus sf0 minus f0. We can again use the iteration over
this n can be Laplace transform f triple dash t and so on. So, Laplace transform of the
n rh derivative of f I sn to the power fs minus sn minus 1 f0 minuss n minus 2f dash
2 and so on fn minus 10.
Now Laplace transform of the integral of a function, suppose Laplace transform of Ft
is Fs, then Laplace transform 0 to t udu is Fs s. So, when we take the integral of a function
then it is Laplace transform gives the Laplace transform of the function F by s. Now, let
us study 2 important properties of the Laplace transform concerning the shifting on s-axis
and shifting on t-axis.
In the first shifting theorem if the Laplace transform of Ft is Fs for s>alpha then, the
Laplace transform of e to the power at . ft is Fs-a means there is a shifting on the x
axis where s greater a+ plus alpha. So, now placing s by s minus a in the transform Fs
corresponds to Ft by e to power at, In the second shifting theorem if L of FT into Fs
then e to the power minus F(s) equal Laplace Transform of F to T where a greater 0 after
that function is defined as a 0.
If t less than a then f(t minus a) if t greater a. So this can be alternately written like
this
You need a step function L of Ft minus at equal e to the power as FS if you use the
unit step function uat is defined as 0 if t less a and 1 if t greater a, so u at into
Ft minus a will be 0 if t less a and ft minus a then t greater a then Ft minus ma into uat
is F t.
Now, Differentiation of Laplace transform: If L of ft equal Fs, then let us see what
happen when we differentiate the function Fs which will be the Laplace transform of
t function the differentiation of Fs with respect to S corresponds to multiplication
of ft by minus t. So, L of tft equal minus d over ds of fs and when you use mathematical
induction you get L of t to the power ft equal minus 1 to the power n Fns where Fns is the
nth derivative with respect to s.
Integration of Laplace Transform if Lft equal Fs let us see what happens in integrate the
function Fs with respect to S. So when you integrate Fs the function ft by t that is
the integration of Laplace transform Fs of a function f corresponds to the division of
the function ft by t. The more property which we have to consider in the Convolution. Another
important property of the Laplace Transform is concerned with the product of transforms.
Quite often, we wish to determine L inverse of Fs into Gs where inverse Laplace transform
of Fs is ft and inverse Laplace transform of Gs is Gt. So, we know the Laplace transform
of Ft and Gt we want to determine the inverse of the product of Fs and Gs.
It turns out that if ht is the inverse of Laplace inverse of Fs into Gs and ht it will
be the convolution of F and G. This convolution of F and G is defined as integral over 0 to
t F tau Gt minus tau d tau and inverse of Fs into Gs is the convolution of F and G that
will be call it as Convolution theorem, now let us look at periodic function. These Functions
occur in many physical problems and mostly they are more complicated than single sin
or cosine function.
Let ft be a piecewise continuous function for all t greater 0 and ft plus p equal ft,
for all t greater 0.Then Lft equal 0, 1 over 1 minus integral 0 to p e to the power minus
st.ftdt where s is greater 0.
Let us consider the mathematical model of forced oscillations of a body of mass m attached
at the lower end of an elastic spring whose upper end is fixed. Let k be the spring modulus
and k0 sin pt be the driving force or input. Then, we have the initial value problem and
my double dash ky equal k0 sin pt, y0 equal 0 y dash equal 0.Let us solve this equation
by using Laplace transform. So, we can write it as, y double dash plus k by m y equal k0
by m sin pt.
Now, let us de0e k by m equal omega 0 square so then y double dash plus omega 0 square
y equal K sin pt. Let us take the Laplace transform L y double dash because Laplace
is a linear operation. So, omega 0 square Laplace transform of yt and then k times Laplace
transform of sin pt and Laplace transform of y double dash will be s square by s minus
s by 0 minus dash 0 plus omega 0 square by s ys is the Laplace transform of yt equal
K times.
Laplace transform of sin pt is p upon s square plus Square now we are given y equal 0 and
y dash equal 0. So, using the initial conditions s square plus Omega 0 Square k times p upon
s square plus p square or we get ysm equal k times p upon s square plus p square s Square
plus omega 0 square where omega 0 equal root kym and capital K equal k0ym We can then find
the inverse Laplace transform. So, what we will do is we can break it into partial sections.
And if k times p then upon s square plus omega 0 Square then upon s square plus p square
by p square minus omega 0 square. ys can be written as kp upon p square minus omega 0
square so omega 0 square plus b square. Now take inverse Laplace transform L inverse of
ys equal kp upon p square minus omega 0 square L inverse of 1 upon s square plus omega 0
square equal L inverse of 1 upon s square plus p square kp upon p square minus mega
0 square plus omega 0 square is the inverse Laplace transform.
The inverse Laplace transform of s square plus omega 0 square is sin omega t by omega
0 T and omega 0 and this inverse Laplace transform 1 over s square plus p square is sin pt by
p. Here we assume that p s is not equal omega 0 square.
Here we get the function yt by t is the inverse Laplace transform of ys, this is yt. So, with
the function yt equal k over omega 0 square p over omega 0 square omega 0 minus sin omega
0 t into pt, p square is 0 equal omega 0 square. Which is the case of no resonance the equation
5 represents a superposition of 2 harmonic oscillations sin omega 0t sin pt whose frequencies
are the natural frequency of the freely vibrating system.
And the frequency of the driving force given by sin pt. So, in the case of resonance, we
have P equal omega 0 and when p equal omega 0 t by P equal omega 0 but we have here p
equal omega 0 ys will become kp upon S Square plus omega 0 S square whole square we get
ys equal k omega 0 s square s square plus omega 0 square. And when we take the inverse
Laplace transform of this we will get yt equal K by omega 0 square sin omega 0 t minus omega
0 t minus cos omega 0 t.
Let us see how we get this so L inverse of ysm equal k omega 0 is a constant. We can
write like this L inverse of then overs square plus omega 0 square. Now here whole square
Let us remember that let us call that L inverse of 1 over s square plus omega 0 square equal
sin omega 0t. Is equal sin omega 0 t by omega 0. Okay, so when we differentiate, suppose
we know Laplace transform of ft equal fs, then differentiation of fs with respect to
s, gives the multiplication of ft by minus t. t.
So, what we have Laplace transform of minus t sin omega 0 t by omega 0. When you differentiate,
whenever s is square and omega 0 square, the sin omega 0 t omega 0 will be multiplied by
minus , so d by ds upon of 1 by s square plus omega 0 square. And this will be minus 1 upon
s square plus omega 0 square the whole square 2s. So, what we get Laplace transform of t
sin omega 0 t diveded by omega 0 equal 2s upon s square plus omega 0 square the whole
square.
Now we want the inverse Laplace transform of 1 upon s square plus omega 0 square the
whole square. So, we have to use the formula which divides the Laplace transform fs by
s and that is the Laplace if ft equal fs then Laplace transform of 0 to t is fudu is fs
by s. So, this means that l inverse of 1 by s square plus omega 0 square the whole square.
This means we are dividing this by s okay. So, this will be integral by 0 to t u sin
omega 0 u/2 omega 0 du.
We are bringing this 2 here, so this becomes t sin omega 0 t by 2 omega 0 and then the
right side becomes s by s square plus omega 0 square the whole square. We divide this
by s and that means that we are integrating this 0 to t, okay. So, this can be either
integrated 1 by 2 omega 0, integral of this is u sin omega 0 u. so u times minus cos omega
0 u by omega 0 minus 0 to t, derivative of u is 1. So, minus cos omega 0 u by omega 0
du and here we put 0 to t.
So, when we put t here, we get what 1 by 2 omega 0 minus t cos omega 0 t by omega 0.
When we put 0 it becomes 0, then here we get plus 0 to t . So,0 to t cos omega 0 u y omega
0, so 1 omega 0 and cos omega 0 u. When you integrate you get sin omega 0 u by omega 0
0 to t. So, what we get 1 by omega 0 minus cos omega 0 t by omega 0 and this becomes
1 by omega 0 sin omega 0 square sin omega 0 t , okay 0 it becomes u becomes 0. So, we
get 1 by 2 omega 0 s, so what do we get the upper omega 0.
You can take outside, so 1 by 2 omega 0 square you can take omega 0 square outside. So what
you get there, you get 1 by 2 omega 0 q. Okay, 2 omega 0 q. what you get there, minus omega
0 t cos omega 0 t plus sin omega 0 t that is the inverse Laplace transform of 1 by s
square plus omega 0 square and this we multiply by p omega 0. So, we get k upon 2 omega 0
square sin omega 0 t minus omega 0 t cos omega 0 t.
When the input and the natural frequencies are same, we have resonance which is of basic
importance in the study of vibrating systems.
The inverse Laplace transform the Laplace transform can be applied to a wide range of
initial values problems. The main difficulty lies in taking the inverse Laplace transform.
Usually we refer to the table of Laplace transforms but if this does not work then we use inversion
formula for the Laplace transform. So, if Lft equal fs, then ft can be obtained from
the formula, ft is equal 1 by 2 pi I integral by a minus sin infinity to a plus sin infinity
fs e to the power of st ds.
To obtain this function ft, the integral is performed along a line parallel to the imaginary
axis in the complex plane such that all the singularities of fs lie to its left.
So, you can see this figure. This is the line AB parallel to the imaginary axis in the complex
plane and this line is drawn in such a way that, all this similarities of fs lie to its
left. Now let us consider the contour c, the contour c includes the line AB and the semicircle
c dash, this BCDA the semicircle c dash. So, integral byAB equal integral by C, e to the
power st fsds minus integral by c dash e to the power st fsds.
Now we have a theorem, this theorem says that if there exists positive constants k and K
such that mod of fs is less than k times r to the power minus k for every point on c
dash. So this is r, the circle is drawn with center at this point with radius r so when
this r goes to infinity. Mod of fs must be less constant times r to the power minus k
for every point on c dash then integral by c dash e to the power st fsds goes to 0 as
r goes to infinity. So, this theorem we shall use in this result, okay.
So, what we have here, when we take the limit r tends to infinity here in this formula,
we get limit r tends to infinity integral by AB, e to the power st fsds is equal integral
limit r tends to infinity integral by a minus ir because this is A is a minus ir B is a
plus ir. So, integral a minus ir to a plus ir e to the power st fsds and this equal this
integral 2 pi i by residue theorem 2 pi I into some of residues at the poses of fs,
and this goes to 0 as r goes to infinity so we have this formula.
Now r we can say ft is equal 1 by 2 pi i. Now ft is 1 by 2 pi i is integral a minus
i infinity e power fsds becomes some of the residues of e to the power st fsds. Sum of
e to the power st fsds the poles of fs.
1 by 2 pi i. 1 by 2 pi i integral by a minus i infinity a plus i infinity, e to the power
st into fsds will be then equal some of residues of e to the power st into fs, the poles of
fs.
Now let us take for example fs equal s plus 1/s s minus 1 the whole square. Fs has poles
that s equal 0 which is simple pole and s is equal 1 pole of order 2. So, let us find
the residue at residue of fs at s is equal residue of e to the power st fsds. So residue
of st fs at s equal 0 so limit s tends to 0, because it is a simple pole s minus 0 fs
e to the power st which is equal limit s tends to 0 s times fs is s plus 1 by s s minus 1
whole square e to the power st.
So when s goes to 0 e to the power 0 is 1 and we get here 1 and here we get minus 1
whole square equal 1. Now residue at s equal 1 which is the pole of order 2 now residue
at a pole of order 1 has the following formula.1 by m minus 1 factorial residue at pole of
order z equal z 0 of order m. We have limit z tends to z 0 at dm minus 1 by dm minus 1
z minus z 0 to the power m fz. If f z has a pole of order m at z equal z 0 then we apply
this formula.
Residue at a pole z equal z 0 of fz of order m. okay. So by this formula here a pole of
order 2minus 2 factorial limit s tends to 1 this is s here d by s. okay s minus 1 whole
square e to the power st fs. So s plus1 by s s minus 1 whole square. This cancels with
this and by 1 factor release 1 limit s tends to 1 d by ds of e to the power st 1 plus 1
by s. So this is limit s tends to 1 derivative of e to the power of st t 1plus 1 by s and
then e to the power st minus 1 by s square.
When s goes to 1 what we get here. So e to the power t and then we get t here 1 plus
1 is 2 and here we get e to the power t and then minus 1. So we get e to the power t 2t
minus 1. And therefore sum of residues is 1 plus e to the power t times 2t minus 1.
So inverse Laplace transform of fs okay and inverse of fs equal ft and ft equal 1 plus
e to the power of t 2t minus 1. This is the inverse Laplace transform of s plus 1 by s
s minus 1 whole square.
Now let us determine L inverse e to the power minus b root s by s where b greater 0. So,
let us take Fs equal e to the power minus b root s by s then L inverse e to the power
minus b root s by s by inverse formula is 1 by 2 pi i integral by a minus i infinity
to a plus i infinity e to the power st e to the power minus b root s by s ds. Now the
integrand has a branch point at the origin. So we take a contour ABCDEF, which encloses
no singularity. Let us see the contour.
This contour we will take at z equal 0 we have trans point that includes the similarity
and the contour is this. A B C D E F A. We take this contour and then we write the integral/this
contour into parts. Integral by AB BC CD DE EF FA e to the power st e to the power minus
b root s by s ds equal 0. Now what do we notice. Let us notice that this Fs okay is e to the
power st e to the power minus b root s by s. Let us apply this theory which we have
earlier stated.
If there exists positive constants k and k such that modem F s less K times r to the
power minus k. for every point on c dash then this integral goes to 0. So what to happens
is that this integral okay the integral BC and FA goes to 0 by that theory.
Because mod of Fs less some constant by S in some constant by R.
And that goes to 0 and R goes to infinity integral by BC and FA goes to 0 and if you
take OC equal rho and OD equal epsilon then 6 implies that integral by a minus ir to a
plus ir e to the power st e to the power minus b root s by s ds and integral by epsilon to
rho this minus 2 pi I equal 0. When we take epsilon to 0 and rho to infinity and u square
equal tR, we come to this okay this is this.
And then what we use is we use this result. Integral by 0 to infinity e power minus u
square cos 2 lu du equal 1 by 2 root pi e power minus l square and the error function.
Error function of x equal 2 by root pi integral by 0 to x e to the power of minus u square
*du and then we obtain value of this integral is equal 1 minus error function of b by 2
root t. So we get the value of this integral okay and inverse of e to the power of minus
b root s by s equal 1 minus error function of b by 2 root t.
Now the Laplace transform can be used to solve certain types of integral equation also. An
integral equation is an equation in which the unknown function, let us call it as y
t, appears under the integral sign. For example, let us say y t equal t plus integral by 0
to t y tau sin t minus tau*d tau. Then by taking Laplace transform of y t we can write
as equal Laplace transform of t is 1 by s square plus Laplace transform of integral
is 0 to t by tau sin t minus tau d tau.
And Laplace transform of ut let us take ys so ys equal 1s square. Here we use convulsion
theorem. So, Laplace transform of 0 to t y times sin t minus tau this is equal Laplace
transform of yt which is ys Laplace transform of sin t which is 1 by s square plus 1. So,
we get Y s times 1 minus 1 by square plus 1 equal 1 by s square and this gives you ys
equal s square plus 1 by s to the power 4 or we can say ys equal 1 by s square plus
1 by s to the power 4.
Inverse Laplace transform of ys let us take equal inverse Laplace transform of 1 by s
square plus Inverse Laplace transform of 1 by s to the power 4. Inverse Laplace Transform
is 1 by s square t plus inverse Laplace transform of 1 by s power 4 is 3 factorial by t cube.
So we get t plus this is t upon 3 factorial so t plus t cube by 6. Because Laplace transform
of t to the power n is n factorial upon s to the power of n plus 1 if n is a positive
integer.
So, inverse Laplace transform of 1 by s to the power of 4 will be t cube by 3 factorial.
Again by taking inverse Laplace transform of yt function equal t plus t cube by 6. With
this I would like to end my lecture. Thank you very for your attention.
