06 - Chapter 3

Download as pdf or txt
Download as pdf or txt
You are on page 1of 82

CHAPTER - 3

Application Of Spline Collocation To The Partial


Differential Equation Of One As Well As Two
Space Variables

3.1 Spline Formula To Solve Parabolic Linear Partial


Differential Equation Having One Space Variable
3.2 Flow of Electricity In A Cable of Transmission Lines
3.3 Spline Solutions With Explicit Method and Implicit Method
3.4 The Heat Conduction Problem
3.5 Spline Solutions With Explicit Method and Implicit Method
3.6 Spline Formula To Solve Parabolic Partial Differential
Equation With Two Space Variables
3.7 Heat Flow In A Thin Rectangular Plate
3.8 Spline Solutions With Explicit Method and Implicit Method
3.9 Spline Formula To Solve Hyperbolic Partial Differential
Ekjuation With One Space Variables
3.10 The Flow Of Electricity In The Transmission Lines.
3.11 Spline Solutions With Explicit Method and Implicit Method
3.12 Vibrating String Problem
3.13 Spline Solutions With Explicit Method and Implicit Method
3.14 Spline Formula To Solve Hyperbolic Partial Differential
Equation With Two Space Variables
3.15 The Problem of Vibrating Membrane
3.16 Spline Solutions With Explicit Method and Implicit Method
3.17 Conclusion

100
, , UNIVERSll

3.1 SPLINE FORMULA TO SOLVE PARABOLIC t^llfEAR/


PARTIAL DIFFERENTIAL EQUATION HAVIN
SPACE VARIABLE :

Consider parabolic PDE having one space variable.


Ut=c^u^^ 0<x<l, t>0 ...(3.1.1)
with Dirichilet b o u n d a r y conditions
u(0,t) = 0
nf^ 0 ...(3.1.1a)
u(l,t) = 0
along with initial condition,
u(x,0) = f(x) ; 0<x<l ...(3.1.1b)
We divide the region 0<x<l into, say n-equal
subinterval each of width h. Let u s denote t h e p o i n t s of
subdivisions byxo,Xi,x2, , x^. Let u s have t h e solution on
h a n d at time jAt at t h e m e s h pointsxp, Xj,X2,.....,Xj^. Let Uj j

denote t h e value of u at t h e i'^ m e s h point at time jAt.


We approximate t h e function u at time jAt by a cubic
spline S(X): calculate t h e value of S''(Xj) by solving set
of s i m u l t a n e o u s e q u a t i o n s given by (2.4.15a) for i = 1, 2, 3 ,
..., n - 1 . We should note t h a t values of u at x = 0 a n d x = 1
are already known. Now discretizing left side of equation
(3.1.1) by forward difference formula a n d replacing right
side by t h e second derivatives S"{Xi) at j * level like explicit
s c h e m e in finite difference. We get
(Ui,j,i-u,j)/At = c^S^j ...(3.1.3)

where S,"j denotes S"(x,) at j * level.

101
S u b s t i t u t e values of S|'j from (3.1.3) into equation

(2.4.15a) a n d get

= (l + 6 r ) u , , . + ( 4 - 1 2 r ) u . j + ( l + 6r)u.,,.
...(3.1.4)
where r = c At/h . This set of s i m u l t a n e o u s e q u a t i o n s c a n
now be solved equation (3.1.4) k n o w n a s cubic spline
Explicit formula to solve equation (3.1.1).
The finite difference replacement of (3.1.1)
corresponding to implicit s c h e m e is

'""'"'^^'''^ = c'(l/2)(S[j+S[j,i) ...(3.1.5)

where S-'j, Sjj+i denote second derivatives at x = Xj at t h e

time level j a n d j + 1 respectively.


However, second derivatives at Q' + l) level c a n n o t be
c o m p u t e d a s the values of u are not k n o w n . We can express
(3.1.4) in t e r m s of u a s follows.
We u s e the relationship (2.14.5a) a n d rewrite it in t h e
following forms.

SMJ+ Ki+^Ui = (6/h2)(uM,r2Ui,j-iii,i,j) ...(3.1.6)

SM,J+I+S"j+i+S|'+i j+i = (6/h )(Ui.i j+i-2ui j+i-Uj+i j+i)


i = l,2,3,....,(n-l)
...(3.1.7)
p u t t i n g t h e value of S" j+j etc from (3.1.5) into (3.1.6), we get

102
= (6/h2)(Ui.i j+i- 2ui j+i+uj+i j+i)

which gives

(l-3r)Ui.i j+i+(4 + 6r)Uij+i+ (l-3r)Ui+i j+i


= (c2At/2)(S;'.i,j+4S:,j+ S;Vi,j) + {u,.,^^+ 4n,^^+ u,,,J
= 3r(Ui.ij- 2 u i j + Ui+i j) + (Ui.ij+ 4ui j + u,+i j)
= (l + 3r)Ui.i j + (4-6r)Uij+ (l + 3r)Ui+ij

...(3.1.8)
where r = c^At/h^ , i = l,2,3,....,(n-l). Above equation
(3.1.8) is k n o w n a s cubic spline implicit formula to solve
equation (3.1.1).
Now UQ j+i and u^j+i sire k n o w n d u e to the prescribed

b o u n d a r y conditions. The set of s i m u l t a n e o u s equations


obtained in explicit a s well a s implicit s c h e m e c o n t a i n s only
(n - 1) u n k n o w n s . These (n - 1) e q u a t i o n s in (n - 1)
u n k n o w n c a n be solved by a n y s t a n d a r d m e t h o d .
Once the values of u are k n o w n at 0 + 1)^^ level, we c a n
proceed to c o m p u t e the next level (j + 2) by repeating t h e
s a m e process. For e a c h set of s i m u l t a n e o u s e q u a t i o n s in
(n - 1) u n k n o w n s give tri-diagonal matrix. It c a n be solved
by a n y s t a n d a r d m e t h o d , t h u s the m e t h o d c a n proceed by
steps.
The convergence a n d stability of t h e s e m e t h o d s totally
d e p e n d u p o n the values of r. For explicit s c h e m e , we have to
choose t h e value of r s u c h t h a t O < r < l / 2 . Implicit s c h e m e

103
h a s advantage that there is no limitation on the value of r
for convergence and stability along with small values are
more accurate. Values much larger than unity are not
desirable. These two methods will be discussed late on by
taking its actual approximation to a problem.

3.2 FLOW OF ELECTRICITY IN A CABLE OF


TRANSMISSION LINES :
This is the study of the derivation of partial differentiaJ
equations from physical principles, we consider the flow of
electricity in a cable of transmission line. We assume the
cable to be imperfectly insulated, so that there is both
capacitance and current leakage to ground. Figure (3.2.1a)
shows such a cable with the electromotive source x = 0 and
the load x = 1.

Q e<x+dx)
X+dx GCX) AAA- Kx+dx)
Kx)
source Load
T
V77777777777777777777/, V77777777777777777777Z,

Figure (3.2.1)
Flow of electricity in a cable

Here the current returns through the earth. Consider


an element PQ of the cable with P at a distance x from the
source e(x, t), the potential i(x, t), the current at p at any
time t. Let R denote the resistance, L inductance, G the

104
c o n d u c t a n c e (leak ness) a n d C t h e capacitance per u n i t
length of the cable. Let Ax be the length of t h e element PQ
a n d i + Ai, e + Ae be t h e c u r r e n t a n d the electromotive force,
respectively, at Q. Figure (3.2.1b) gives t h e electrical circuit
for t h e element PQ. Since t h e drop in voltage from P to Q is
d u e to t h e resistance RAx a n d i n d u c t a n c e LAx, we have
-Ae = (RAx)i + (LAx) a / a ...(3.2.1)
Agaiin, since t h e drop in the c u r r e n t -At is d u e to t h e
c o n d u c t a n c e GAx a n d capacitance CAx , we have
-Ai = (GAx)e + ( C A x ) a e / a ...(3.2.2)
Dividing e q u a t i o n s (3.2.1) a n d (3.2.2) by Ax a n d taking the
limit a s Ax -^ 0, we get
ai/ax + L(ai/a)+Ri = o ...(3.2.3)
and
di/dx + C(de/dt) + GQ = 0 ...(3.2.4)

Differentiating equation (3.2.3) with respect to x a n d


e q u a t i o n (3.2.4) with respect to t, we get
d^e / dx^ +L{d^i / dxdt) + Rdi / dx = 0
and
a^i/axa+c(a^e/a^) + Gae/a = 0
If we eliminate t h e t e r m dH/ dxdt between t h e s e two
equations and then s u b s t i t u t e for 51/a from equation
(3.2.3), we obtain
d^e/dx^= LeCa^e/ a ^ ) + (Re + LG) de/dt + RGe ...(3.2.5)
Similarly we obtain
dh/dx^= LeCa^i/ a 2 ) + (Re + LG) a i / a + RGi ...(3.2.6)

105
E q u a t i o n (3.2.5) a n d (3.2.6) are k n o w n a s t h e telephone
equations.
Two special case of t h e telephone e q u a t i o n s are
a^e/ax^zz Re a e / a ...(3.2.7)
d^i/dx^=Rc a/dt ...(3.2.8)
Equation (3.2.7) and (3.2.8) are obtained from
e q u a t i o n s (3.2.5) a n d (3.2.6) by taking G = L = 0 i.e. leakage
a n d i n d u c t a n c e are negligible, for example, in telegraphic
t r a n s m i s s i o n t h r o u g h s u b m a r i n e cables. These e q u a t i o n s
are k n o w n a s telegraphic e q u a t i o n s . Mathematically t h e s e
e q u a t i o n s are similar to parabolic type linear PDE having
one space variable. E q u a t i o n s (3.2.7) a n d (3.2.8) c a n be
rewrite a s
u, = ( l / R C ) u , ,
where u(x, t) s t a n d s for either i(x, t) or e(x, t).

3.3 (A) SPLINE SOLUTIONS WITH EXPLICIT SCHEME :


Let u^ = 1/Rc u^^ ...(3.3.1)
Let the initial c u r r e n t in the cable be f(x), t h e n the
initial condition is
u(x, 0) = f(x) ...(3.3.2)
where f(x) is the given function. Also t h e b o u n d a r y
conditions are
u(0, t) = 0 a n d u(l, t) = 0 V t ...(3.3.3)
We shedl determine the solution of t h e equation (3.3.1)
satisfying e q u a t i o n s (3.3.2) a n d (3.3.3). S u p p o s e the length
of t h e cable is 1 a n d t h e given function is
f(x) = sin;r ; 0 < x < l ...(3.3.4)

106
Let length of the cable is subdivided into 20
s u b i n t e r v a l s . We g e t h = Ax = l / 2 0 . Also let At = 1/400 a n d
RC = 1 / 1 0 0 , w e g e t r = 0 . 0 1 w h i c h gives
l + 6 r = 1.06 a n d 4 - 12r = 3 . 8 8
S u b s t i t u t i n g t h e v a l u e s of 1 + 6 r a n d 4 - 12r in
e q u a t i o n (3.1.4) a n d u s i n g initial c o n d i t i o n s , w e get
For j = 0

i = l, U o j + 4 u i i+U2,i = (1.06)uo,o+(3.88)ui o+(l-06)u2,o


= 0.934523
S i n c e UQ j = 0 4 u i i + U21 = 0.934523

i = 2, U i i + 4 u 2 , i + U 3 i = ( 1 . 0 6 ) u i o + ( 3 . 8 8 ) u 2 , o + ( l - 0 6 ) u 3 o
= 1.846040

i = 3, U 2 J + 4 U 3 1 + U 4 1 = (1.06)u2,o + (3.88)u3o+(1.06)u4^o
= 2.712093
i = 4, U3 1 + 4 u 4 I+U5 I = 3.511370

i = 5, U4,i+4u5,i+U6,i = 4.224184

i = 6, U5^i+4u6,i+U7i = 4.832986

i = 7, U6,i+4u7^i+U8^i = 5.322783

i = 8, U7 1 + 4 u 8 1+U9 1 = 5.621516

i = 9, U 8 i + 4 u 9 i + U i o , i = 5.900351

i = 10, U 9 I + 4 U I O , , + U H , I = 5.973899

S i n c e U9 j and Uu j a r e s y m m e t r i c , w e g e t

2u9,i+4u2io,i =5.973899

H e r e w e g e t 10 a l g e b r a i c e q u a t i o n s i n 10 u n k n o w n s
with t r i - d i a g o n a l m a t r i x . T h i s s y s t e m of e q u a t i o n s is
solved e a s i l y b y a n y well k n o w n m e t h o d like c r o u t ' s m e t h o d ,

107
similarly for j = 1, we get another 10 algebraic equations.
This can be solved easily by above method. Proceeding in
this way, the results obtained by explicit method are shown
in table (3.3.1) and plotted in figure (3.3.1).
Table (3.3.1)
Current distribution in the cable through cubic spline
explicit method

Current in the cable u ->


X t=0.0 t= 1/400 t=l/200 t=3/400 t=l/100 t = 1/80
0.0 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000
0.05 0.156434 0.156396 0.156357 0.156318 0.156280 0.156241
0.10 0.309017 0.308941 0.308864 0.308788 0.308711 0.308635
0.15 0.453991 0.453878 0.453766 0.453654 0.453542 0.453429
0.20 0.587785 0.587640 0.587495 0.587349 0.587204 0.587059
0.25 0.707107 0.706932 0.706757 0.706582 0.706407 0.706233
0.30 0.809017 0.808817 0.808617 0.808417 0.808217 0.808017
0.35 0.891007 0.890786 0.890566 0.890346 0.890126 0.889906
0.40 0.951057 0.950821 0.950586 0.950351 0.950116 0.949882
0.45 0.987688 0.987444 0.987200 0.986956 0.986712 0.986468
0.50 1.000000 0.999753 0.999506 0.999258 0.999011 0.998764

1 9nnnnn • t ' . •.- • • . •

1.000000 ' - • - t = 0.0 I


T 0.800000 -•-t = 1/400
1 t= 1/200
3 0.600000 1 = 3^400
0.400000 -«-t = 1/100i
-•-t=1/80
0.200000
nmrmni
0.0 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50
X

Figure (3.3.1)
Current distribution in the cable through cubic spline
explicit method

108
3 . 3 ( B ) SPLINE SOLUTIONS WITH IMPLICIT METHOD :
I n t h i s s e c t i o n w e d i s c u s s t h e s o l u t i o n of e q u a t i o n
(3.3.1) b y implicit s c h e m e . S u b s t i t u t e t h e v a l u e s of r a n d
u s i n g initial c o n d i t i o n s i n (3.1.8), w e get,
Forj = 0 i.e. at t = 1/400
i = l, (0.97)uo,i+(4.06)uii+(0.97)u2,i = 0.934639

since u^, = 0, w e g e t

(4.06) u i i + ( 0 . 9 7 ) U2,i = 0.934639

i = 2, (0.97)ui 1+(4.06) U2J+(0.97) U31 = 1.846264

= 3, ( 0 . 9 7 ) u 2 , i + ( 4 . 0 6 ) u 3 i + ( 0 . 9 7 ) u 4 i = 2.712429

= 4, (0.97) U31+(4.06) U41+(0.97) U51 = 3.511804

= 5, (0.97) U41+(4.06) U51+(0.97) ug 1 = 4.22407

= 6, (0.97) U51+(4.06) U61+(0.97) U71 = 4.833583

= 7, (0.97)U6,+(4.06) U7]+(0.97) Us 1 = 5.323441

= 8, (0.97) U71+(4.06) U81+(0.97) U91 = 5.682218

= 9, (0.97)u8,i + (4.06) U91+(0.97) Uio,i = 5.901080

i = 10, (0.97)U9 1+(4.06) Uio,i+(0.97) uii 1 = 5.974638

s i n c e U9 j and Ujj ^ a r e s y m m e t r i c , w e get,

(0.97) U91+(4.06) uioj = 5.974638

H e r e w e g e t 10 a l g e b r a i c e q u a t i o n s i n 10 u n k n o w n s
with t r i - d i a g o n a l m a t r i x . T h i s c a n b e solved b y a n y
s t a n d a r d m e t h o d . Similarly, applying above p r o c e s s are, we
get t h e solution at t = 1 / 2 0 0 , 3 / 4 0 0 , 1 / 1 0 0 , 1/80 etc. a n d
t h e y a r e s h o w n i n t a b l e (3.3.2) a n d p l o t t e d i n figure (3.3.2).

109
Table (3.3.2)
Current distribution in the cable through cubic spline
implicit method

Current in the cable u -^


X t=0.0 t= 1/400 t= 1/200 t=3/400 t=l/100 t = 1/80
0.0 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000
0.05 0.156434 0.156396 0.156357 0.156318 0.156280 0.156241
0.10 0.309017 0.308941 0.308866 0.308789 0.308713 0.308636
0.15 0.453991 0.453878 0.453759 0.453648 0.453536 0.453430
0.20 0.587785 0.587640 0.587496 0.587351 0.587205 0.587060
0.25 0.707107 0.706932 0.706757 0.706582 0.706408 0.706233
0.30 0.809017 0.808817 0.808617 0.808417 0.808217 0.808018
0.35 0.891007 0.890786 0.890566 0.890346 0.890126 0.889906
0.40 0.951057 0.950821 0.950586 0.950352 0.950117 0.949882
0.45 0.987688 0.987444 0.987200 0.986956 0.986712 0.986468
0.50 1.000000 0.999753 0.999506 0.999258 0.999011 0.998764

1.200000 1 \WS|i'l',t:-i'H-*- •;•;

1.000000
-•-t = 0 0
t 0.800000 - ^ t = 1/400
_^-"'*^^^
1 t = 1/2D0
3 oeooooo X ^ — 1 = 3^400^
Q400000 y ^ -*-t = 1/100
MT -•-t = 1/80 1
02D0000

nnfYYYYi 1 X
00 0.05 0 1 0 015 O 2 0 0 2 5 0 3 0 035 0.40 0 4 5 O 5 0
X •

Figure (3.3.2)
Current distribution in the cable through cubic
spline implicit method

no
3.3(C) DISCUSSION OF RESULTS :
Here table (3.3.3) gives the comparison of both the
spline solutions namely explicit and implicit with exact
solutions.
From the table (3.3.3) it is clear that, the spline
solutions are fairly agree with exact solutions up to five
digits of decimal points. Figure (3.3.3(a)) indicates the error
analysis which compares the exact solution with spline
solution obtained by both the methods at t = 1/80. The
figure (3.3.3(b)) gives good agreement of curves presenting
exact and approximate solutions obtained by cubic spline
method.
Table (3.3.3)
Error analysis
Current: in the cable u -> t = 1/80
X USE USI UEXT USE-UEX USI-UEX
0.0 0.000000 0.000000 0.000000 0.000000 0.000000
0.05 0.156241 0.156241 0.156242 0.000001 0.000001
0.10 0.308635 0.308636 0.308636 0.000001 0.000000
0.15 0.453429 0.453430 0.453431 0.000002 0.000001
0.20 0.587059 0.587060 0.587060 0.000001 0.000000
0.25 0.706233 0.706233 0.706235 0.000002 0.000002
0.30 0.808017 0.808078 0.808019 0.000002 0.000001
0.35 0.889906 0.889906 0.889908 0.000002 0.000002
0.40 0.949882 0.949882 0.949884 0.000002 0.000002
0.45 0.986468 0.986468 0.986471 0.000003 0.000003
0.50 0.998764 0.998764 0.998767 0.000003 0.000003

[USE and USI are current in the cable obtained by


using explicit, implicit method respectively. UEXT
denote the exact solution.]

11
0.000004

0.000003

0.000003

DC 0.000002 -•—USE-UEXT
O
0.000002 •«—USI-UEXT
LU

0.000001

0.000001

0.000000
0.0 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50

Figure (3.3.3(a))
Error analysis (at t = l / 8 0 )

112
1.200000

•^—USE
!
3 -•—USI
UEX

0.000000 A
0.0 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50

X - "

Figure (3.3.3(b))

Current distribution in the cable through cubic spline method


(at t=l/80)

113
3.4 THE HEAT CONDUCTION PROBLEM :
Consider a thin long rod surrounded except at
the ends with a material impervious to heat unless all the
points of the rod are at the same temperature, heat will flow
along the rod. If the rod is homogeneous and of the same
cross section throughout, we may schematically regard the
rod as a line, since the temperature of all the points of any
cross section will sensibly the same. When heat is flowing
uniformly, it is experimentally known that the amount of
heat flow across any portion of the rod is proportional to the
difference of temperatures of the end points of the portion,
to the area of cross - section and to the end points of the
flow and inversely proportional to the length of the portion
considered. Taking the limiting case when the length of the
portion considered tends to zero, we obtain the quamtity of
heat Qi that flows across any section of the rod as.

Qi = -k ^ A per sec .,.(3.4.1)


&

where, k : Coefficient of conductivity


u : The temperature at a distance x from some
fixed point
on the rod
and A : Area of cross section
The negative being attached as the heat flows from a
higher to lower temperature. If we take the section at a
pointx + &:, then a quantity of a heat that flow out across
this section is given by

114
Sa
Q2 = -k A per sec ...(3.4.2)
X+SK

t h u s from (3.4.1) a n d (3.4.2) above the quamtity of h e a t


gained by this section per sec is

Qi - Q2 = k — ...(3.4.3)
X+SK X_

The rate of rise of t e m p e r a t u r e is di/S t. Therefore Qj - Qj is


also given by
(J
Q, - Q2 = S.P. SK .A. — ...(3.4.4)

where S = Specific h e a t
and P = Density of the rod
e q u a t i n g the values of Qj - Q2 from (3.4.3) a n d (3.4.4) a n d
dividing by Ss., we have
Sa Sa
Sa Si x+Sx Si
S.P. ,(3.4.5)
St SK
taking t h e limit of this e q u a t i o n a s Ss.-^ 0, we get

S.P. ^ =k^'^
St ^2
Sa k S^u
or
St S.P. &2

writing a = t h e eequat]
q u a t i o n of one dimension h e a t flow is
S.P. ^
Sa S^a
.(3.4.6)
c^t &2

In shortly, we summarize these phenomena by


considering a h o m o g e n e o u s rod of length L. The rod is
sufficiently thin, so t h a t the h e a t is d i s t u r b e d equally over

15
t h e cross section at time t. The surface of t h e rod is
i n s u l a t e d a n d therefore there is no h e a t loss t h r o u g h the
b o u n d a r y . The t e m p e r a t u r e distribution of t h e rod is given
by t h e solution of initial b o u n d a r y value problem
Ut = aUxx , 0<x<L t>0
u(0, t) = 0, t >0
,.(3.4.7)
u(L,t) = 0, t<0
u(x,0) = f(x), 0 < x < L

3 . 5 (A) SPLINE SOLUTIONS WITH EXPLICIT METHOD :


We illustrate the problem by considering a
h o m o g e n e o u s rod of length 1 meter a n d fluid having u n i t
viscosity (v) a n d the t e m p e r a t u r e is given by function f(x) =
X (1 - x), t h e n p h e n o m e n a c a n be written a s ,
Ut=auxx 0<x<l; t>0 ...(3.5.1)
u(0,t) = 0 1
h^O ...(3.5.2)
u(l,0) = 0 J
and
u(x,0) = x ( l - x ) 0<x<l ...(3.5.3)
We shall determine the solution of the equation (3.5.1)
satisfying equation (3.5.2) a n d (3.5.3).

Let u s take Ax= h = — , A t = k =


10 500
a = 1 and r = 0.2 which gives,
1 + 6r = 2.2 and 4-12r=1.6
S u b s t i t u t i n g the values of (1 + 6r) a n d (4 - 12r) in
equation (3.1.4) a n d u s i n g initial conditions, we get
for j = 0, i = 1,

116
U o , i + 4 u i , + U 2 , i = (2.2)uo,o+(l-6)uio+(2.2)u2,o = 0.498

since UQ i = 0

4 u i , i + U 2 , i = 0.498

i = 2,
U i , i + 4 u 2 , i + U 3 i = ( 2 . 2 ) u i o + ( l - 6 ) u 2 , o + ( 2 . 2 ) u 3 o = 0.916

i = 3,
U2,l+4U3,1+U4i = (2.2)U2,o+(l-6)U3o+(2.2)U4o = 1.216

i = 4,
U31+4U41+U51 = ( 2 . 2 ) u 3 o + ( 1 . 6 ) u 4 o + ( 2 . 2 ) u 5 o = 1.396

i = 5,
U4,i+4u5^i+U6^i = (2.2) U4^o+(l-6)115,0+(2.2) ug^o = 1-456

U 5 , i + 4 u 6 i + U 7 i = (2.2)u5o+(1.6)u6,o+(2.2)u7o = 1.396

i = 6,
U5,i+4u6,i+U7i = (2.2)u5o+(1.6)u6,o+(2.2)u7o = 1.396

i = 6,
U5,i+4u6,i+U7i = (2.2)u5o+(1.6)u6,o+(2.2)u7o = 1.396

U 6 , i + 4 u 7 j + U 8 i = (2.2)u6o+(1.6)u7o+(2.2)u8,o = 1-216

i = 8,
U7,i+4u8,i+U9 J = (2.2)u7o+(1.6)u8,o+(2.2)u9o = 0.916

i = 9,

U8,i+4u9i+Uio,i = (2.2)u8o+(1.6)u9o+(2.2)uio,o
U8,i+4u9,i=0.408

s i n c e UJQ 1 = 0

117
Hence we get 9 algebraic equations in 9 unknowns
with tri-diagonal matrix. This system of equation is solved
by any well known method. Similarly, for j = 1, we get
another 9 algebraic equations. This can be solved by above
method. Proceeding in this way, the results obtsdned by
explicit method are shown in table (3.5.1) and plotted in
figure (3.5.1).

Table (3.5.1)
Temperature in a rod through cubic spline solution by
explicit method

Temperature u(x, t)

X t=0.0 t=0.002 t=0.004 t=0.006 t=0.008 t=0.01

0.0 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000


0.1 0.09 0.086 0.0828 0.08008 0.07768 0.07551
0.2 0.16 0.156 0.152 0.14816 0.144512 0.14104
0.3 0.21 0.206 0.202 0.198 0.194032 0.19012
0.4 0.24 0.236 0.232 0.228 0.224 0.22000
0.5 0.25 0.246 0.242 0.238 0.234 0.23
0.6 0.24 0.236 0.232 0.228 0.224 0.22000
0.7 0.21 0.206 0.202 0.198 0.194032 0.19012
0.8 0.16 0.156 0.152 0.14816 0.144512 0.14104
0.9 0.09 0.086 0.0828 0.08008 0.07768 0.07551
1.0 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000

118
nTimnn
\J.\XKXkAJ

0.250000
-•-t=Q0 i
1 0200000 -•-t=Q002
t=Q004i
3 0150000
-^^ t=0.006
Q100000 HK-t=o.ooe \
- • - t = Q01|
OQBDOOO

nmnrmi
0.0 01 02 03 0.4 05 06 0.7 08 09 1.0

X •

Figure (3.5.1)
Temperature in t h e rod t h r o u g h cubic s p l i n e explicit
method

3 . 5 (B) SPLINE SOLUTIONS WITH IMPLICIT METHOD :


In t h i s section we d i s c u s s the solution of (3.5.1) by
implicit s c h e m e . S u b s t i t u t e t h e values of r a n d u s i n g initial
conditions in (3.1.8), we get,
For j = 0 i.e. t = 0.002
i = 1 (0.4)uo,i +(5.2)ui,, +(0.4)U2,, = 0.508

Since Uo,,= 0, (5.2)ui,+(0.4)U2,i = 0.508

i= 2 (0.4) u,, + (5.2) U2,1 + (0.4) u^^, = 0.928

i= 3 (0.4)u2^i +(5.2)U3 1 +(0.4)U4 , = 1.228

i= 4 (0.4)U3, +(5.2)U4 1 +(0.4)U5 , = 1.408

i= 5 (0.4)U4 , +(5.2)U5 1 +(0.4)U6,, = 1.468

119
i =6 (0.4)u51 +(5.2)u6 1 +(0.4)u7 1 = 1.408

i =7 (0.4)u6,i +(5.2)u7 1 +(0.4)u8,i = 1.228

i =8 (0.4)u7 1 +(5.2)u81 +(0.4)u9 1 = 0.928

i =9 (0.4)u8i+(5.2)u9i+(0.4)uio,i = 0.508

Since Uio,i= 0

(0.4)u8i+ (5.2) U91 = 0.508

Hence we get 9 algebraic e q u a t i o n s in 10 u n k n o w n s


with tri-diagonal matrix. This system of e q u a t i o n s is
solved by any well k n o w n m e t h o d . Similarly, for j = 1, we get
a n o t h e r 9 algebraic e q u a t i o n s . This c a n be solved by above
m e t h o d . Proceeding in this way, t h e r e s u l t s obtained by
implicit m e t h o d are s h o w n in table (3.5.2) a n d plotted in
figure (3.5.2).
Table (3.5.2)
Temperature in a rod t h r o u g h cubic s p l i n e s o l u t i o n by
implicit m e t h o d

Temperature u(Xy t)
X t = 0.0 t= 0.002 t =0.004 t =0.006 t =0.008 t = 0.01
0.0 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000
0.1 0.09000 0.08492 0.08208 0.07914 0.07703 0.07464
0.2 0.16000 0.15628 0.15137 0.14777 0.14357 0.14055
0.3 0.21000 0.20592 0.20233 0.19868 0.19411 0.18939
0.4 0.24000 0.23602 0.23185 0.22802 0.22366 0.22040
0.5 0.25000 0.24598 0.24209 0.23771 0.23440 0.22959
0.6 0.24000 0.23602 0.23185 0.22802 0.22366 0.22040
0.7 0.21000 0.20592 0.20233 0.19868 0.19411 0.18939
0.8 0.16000 0.15628 0.15137 0.14777 0.14357 0.14055
0.9 0.09000 0.08492 0.08208 0.07914 0.07703 0.07464
1.0 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000

120
H-t=Q(]0 ^
-^t=Q(]02j
t=Q004^

HK-t=Q(]06|
-•-t=Q01 :

QO 0.1 0.2 03 0.4 05 0.6 07 08 09 1.0

Figure (3.5.2)
Temperature in rod through cubic spline implicit
method

3.5(C) DISCUSSION OF RESULTS :


Table (3.5.3) gives the comparison of both the spline
solutions namely explicit and implicit with exact solution.
From table (3.5.3) it is clear that, the spline solutions are
fairly agree with exact solutions u p to three digits of decimal
points, figure (3.5.3(a)) indicates the error analysis which
compares the exact solution with spline solution obtained
by both the methods t = 0.008. The figure (3.5.3(b)) gives
good agreement of curves presenting exact and approximate
solutions obtained by cubic spline method.

121
Table (3.5.3)
Error analysis

Temperature u(x, t) Att = 0.006

X USE USl UEX USE-UEX USI-UEX

0.0 0.000000 0.000000 0.000000 0.000000 0.000000

0.1 0.080080 0.079149 0.079683 0.000397 0.000534

0.2 0.148160 0.147771 0.148269 0.000109 0.000498

0.3 0.198000 0.198685 0.198467 0.000467 0.000218

0.4 0.228000 0.228024 0.227979 0.000021 0.000045

0.5 0.238000 0.237713 0.237569 0.000431 0.000144

0.6 0.228000 0.228024 0.227979 0.000021 0.000045

0.7 0.198000 0.198685 0.198467 0.000467 0.000218

0.8 0.148160 0.147771 0.148269 0.000109 0.000498

0.9 0.080080 0.079149 0.079683 0.000397 0.000534

1.0 0.000000 0.000000 0.000000 0.000000 0.000000

122
0.000600

0.000500

0.000400
DC
O -•—USE-UEX
0.000300
oc -•—USI-UEX
lU
0.000200

0.000100

0.000000
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

Figure (3.5.3(a))
Error analysis (at t=0.006)

123
0.050000

0.000000 A
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

Figure (3.5.3 (b))


Temperature distribution in the rod through cubic spline method

124
3 . 6 SPLINE FORMULA TO SOLVE PARABOLIC PARTIAL
DIFFERENTIAL EQUATION WITH TWO SPACE
VARIABLES :
Consider the parabolic differential equation having two
space variables.
Uj = c^ (Uxx + Uyy) ; 0<x<a,0<y<b,t>0 ...(3.6.1)

With Dirichilet conditions prescribed on the


b o u n d a r i e s x = 0, x = a, y = 0, y = b . We should
subdivide the region 0 < x < a into, say M intervals, each of
width Ax a n d 0 < y < b into N intervals of width Ay s u c h t h a t
MAx = a and NAy = b . Let us denote the points of

subdivisions by Xo,Xi,X2, , xM a n d yo,yi,y25 , yN. Let

Uj j.k denote the value of u' at the (i, j ) * m e s h point at the

time kAt. For simplicity, let u s take a square region


0 < x , y < a , M = N a n d Ax = Ay = h(say).
We approximate t h e function u at time At by a cubic
spline S(x). Discretizing the left side of equation (3.6.1) by
forward difference formula a n d replacing right side by twice
t h e second derivative i.e. 2S"(Xjj) at k^*^ level like explicit
s c h e m e in finite difference, we get

K k . i - U y , k ) / A t =c2(2S;,k) ...(3.6.2)

where is SjJ ^ d e n o t e s S^x^) at k*^

level. #
Now with the help of equation (2.4.15a) a n d t h e value
of Sjj k obtained from equation (3.6.2) we get,

125
Ui-lj,k+l " " i - l j , k Uy 1^+1 - Ujj ]^ Uj_^ij k+1 - Uj+ij k

2c^At + 2c^At ' 2c^At


= ( 6 / h ^ ) (uj.ij^k- 2 u i j k + Ui+ijk)

At last we get,
Ui-lj,k+l + 4 U y j^+l + Uj+ij k+l

= (l + 12r)Ui.ij,k+ (4-24r) u^^^^ + (l + 12r)u,^ij^k)


i,j = l,2, ,N-1
...(3.6.3)

where r = c A t / h t h e s e set of (N - 1) x (N - 1) equation in


(N - 1) X (N - 1) u n k n o w n s can be solved by any well -
k n o w n method. The above set of simultEineous e q u a t i o n s
gives s q u a r e matrix. The equation (3.6.3) is k n o w n a s cubic
spline explicit formula to solve equation (3.6.1). For this
m e t h o d , the m a x i m u m possible value of r is % b u t the
equation (3.6.1) having two space variables a n d equal grid
spacing, hence r < 1/6 is required for stability and
convergence. The difficulty while u s i n g the explicit s c h e m e
is t h a t the restriction on At requires inordinately m a n y rows
of calculations. In s u c h case one looks for a m e t h o d in
which At c a n be m a d e larger without lost of stability. The
implicit m e t h o d w a s s u c h a method.

The implicit m e t h o d , the finite difference s c h e m e of


equation (3.6.1) is

K k . i - Uij,k)/At = c^ (S;;,, +S;,k.i) ...(3.6.4)

126
where SjJ j^ and S,j ^+1 d e n o t e s e c o n d d e r i v a t i v e s of S(x) a t

X = Xjj a t t h e t i m e i n t e r v a l k a n d k + 1 r e s p e c t i v e l y .

W e c a n e x p r e s s (3.6.1) i n t e r m s of u a s follows. W e u s e
t h e r e l a t i o n s h i p ( 2 . 4 . 1 5 a ) a n d r e w r i t e it a s

.(3.6.5)
= (6/h') ( u , , j , - 2 u . j , + u . , , . J

c" -4- 4 Q ' ' 4- <;"

•^i-lj.k + l ^ ^"^ij.k + l ^ '^i+lj.k + l

= (6/h^) (uMo,k.i - 2u.,^,,„ + u,,,,^,,,,) ...(3.6.6)


i j = l,2, ,N-1
w i t h t h e h e l p of e q u a t i o n s (3.6.5) a n d (3.6.6) u s i n g t h e
v a l u e of Sjj k+i from e q u a t i o n (3.6.4) w e get,

1/c'At (Ui.,,j,,„ - Ui.i^j,,) - S:.,,j,, + 4{(l/c'At)(Ui,j,,,, - u,,j,k)- S:,,, }


+ 4{(l/c'At)(Ui,,,j,k.i-u„,,j,,)-S:,,,j,, }
= (6/h') (u,., j k^, - 2Uij k+, + Ui^,jk+i)

t h i s gives

(l-6r)u..,j ,„ + (4 + 12r) u , j , , , + ( l - 6 r ) U i , , ^ , ^
= (l + 6 r ) u , , , , + (4-12r) u , , , + (l + 6r)u.„ .,,)

w h e r e r = c^At/h^ a n d i,j = l , 2 , ,N-1

Above equation (3.6.7) is known as cubic spline


i m p l i c i t f o r m u l a to solve e q u a t i o n (3.6.1). Like s c h e m e , w e
get (N - 1) X (N - 1) s i m u l t a n e o u s e q u a t i o n s i n (N - 1) x
( N - 1 ) u n k n o w n s . These equations with s q u a r e matrix can
b e solved b y a n y s t a n d a r d m e t h o d .

127
In b o t h the m e t h o d s , once the values of u are k n o w n at
(k + 1)* level, we c a n proceed to c o m p u t e t h e next level (k +
2) by s a m e t e c h n i q u e s a s above. The coefficient matrix of
the combined equation is a s q u a r e matrix, however, t h e
system becomes a tri-diagonal one w h e n s e p a r a t e c a s e s are
h a n d l e d . These two m e t h o d s will be d i s c u s s e d later on by
taking its a c t u a l approximation to a problem.

3.7 HEAT FLOW IN A THIN RECTANGULAR PLATE :

Consider the flow of h e a t in a thin r e c t a n g u l a r plate


with sides of length Ax and Ay along co-ordinals x a n d y.

y^

Figure (3.7.1)
Thin rectangular plate

The a m o u n t of h e a t entering the element t h r o u g h the


side AB in time At is
-K(Ay)(ay/ax)^ At
a n d t h a t leaving t h e element t h r o u g h t h e opposite side CD
is
-K(Ay)(ay/ax),,^, At

128
where K is the t h e r m a l s conductional of t h e material a n d
u(x, y, t) is the t e m p e r a t u r e function. The negative signs are
t a k e n b e c a u s e the h e a t flows in the direction of decreasing
t e m p e r a t u r e . Hence t h e quantity of h e a t remaining in the
plate a s a result of entry t h r o u g h t h e side AB a n d exit
t h r o u g h t h e side CD is
K {(du/dx\ , ^, - (au/ax), } AyAt
= K{(6Vax^)Ax}AyAt
u p t o a first approximation.
Similarly corresponding difference in t h e h e a t entering
a n d leaving t h r o u g h t h e r e m a i n i n g pair of opposite side is
K{(aV5x2)Ay}AxAt ...(3.7.2)
Hence the total h e a t retained by t h e plate in time At is
t h e s u m of r e s u l t s (3.7.1) a n d (3.7.2), which is equal to the
h e a t required to raise the t e m p e r a t u r e of the element by Au .
T h u s we have
K{(a^u/ax^) + ( a V V ) } A x A y A t = (pAxAy)SAu ...(3.7.3)
where p is t h e density a n d S be the specific h e a t of the
plate. Dividing the equation (3.7.3) by AxAySp a n d taking
limit At ^^ 0, we get
au/a = c^ (d^u/dx^+d\/dy^)

where c^ = KJSp ...(3.7.4)


which is parabolic PDE having two space variables x a n d y
a n d time variable t.
Consider the edges of thin s q u a r e plate of side 1 (figure
3.7.2) are kept at t e m p e r a t u r e zero a n d faces are perfectly
insulated.

129
0 1 X
Figure (3.7.2)
Thin square plate

Hence t h e flow of h e a t in the plate is governed by


equation (3.7.3) with b o u n d a r y with b o u n d a r y conditions.
u(x,0,t) = 0 0<x,y<l
...(3.7.5)
u(0, y, t) = 0
a n d let initial t e m p e r a t u r e distribution in t h e plate be
u(x,y,0) = sin;rx sin;ry 0<x,y<l ...(3.7.6)
The given problem with boundary and initial
conditions is solved by explicit a s well a s implicit m e t h o d a s
follows.

3 . 8 (A) SPLINE SOLUTIONS WITH EXPLICIT METHOD :


We shall determine the solution of equation (3.7.3)
satisfying b o u n d a r y conditions (3.7.5) a n d initial conditions
given by the equation (3.7.6) respectively, by u s i n g the
explicit formula given in equation (3.6.3).
Let h = l/20,c^ = 0.001 and At = l/400which gives
R = 0.001
Hence 1 + 12r = 1.012

130
4-24r = 3.976
S u b s t i t u t i n g the values of 1 + 12r, 4 - 2 4 r with initial
a n d b o u n d a r y conditions in equation (3.6.3), we get
For k = 0
For j = 1
i = 1 Uo,i,i + 4ui^i,, + U2,i,i = 0.146220

Since UQ i i = 0 we have

4ui,i,i+U2ii = 0.146220

i = 2 ui^i^i + 4u2,i,i+U3,i,i = 0.288841

i = 3 U2,i,i + 4u3ii+U4^i,i = 0.424349

Similarly we get 19 x 19 s i m u l t a n e o u s e q u a t i o n s in 19
X 19 u n k n o w n s , for j = 1, 2, 3 , , 19 where i = 1, 2, 3 ,
, 19. This c a n be solved by any s t a n d a r d m e t h o d . Once
t h e r e s u l t s are obtained, for r e s u l t s for (k + 1)* level, r e s u l t s

for (k + 2)* level are obtained in similar m a n n e r d i s c u s s e d


a s above. Due to t h e symmetry of the solutions, the r e s u l t s
are given for 0 < x < 0 . 5 , 0 < y < 0 . 5 at t = 1/400, 3 / 4 0 0 a n d
1/80 respectively in t h e table (3.8.1(a)) - (3.8.1(c)). Results
are obtained by explicit m e t h o d for y = 0.25 plotted in the
figures (3.8.1).

131
Table (3.8.1(a))
Temperature distribution in thin rectangular plate through explicit method

Temperature in thin rectangular plate u -> at t = 1/400

X
0.0
0.0 0.05
0.000000 0.000000
0.10 0.15
0.000000 0.000000
0.20
0.000000
0.25
0.000000
0.30
0.000000
0.35 0.40
0.000000 0.000000
0.45
0.000000
0.50
0.000000

0.05 0.000000 0.024470 0.048339 0.071016 0.091945 0.110610 0.126552 0.139377 0.148770 0.154500 0.156427

0.10 0.000000 0.048339 0.095487 0.140284 0.181627 0.218497 0.249988 0.275323 0.293878 0.305197 0.309002

0.15 0.000000 0.071016 0.140284 0.206097 0.266836 0.321004 0.367268 0.404489 0.431749 0.448379 0.453968

0.20 0.000000 0.091945 0.181627 0.266836 0.345474 0.415606 0.475505 0.523695 0.558989 0.580520 0.587756

0.25 0.000000 0.110610 0.218497 0.321004 0.415606 0.499975 0.572033 0.630006 0.672465 0.698367 0.707072

0.30 0.000000 0.126552 0.249988 0.367268 0.475505 0.572033 0.654476 0.720804 0.769383 0.799017 0.808977

0.35 0.000000 0.139377 0.275323 0.404489 0.523695 0.630006 0.720804 0.793853 0.847356 0.879993 0.890962

0.40 0.000000 0.148770 0.293878 0.431749 0.558989 0.672465 0.769383 0.847356 0.904464 0.939301 0.951009

0.45 0.000000 0.154500 0.305197 0.448379 0.580520 0.698367 0.799017 0.879993 0.939301 0.975480 0.987640

0.50 0.000000 0.156427 0.309002 0.453968 0.587756 0.707072 0.808977 0.890962 0.951009 0.987640 0.999951

132
Table (3.8.1(b))
Temperature distribution in thin rectangular plate through explicit method

Temperature in thin rectangular plate u ^ at t = 3/400

0.0 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0,45 0.50
^
0.0 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000

0.05 0.000000 0.024468 0.048334 0.071009 0.091936 0.110599 0.126539 0.139363 0.148756 0.154486 0.156411

0.10 0.000000 0.048364 0.095477 0.140270 0.181609 0.218476 0.249963 0.275295 0.293849 0.305167 0.308971

0.15 0.000000 0.071009 0.140270 0.206077 0.266809 0.320975 0.367229 0.404052 0.431704 0.448335 0.453923

0.20 0.000000 0.091936 0.181609 0.266809 0.345440 0.415565 0.475458 0.523643 0.558934 0.580463 0.587698

0.25 0.000000 0.110599 0.218476 0.320975 0.415565 0.499926 0.571977 0.629943 0.672399 0.698298 0.707002

0.30 0.000000 0.126539 0.249963 0.367229 0.475458 0.571977 0.654412 0.720733 0,769307 0.798938 0.808897

0.35 0.000000 0.139363 0.275295 0.404452 0.523643 0.629943 0.720733 0.793775 0.847272 0.879906 0.890874

0.40 0.000000 0.148754 0.293849 0.431704 0.558934 0.672399 0.769307 0.847272 0.904374 0.939208 0.950915

0.45 0.000000 0.154486 0.305167 0.448335 0.580463 0.698298 0.798938 0.879906 0.939208 0.975383 0.987542

0.50 0.000000 0.156411 0.308971 0.453923 0.587698 0.707002 0.808897 0.890874 0.950915 0.987542 0.999852

133
Table (3.8.1(c))
Temperature distribution in thin rectangular plate through explicit method

Temperature in thin rectangular plate u -> at t = 1/80

0.0 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50
0.0 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000

0.05 0.000000 0.024466 0.048329 0.071002 0.091927 0.110588 0.126527 0.139350 0.148741 0.154470 0.156395

0.10 0.000000 0.048329 0.095468 0.140256 0.181591 0.218454 0.249938 0.275268 0.293820 0.305137 0.308940

0.15 0.000000 0.071002 0.140256 0.206057 0.266783 0.320942 0.367189 0.404019 0.431657 0.448292 0.453877

0.20 0.000000 0.091927 0.181591 0.266729 0.345406 0.415524 0.475411 0.523591 0.558879 0.580405 0.587640

0.25 0.000000 0.110588 0.218454 0.320942 0.415524 0.499876 0.571920 0.629881 0.672332 0.698229 0.706932

0.30 0.000000 0.126527 0.249938 0.367189 0.475411 0.571920 0.654347 0.720662 0.769231 0.798860 0.808816

0.35 0.000000 0.139350 0.275268 0.404409 0.523591 0.629881 0.720662 0.793697 0.847188 0.879820 0.890786

0.40 0.000000 0.148751 0.293820 0.431657 0.558879 0.672332 0.769231 0.847188 0.904285 0.939116 0.950821

0.45 0.000000 0.154470 0.305137 0.448292 0.580405 0.698229 0.798860 0.879820 0.939116 0.975287 0.987444

0.50 0.000000 0.156395 0.308940 0.453877 0.587640 0.706932 0.808816 0.890786 0.950821 0.987444 0.999753

134
0.800000
0.700000
0.600000
0.500000 ••-t=1/400
t 0.400000 -»-t=3/400
0.300000 t=1/80

0.200000
0.100000

0.000000 rV
0.0 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50

Figure (3.8.1)
Temperature distribution in tliin rectangular plate through cubic spline explicit method
(at y=0.25)

135
3.8(B) SPLINE SOLUTIONS WITH IMPLICIT METHOD :

Using implicit formula given by equation (3.6.7), the


solution of equation (3.7.3) satisfying b o u n d a r y a n d initial
conditions, t h o s e are given in section (3.7) is obtadned a s
follows :
For k = 0
For j = 1
i = 1, (0.994)uo,i,i + (4.012)uiii+ (0.994)u2,i,i = 0.146224

i = 2, (0.994)uiii + (4.012)u2,i,i+ (0.994)u3i j = 0.288848

i = 3,(0.994)u2,i,i + (4.012)u3ii+ (0.994)u4ii = 0.424359

Proceeding in t h i s way, we get 19 x 19 s i m u l t a n e o u s


e q u a t i o n s in 19 x 19 u n k n o w n s , for j = 1, 2, 3, , 19
where i = 1, 2, 3, , 19. Solving t h e set of e q u a t i o n s by
a n y well k n o w n m e t h o d , the t e m p e r a t u r e distribution in the
plate is obtained. Once the r e s u l t s are obtained, for r e s u l t s
for (k + 1)* level, the r e s u l t s for (k + 2)* level are obtained in
similar m a n n e r d i s c u s s e d a s above. Due to t h e symmetry of
t h e solutions, the r e s u l t s are given for 0 < x < 0 . 5 , 0 < y < 0 . 5 at
t = 1/400, 3 / 4 0 0 and 1/80 respectively in the tables
(3.8.2(a)) - (3.8.2(c)). Results are obtained by implicit
m e t h o d for y = 0.25 plotted in t h e figures (3.8.2(a)).

136
Table (3.8.2(a))
Temperature distribution in thin rectangular plate through implicit method

Temperature in thin rectangular plate u -> at t = 1/400

0.0 o.os 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50
0.0 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000

0.05 0.000000 0.024470 0.048339 0.071016 0.091945 0.110610 0.126551 0.139377 0.148771 0.154501 0.156427

0.10 0.000000 0.048339 0.095487 0.140284 0.181627 0.218497 0.249988 0.275323 0.293878 0.305197 0.309002

0.15 0.000000 0.071016 0.140284 0.206097 0.266836 0.321004 0.367268 0.404489 0.431749 0.448379 0.453968

0.20 0.000000 0.091945 0.181627 0.266836 0.345474 0.415606 0.475505 0.523695 0.558989 0.580520 0.587756

0.25 0.000000 0.110610 0.218497 0.321004 0.415606 0.499975 0.572033 0.630006 0.672465 0.698367 0.707072

0.30 0.000000 0.126552 0.249988 0.367268 0.475505 0.572033 0.654476 0.720804 0.769383 0.799017 0.808977

0.35 0.000000 0.139377 0.275323 0.404489 0.523695 0.630006 0.720804 0.793853 0.847356 0.879993 0.890962

0.40 0.000000 0.148771 0.293878 0.431749 0.558989 0.672465 0.769383 0.847356 0.904464 0.939301 0.951009

0.45 0.000000 0.154500 0.305197 0.448379 0.580520 0.698367 0.799117 0.879993 0.939301 0.975480 0.987640

0.50 0.000000 0.156427 0.309002 0.453968 0.587756 0.707072 0.808977 0.890962 0.951009 0.987640 0.999951

137
Table (3.8.2(b))
Temperature distribution in thin rectangular plate through implicit method

Temperature in thin rectangular plate u -^ at t = 3/400

X
0.0
0.0
0.000000
0.05 0.10
0.000000 0.000000
0.15
0.000000
0.20 0.25
0.000000 0.000000
0.30
0.000000
0.35
0.000000
0.40
0.000000
0.45
0.000000
0.50
0.000000

0.05 0.000000 0.024468 0.048334 0.071009 0.091936 0.110599 0.126539 0.139363 0.148756 0.154486 0.156411

0.10 0.000000 0.048364 0.095477 0.140270 0.181600 0.218476 0.249963 0.275295 0.293849 0.305167 0.308971

0.15 0.000000 0.071009 0.140270 0.206077 0.266809 0.320972 0.367231 0.404449 0.431707 0.448335 0.453923

0.20 0.000000 0.091936 0.181600 0.266809 0.345440 0.415565 0.475458 0.523643 0.558934 0.580463 0.587698

0.25 0.000000 0.110599 0.218476 0.320972 0.415565 0.499926 0,571977 0.629943 0.672399 0.698298 0.707002

0.30 0.000000 0.126539 0.249963 0.367231 0.475458 0.571977 0.654411 0.720733 0.769307 0.798938 0.808897

0.35 0.000000 0.139363 0.275295 0.404449 0.523643 0.629943 0.720733 0.793775 0.847272 0.879906 0.890874

0.40 0.000000 0.148754 0.293849 0.431707 0.558934 0.672399 0.769307 0.847272 0.904374 0.939208 0.950915

0.45 0.000000 0.154486 0.305167 0.448335 0.580463 0.698298 0.798938 0.879906 0.939208 0.975383 0.987542

0.50 0.000000 0.156411 0.308971 0.453923 0.587698 0.707002 0.808897 0.890874 0.950915 0.987542 0.999852

138
Table (3.8.2(c))
Temperature distribution in thin rectangular plate through implicit method

Temperature in thin rectangular plate u -> at t = 1/80

0.0 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50

0.0 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000

0.05 0.000000 0.024465 0.048329 0.071002 0.091927 0.110588 0.126527 0.139350 0.148741 0.154470 0.156395

0.10 0.000000 0.048329 0.095468 0.140256 0.181591 0.218454 0.249938 0.275268 0.293820 0.305137 0.308940

0.15 0.000000 0.071002 0.140256 0.206057 0.266783 0.320941 0.367195 0.404409 0.431664 0.448290 0.453878

0.20 0.000000 0.091927 0.181591 0.266729 0.345406 0.415524 0.475410 0.523591 0.558879 0.580405 0.587640

0.25 0.000000 0.110588 0.218454 0.320941 0.415524 0.499877 0.571920 0.629881 0.672332 0.698229 0.706932

0.30 0.000000 0.126527 0.249938 0.367195 0.475410 0.571920 0.654347 0.720661 0.769231 0.798859 0.808817

0.35 0.000000 0.139350 0.275268 0.404409 0.523591 0.629881 0.720661 0.793699 0.847188 0.879820 0.890786

0.40 0.000000 0.148751 0.293820 0.431664 0.558879 0.672332 0.769231 0.847188 0.904285 0.939116 0.950821

0.45 0.000000 0.154470 0.305137 0.448290 0.580405 0.698229 0.798859 0.879820 0.939116 0.975287 0.987444

0.50 0.000000 0.156395 0.308940 0.453878 0.587640 0.706932 0.808817 0.890786 0.950821 0.987444 0.999753

139
0.800000
0.700000
0.600000
0.500000
0.400000
0.300000
0.200000
0.100000
0.000000 ^
0.0 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50

Figure (3.8.2)
Temperature distribution in thin rectangular plate through cubic spline implicit method
(at y=0.25)

140
3.8 (C) DISCUSSION OF RESULTS :
The solutions of equation (3.7.3) obtained by implicit
as well as explicit solutions are comparing with exact
solutions as follows in table (3.8.3). Clearly the results are
accurate upto five digits of decimal places. Figure (3.8.3(a))
indicates the error analysis which compares the solutions
with spline solutions and figure (3.8.3(b)) gives the single
curve which shows that spline solutions are quite accurate
and reliable.

Table (3.8.3)

Error analysis

Temperature in thin rectangular plate u -^


at y = 0.25 fie t = 1 / 8 0

X USE USI UEXT UEXT-USE UEXT-USI

0.0 0.000000 0.000000 0.000000 0.000000 0.000000


0.05 0.110588 0.110588 0.110589 0.000001 0.000001
0.10 0.218454 0.218454 0.218454 0.000000 0.,000000
0.15 0.320942 0.320941 0.320941 0.000001 0.000000
0.20 0.415524 0.415524 0.415524 0.000000 0.000000
0.25 0.499876 0.499877 0.499877 0.000001 0.000000
0.30 0.571920 0.571920 0.571920 0.000000 0.000000
0.35 0.629881 0.629881 0.629881 0.000000 0.000000
0.40 0.672332 0.672332 0.672333 0.000001 0.000001
0.45 0.698229 0.698229 0.698228 0.000001 0.000001
0.50 0.706932 0.706932 0.706932 0.000000 0.000000

141
0.000001

0.000001

t 0.000001

O H#—USE-UEXT
0.000001
-^-USI-UEXT
UJ
0.000000

0.000000

0.000000 !" ' I

0.0 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50

X —

Figure (3.8.3(a))
Error Analysis (at y = 0.25 & t = 1/80)

142
0 800000

0.700000 „ ^ 1
0.600000
n
0.500000 n —•—USE
t 0.400000 n -•—USI
3
UEXT
0.300000

0.200000

0.100000
V
0 000000 F 1 : 1 i 1 1 1 1 r ""

0.0 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50

X —•

Figure (3.8.3(b))
Temperature distribution in thin rectangular plate through cubic spline implicit method
(att=l/80&y=0.25)

143
3.9 SPLINE FORMULA TO SOLVE HYPERBOLIC
PARTIAL DIFFERENTIAL EQUATION WITH ONE
SPACE VARIABLES :
The general form of hyperbolic PDE with one space
variable x a n d time variable t is given by

a^u/a^t = c^/a\/5x^ ; O<X<L, t>o ...(3.9.1)


with a Dirichilet b o u n d a r y conditions, namely
u(0, t) = 0
u(L, t) = 0 ...(3.9.2)
a n d two initial conditions at t = 0 (Cauchy conditions)
u(x, 0) = f(x)
u,ix,0) = g{x) ...(3.9.3)

In equation (3.9.1), c^ is a c o n s t a n t term, it d e p e n d s


u p o n some physical quantities in case of different problems.
Divide t h e region 0 < x < L into say n s u b - intervals
each of width Ax (= h) s u c h t h a t nAx = L. The s u b s c r i p t j
denotes time and i for the positions. The points of
subdivisions are Xj; i = 0 ( l ) n . Let Ujj denote the solution

of equation (3.9.1) at (ij)^*^ m e s h point. Discretize the left


h a n d side of PDE (3.9.1) by t h e central difference formula
like finite difference a n d right side by second derivative of
cubic spline S(x) i.e. S"(x^) at t h e (i, j ) * m e s h point, one can
get
( u , j ^ , - 2ui,j +Ui J.I) / (At)2 = c' (S:,j) ...(3.9.4)

where S-j = second derivative of cubic spline S(x) i. e. S"(Xj)

144
at time jAt s u b s t i t u t i n g t h e values of S" j from equation

(3.9.4) into (2.4.15a), t h e following relation is obtained.


(Ui-i,j+i+4uij^i+Ui+ij+i) =(2 + 6r2)u,.ij+(8-12r2)Uij + (2 + 6r2)Ui^i j
- [UMJ-l+4Uij.i+Ui+l j . l ]
...(3.9.5)
where r = cAt / h i=l(l)n-l
Above formula is k n o w n as cubic spline explicit
formula at to solve hyperbolic PDE of the form (3.9.1). It is
clear t h a t above formula is applied for all values of j > l .
However, for j = 0, it b e c o m e s

Ui-i,i+4ui,i+Ui^ii) =(2 + 6r2)Ui.io+(8-12r2)Uio+(2 + 6r^)Ui^io


- [Ui.i,.i+4ui.i+Ui+i .il

...(3.9.6)
where i = 1(1) n - 1
it involves the term Uj+j .j , Uj .^ a n d Uj.j .j which are

u n k n o w n s . To deal with t h e m , we consider the function u =


u(x, t) to be extended b a c k w a r d in time, the t e r m t = t.i m a k e
a good s e n s e . Most of t h e time, we get periodic functions for
u v e r s u s at a given point, we c a n consider zero time a s a n
arbitrary point at which we k n o w t h e value of u . So, to get
t h e values for the fictitious p o i n t s at t = t.i we u s e the initial
condition (initial velocity)
au/a = g(x) at t = o
By central difference approximation, we have
Ut(x,0) = (Uii -Ui.i)/2At = g(x.)

giving uj.i = uji - 2g(Xi)At at t = 0 only

i = 1(1) n - 1 ...(3.9.7)

145
The values of Uj .j ; i = 0(l)n a n d u s i n g initial and

b o u n d a r y conditions with equation (3.9.6) gives system of


(n - 1) s i m u l t a n e o u s linear e q u a t i o n s in (n - 1) u n k n o w n s .
The system h a s a tri-diagonal miatrix, which c a n be solved
by a n y well-known m e t h o d . After calculating the values of u
for j = 0, we apply equation (3.9.5) for j > l , we get (n - 1)
s i m u l t a n e o u s linear e q u a t i o n s in (n - 1) u n k n o w n s with
tri-diagonal matrix, again r < l is the required condition for
convergence and stability of this cubic spline explicit
method.
J u s t like the implicit s c h e m e for obtaining solution of
parabolic partial differential equations, we have implicit
scheme for hyperbolic differential equations. Implicit
s c h e m e is unconditionally stable i.e. stable for all t h e values
for r. In implicit m e t h o d t h e discretization of t h e differential
equation at gmy m e s h point (i, j) is done by replacing time
derivative by the central difference formula a s done in
explicit s c h e m e a n d t h e space derivative is replaced by
average of second derivatives of cubic spline S(x) at the

G-1)* a n d G + 1)* level i.e. (S-j.i+S;;J+I)/2, equation (3.9.1)


becomes,

(Ui,j,i - 2 u , j + u , j . i ) / ( A t ) ' = c^S[j.i+S[j,i)/2 ...(3.9.8)


The values of Sy+j obtained from equation (3.11.8) a n d

with t h e help of equation (2.4.15a) t h e following relation is


obtained.

146
= (3r2 -1) Ui,,^ j - (6r2 + 4) u,^.^ + (Sr^ -1) u^.,, j.^
+ 2(Ui^i j + 4ui^j + Ui.ij)

...(3.9.9)
where r = cAt/h ; i = l(l)n-l
The above equation (3.9.9) is k n o w n a s cubic spline
implicit formula to solve hyperbolic PDE of t h e form (3.9.1).
Like explicit s c h e m e , described a s above, the equation
(3.9.9) gives (n - 1) s i m u l t a n e o u s linear e q u a t i o n s in (n - 1)
u n k n o w n s with the coefficient matrix of tri-diagonal form.
For j = 0, here we c a n also u s e the initial condition in
similar m a n n e r described a s above.

3 . 1 0 THE FLOW OF ELECTRICITY IN THE


TRANSMISSION LINES :
The problem is already d i s c u s s e d in c h a p t e r - 3 in
section 3.2. In addition to t h a t , for higher frequency the
effect of R a n d G are negligible a n d e q u a t i o n s (3.2.4) and
(3.2.5) r e d u c e to

a^e/ax^ = (LQa^e/at^

i.e. a^u/ax^ = (LC)a^u/at^ ...(3.10.1)


where u s t a n d s for either i(x, t). L is the i n d u c t a n c e a n d C is
t h e capacitance per u n i t length of t h e t r a n s m i s s i o n line i.e.
cable. This equation is k n o w n a s hyperbolic PDE with one
space variable x a n d time variable t.

147
Let 1 = length of the t r a n s m i s s i o n line = 1 a n d u s i n g
following initial a n d b o u n d a r y conditions, t h e solution of
above equation (3.10.1) i.e. t h e c u r r e n t distribution is
obtained a s follows.
Dirichilet b o u n d a r y conditions :
u(0, t) = 0
u(l, t) = 0 ...(3.10.2)
Initial conditions (cauchy conditions at t = 0)
u(x, 0) = sin ;r X ; 0<x<1
U(x, 0) = 0 ...(3.10.3)

3.11(A) SPLINE SOLUTIONS WITH EXPLICIT METHOD :


Using initial a n d b o u n d a r y conditions described by
e q u a t i o n s (3.10.2) a n d (3.10.3), the solutions of equation
(3.10.1) are obtained by explicit formula given in the
e q u a t i o n (3.9.5).
Let the length of t h e t r a n s m i s s i o n line i.e. 1 = 1 divide
the region 0 < x < 1 into 10 s u b - intervals each of width h =
0 . 1 . LetL.C = 0 . 0 1 , a n d At = 0.01. These gives r = 0 . 0 1 .
Hence, 2+ 6r^= 2.0006 a n d 8 - H r ^ =7.9988
Having calculated the values of u^,•^ ; i = l(l)n from

e q u a t i o n (3.9.7) with initial conditions, s u b s t i t u t e t h e values


of Uj .1 a n d 2 + 6r^, 8-12r^ into t h e equation (3.9.6) we get,

For j = 0
i= 1

148
Uo,, + 4Uj J + U2,i

= {(2.0006)Uo,o+ (7.9988)ui 0 + (2.0006)U2,o}/2


= 1.823844
S i n c e UQ j = 0

4ui,i+ U2J = 1.823844


i =2 Ui 1+4u2,i+ U31 = 3.469157

i =3 U2J + 4u31 + U4 1 = 4.774886

i =4 U3 1 + 4u4 1 + U51 = 5.613215

i =5 U4 1 + 4u5 1 + U6,i = 5.902084


But U4 J and U(5j are s a m e d u e to symmetry of the

results
2U41+ 4u5i = 5.902084

The above five simultaneous equations in five


u n k n o w n s with in diagonal coefficient matrix c a n be easily
solved. Once the values of u are k n o w n at the first level of
time, t h e process c a n be repeated for second time level a n d
so on. It should be noted t h a t t h e n u m b e r of u n k n o w n s are
r e d u c e d b e c a u s e of the s y m m e t i y of the problem otherwise
we would have obtained nine s i m u l t a n e o u s e q u a t i o n s for
nine internal m e s h points. The r e s u l t s obtained by the
explicit m e t h o d are given in the table (3.11.1) a n d aire
plotted in t h e figures (3.11.1). D u e to the symmetry of t h e
problem, r e s u l t s are given for 0 < x < 0 . 5 only.

149
Table (3.11.1)
Current distribution in the cable through cubic spline
explicit method

Current Distribution in the cable -^


X 1 = 0.0 t = 0.01 t = 0.02 t = 0.03 t = 0.04 t = 0.05
0.0 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000
0.1 0.309017 0.309015 0.309011 0.309003 0.308992 0.308978
0.2 0.587785 0.587782 0.587773 0.587759 0.587738 0.587712
0.3 0.809017 0.809013 0.809000 0.808981 0.808953 0.808916
0.4 0.951057 0.951052 0.951038 0.951015 0.950982 0.950940
0.5 1.000000 0.999995 0.999978 0.999955 0.999920 0.999875

1.200000

1.000000
-^\ = 0.0

0.800000 - • - t = 0.01
t t = 0.02
0.600000
~4f-t = 0.03
0.400000 - * - t = 0.04
- • ~ t = 0.05
0.200000

0.000000
0.0 0.1 0.2 0.3 0.4 0.5

Figure (3.11.1)
Current distribution in the cable through cubic spline
explicit method

150
3 . 1 1 (B) SPLINE SOLUTIONS WITH IMPLICIT METHOD:
Using the s a m e initial a n d b o u n d a r y conditions a n d
t h e values of other p a r a m e t e r s d i s c u s s e d in section 3.11(A),
t h e solution of equation (3.11.1) is obtained by implicit
formula t h a t is given by the equation (3.9.9), derived a s
follows.
For j = 0,
i = 1 (0.9997)uo,i + (4.0006)ui 1 + (0.9997)u2,i = 1.823853
S i n c e UQ i = 0

(4.0006)uii+ (0.9997)u2,i = 1.823853

i =3 (0.9997)u2,i + (4.0006)u31 + (0.9997)u4 j = 4.774909

i =4 (0.9997)U3 1 + (4.0006)u4 j + (0.9997)u5 j = 5.613243

i =5 (0.9997)U4 1 + (4.0006)u5 j + (0.9997)u6,i = 5.902113

Since U4 j and U5 1 are s a m e ,

(1.9994)u4i+ (4.0006)u5i = 5.902113

The above five simultaneous equations in five


u n k n o w n s with coefficient m a t r i x tri-diagonal one, is solved
by a n y well-known m e t h o d . Like the explicit s c h e m e , once
t h e values of u are k n o w n at the first level of time, the
process c a n be repeated for second time level a n d so on.
Due to symmetry of the problem the n u m b e r s of u n k n o w n s
are r e d u c e d otherwise we get nine equations in nine
u n k n o w n s . The r e s u l t s obtained by the implicit m e t h o d are
given in t h e table (3.11.2) a n d are plotted in the figures
(3.11.2). Also d u e to the symmetry of the process r e s u l t s are
given for 0 < x < 0.5 only.

51
Table (3.11.2)
Current distribution in the cable through cubic spline
implicit method

Current distribution in the cable -^


X t=0.0 t = 0.01 t = 0.02 t = 0.03 t = 0.04 t = 0.05
0.0 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000
0.1 0.309017 0.309015 0.309011 0.309003 0.308992 0.308978
0.2 0.587785 0.587782 0.587773 0.587759 0.587738 0.587711
0.3 0.809017 0.809013 0.809000 0.808981 0.808953 0.808917
0.4 0.951057 0.951052 0.951038 0.951014 0.950981 0.950939
0.5 1.000000 0.999995 0.999978 0.999955 0.999920 0.999875

• ^ t = 0.0
•«-t = 0.01
t t = 0.02
•4f-1 = 0.03
• ^ t = 0.04
• ^ t = 0.05

0.000000
0.0 0.1

Figure (3.11.2)
Current distribution in the cable through cubic spline
implicit method

152
3.11 (C) DISCUSSION OF RESULTS :
The current distributions obtained by the solution of
equation (3.10.1) by explicit as well as implicit methods are
compared with the exact solutions at t = 0.04 in the table
(3.11.3). Figure (3.11.3) indicates the error analysis, which
compares the exact solutions with spline solutions, obtained
by both the methods, Explicit as well as Implicit, at t =
0.04. From the table (3.11.3) and figure (3.11.3), it is clear
that solutions obtained by both the methods are fairly good
and correct u p to five decimal places. Also implicit scheme
gives better results than the explicit one.

Table (3.11.3)
Error analysis
Current distribution u ->
X USE USl UEXT USE-UEXT USI-UEXT

0.0 0.000000 0.000000 0.000000 0.000000 0.000000


0.1 0.308992 0.308992 0.308992 0.000000 0.000000

0.2 0.587738 0.587738 0.587739 0.000001 0.000001


0.3 0.808953 0.808953 0.808953 0.000000 0.000000
0.4 0.950982 0.950981 0.950981 0.000001 0.000000
0.5 0.999920 0.999920 0.999921 0.000001 0.000001

[USE, USl and UEXT denotes cubic spline explicit, cubic


spline implicit and exact solutions]

:53
0.0000012

0.0000010

1 0.0000008
-•—USE-UEXT
2 0.0000006
-m— USI-UEXT
HI

0.0000004

0.0000002

0.0000000 0.0 0.1 0.2 0.3 0.4 0.5

Figure (3.11.3(a))
Error analysis (at t = 0.04)

154
1.200000

-•—USE
t •«—USI
UEXT

0.000000 A
0.0 0.1 0.2 0.3 0.4 0.5

Figure (3.11.2(b)

Current Distribution in t h e cable through cubic spline method


(at t = 0.04)

155
3 . 1 2 VIBRATING STRING PROBLEM :
Many problems in m a t h e m a t i c a l physics r e d u c e to t h e
solving of partial differential equations. The partial
differential e q u a t i o n s play a n i m p o r t a n t role in the theory of
elasticity, hydraulics and other some branches of
engineering. One of t h e most important problems in
m a t h e m a t i c a l physics is the vibration of a stretched string.
Simplicity a n d frequent occurrence in m a n y b r a n c h e s of
m a t h e m a t i c a l physics m a k e it classical example in t h e
theory of partial differential e q u a t i o n s .
Let u s consider a stretched string of length L fixed at
t h e end points. The problem here is to determine the
equation of motion which characterizes the position u(x, t)
of the string at time t after a n initial d i s t u r b a n c e is given.
In order to obtain a simple equation we m a k e the
following a s s u m p t i o n s .
(1) The string is flexible a n d elastic t h a t is t h e string
c a n n o t resist bending m o m e n t s a n d t h u s the tension
in t h e string is always in t h e direction of t h e t a n g e n t to
t h e existing profile of t h e string.
(2) There is n o elongation of a single segment of t h e string
a n d h e n c e by Hook's law t h e tension is c o n s t a n t .
(3) The weight of the string is small compared with t h e
tension in the string.
(4) The deflection is small c o m p a r e d with t h e length of t h e
string.
(5) The slope of the displaced string at amy point is small
c o m p a r e d with unity.

156
(6) There is only pure transverse vibration.

Figure (3.12.1)
String position and tension

Let T be the tension at the end points as shown in


figure (3.12.1). Consider the vibration of an elastic string of
length L in the vertical plane, the ends of which are fixed.
Take the origin at one fixed end, the x-axis along the length
of the string (when undisturbed) and the y-axis
perpendicular to it.
The displacement u of any point of the string is a
function of two variables, x its distance from o, and the time
t. To obtain the relation between u, x and t, take a small
element <St of the string at a distance x from the origin of
which the displacement by y at time t. Let m be the mass
per unit length of the string. Then considering vertical
displacement of these elements, its acceleration isS'^u / S t'^.
If ^ and ^ + S(^ be the angle made by the tangents at the
extremities of this elements (along which the tension t at

157
each point acts), with the x-axis then the vertical component
of the force to which this element is subjected is
T sm((f> + S<^) - Tsin^
and as ^ is smiall, we can write
tan ^ = sin ^, so these becomes
T tan{(/> + S(p) - Ttan^

Sa
T<^
x+Sx.

or
da
- X+&
& (3.12.1)
c5x

Thus for the vertical motion of this element, we have


the equation of motion as
"" /
'da' 'da' 11
^^u .&. X+& _ & . »jj
(m.&) ^& ...(3.12.2)
di^ dK
canceling ds. on both sides and taking the limiit of the
equation (3.12.2), we have
d\ d'u
m
dt' dx'
or
d^ d'u
where c^= T/m ...(3.12.3)
dt' dx'
T : Tension in a string
m : Mass per unit length

158
This is the PDE having the vibrations of the string a n d
is called t h e WAVE e q u a t i o n s , which is of HYPERBOLIC
TYPE.
S u p p o s e t h e string of length 2 meter fixed at its b o t h
e n d s , t h e initial velocity g(x) is to be t a k e n 0 a n d from t h e
b a s e d on m a t h e m a t i c a l physics it c a n be k n o w n t h a t t h e
initial displacement f(x) is sin(;zx/L). We also a s s u m e t h a t
c o n s t a n t C t a k e s value 1, t h e n t h e s e problem becomes a
initial b o u n d a r y value problem defined a s
Utt ^ "xx ; 0 < x < 2 ; t>0 ...(3.12.3)

u(x, 0) = sin 0<x<2 ...(3.12.3a)

ut(x,0) = 0 ; 0 < x < 2


u(0, t) =0 t>0 ...(3.12.3b)
u(2,t) = 0 1t > 0

3.13(A) SPLINE SOLUTION WITH EXPLICIT METHOD :


Using initial a n d b o u n d a r y conditions described by
equation (3.12.3a) a n d (3.12.3b), the solution of equation
(3.12.3) is obtained by explicit formula given in equation
(3.9.5).
Let t h e string of length 2 meter fixed at its b o t h e n d s

a n d divide t h e region 0 < x < 2 with h = - = 0.2 At = — = 0.05 .


5 20
These gives r = 0.25.
Hence 2 + 6r^ = 2.375 and S-^r^ = 7.25.

59
Having calculated the values of Uj .j ; i = l(l)n from

equation (3.9.7) with initial condition, s u b s t i t u t e t h e values

Uj .1 a n d 2 + 6r , 8-12r into equation (3.9.6), we get

For j = 0 a n d i = 1

Uo,i+ 4ui,i+ U2,i = i{(2.375)uo,o+(7.25)ui,o+(2.375)u2,o}


= 1.818177
Since UQ i = 0

4uii+U2j = 1.818177
i =2 uj 1+ 4u2,i+ U31 = 3.458383

i =3 U2,i+ 4u3 J + U41 = 4.760060

i =4 U31+4U41+U51 = 5.595787

Hence we get nine e q u a t i o n s in nine u n k n o w n s for


i = l , 2 , . . . , 1 0 £ind j= 1,2,3,4,5. The above nine s i m u l t a n e o u s
e q u a t i o n s in nine u n k n o w n s with in diagonal co-efficient
m a t r i x c a n be easily solved. Once the value of u are k n o w n
at first level of time, t h e process can be repeated for second
time level a n d so on. The r e s u l t s obtained by explicit
m e t h o d are given in table (3.13.1) a n d are plotted in figures
(3.13.1). Due to t h e symmetry of t h e problem r e s u l t s are
given for 0 < x < 1 only.

160
Table (3.13.1)
Displacement of vibrating string through cubic spline
explicit method

Displacement u(x, t)
X t=0.0 t= 0.05 t=0.1 t= 0.15 t= 0.20 t= 0.25
0.0 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000
0.2 0.309075 0.308058 0.305190 0.300430 0.2938 0.285341
0.4 0.587786 0.585962 0.580507 0.571453 0.558897 0.542938
0.6 0.809016 0.806508 0.799000 0.786838 0.769760 0.747802
0.8 0.951056 0.948108 0.939282 0.924632 0.904292 0.878408
1.0 1.000000 0.996899 0.987618 0.972214 0.950766 0.923410
1.2 0.951056 0.948108 0.939282 0.924632 0.904292 0.878408
1.4 0.809016 0.806508 0.799000 0.786838 0.769760 0.747802
1.6 0.587786 0.585962 0.580507 0.571453 0.558897 0.542938
1.8 0.309075 0.308058 0.305190 0.300430 0.2938 0.285341
2.0 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000

1.20OOOO

-^t=QO I
•<-t=Q05
t=0.1 j
-^t=Q15|
^K-t=Q20l
•*-t=Q25i

0.000000
GO 02 0.4 0.6 08 1.0 1.2 1.4 1.6 1.8 20
X—•

Figure (3.13.1)
Displacement of vibrating string through cubic spline
explicit method

161
3 . 1 3 (B) SPLINE SOLUTIONS WITH IMPLICIT METHOD :
Using t h e s a m e initial a n d b o u n d a r y conditions a n d
t h e v a l u e s of o t h e r p a r a m e t e r s d i s c u s s e d in section (3.13A)
t h e solution of e q u a t i o n (3.12.3) is o b t a i n e d by implicit
formula t h a t is given by t h e e q u a t i o n (3.9.9), derived a s
follows.
For j = 0
i= 1
(0.8125)uo,i+(4.375)ui,i+(0.8125)u2,i

= - { ^ ( " 2 , 0 + 4 u i 0 + Uo,o)}
= 1.823849
S i n c e UQ i = 0

(4.375)ui i+(0.8125)u2,i = 1.823849

i =2 (0.8125)uii+(4.375)u2,i+(0.8125)u3i= 3.469172

i=3 (0.8125)u2,i+(4.375)u3i+(0.8125)u4i= 4.774909

i =4 (0.8125)u3i+(4.375)u4i+(0.8125)u5i= 5.613242

i =5 (0.8125)u4i+(4.375)u5i+(0.8125)u6,i- 5.902113

i =6 (0.8125)u5,+(4.375)u6i+(0.8125)u7i= 5.613242

i =7 (0.8125)u6^i+(4.375)u7i+(0.8125)u8i= 4.774909

i =8 (0.8125)u7i+(4.375)u8i+(0.8125)u9i= 3.469172

i =9 (0.8125)u8i+(4.375)u9i+(0.8125)uio,i= 1.823849

S i n c e Ujo^i = 0

(0.8125)u8i+(4.375)u9i = 1.823849

T h e a b o v e s y s t e m of n i n e e q u a t i o n s i n n i n e u n k n o w n s
with co-efficient m a t r i x t r i - d i a g o n a l o n e , is solved b y

'62
any well known method. Like explicit scheme, once the
value of u are known at first level of time, the process can
be repeated for second time level and so on. The results
obtained by implicit method are given in the table (3.13.2)
and are plotted in figures (3.13.2). Due to the symmetry of
the process results are given for 0 < x < 1 only.

Table (3.13.2)
Displacement of vibrating string through cubic spline
implicit method

Displacement u(x, t)

X t=0.0 t= 0.05 t=0.1 t= 0.15 t= 0.20 t= 0.25

0.0 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000


0.2 0.309016 0.308058 0.305190 0.300430 0.2938 0.285341
0.4 0.587785 0.585962 0.580507 0.571453 0.558897 0.542938
0.6 0.809016 0.806508 0.799000 0.786838 0.769760 0.747802
0.8 0.951056 0.948108 0.939282 0.924632 0.904292 0.878408
1.0 1.000000 0.996899 0.987618 0.972214 0.950766 0.923410
1.2 0.951056 0.948108 0.939282 0.924632 0.904292 0.878408
1.4 0.809016 0.806508 0.799000 0.786838 0.769760 0.747802
1.6 0.587785 0.585962 0.580507 0.571453 0.558897 0.542938
1.8 0.309016 0.308058 0.305190 0.300430 0.2938 0.285341
2.0 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000

163
1.20000O

0000000
QO Q2 0.4 Q6 08 1.0 1.2 1.4 1.6 1.8 20

X •

Figure (3.13.2)
Displacement of vibrating string through cubic spline
implicit method

3 . 1 3 (C) DISCUSSION OF RESULTS :


The result presented in table (3.13.1), (3.13.2) is
obtained by explicit and implicit methods are compared with
the exact solutions at t = 0.15 in the table (3.13.3).
Figure (3.13.3(a)) indicates the error analysis, which
compares the exact solutions with spline solutions, obtained
by both the methods. Explicit as well as Implicit, at
t = 0.15. From the table (3.13.3) and figure (3.13.3(a)), it is
clear that solutions obtained by both the methods are fairly
good and correct u p to three decimal places. Figure
(3.13.3(b)) shows the single curve which gives accuracy of
spline solutions. Also implicit method gives better results
t h a n the explicit one.

164
Table 3.13.3
Error analysis

X USE USI UEX UEX-USE UEX-USI

0.0 0.000000 0.000000 0.000000 0.000000 0.000000


0.2 0.300404 0.300430 0.300477 0.000073 0.000047
0.4 0.571402 0.571453 0.571538 0.000136 0.000085
0.6 0.786468 0.786838 0.786659 0.000191 0.000179
0.8 0.924548 0.924632 0.904506 0.020042 0.020126
1.0 0.972129 0.972214 0.951056 0.021073 0.021158
1.2 0.924548 0.924632 0.904506 0.020042 0.020126
1.4 0.786468 0.786838 0.786659 0.000191 0.000179
1.6 0.571402 0.571453 0.571538 0.000136 0.000085
1.8 0.300404 0.300430 0.300477 0.000073 0.000047
2.0 0.000000 0.000000 0.000000 0.000000 0.000000

165
0.025000

0.020000

1 0.015000
o ••-UEX-USE
•m— UEX-USI
UJ 0.010000

0.005000

0.000000 i
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0

Figure (3.13.3(a))
Error Analysis (at t = 0.15)

166
1.200000

1.000000

I 0.800000
••-USE
S 0.600000 n n
-»—UEXT
UJ USI
0.400000
n n
0.200000

0.000000 A -fx

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0

Figure (3.13.3(b))
Displacement of vibrating string through cubic spline method
(at t = 0.15)

167
3 . 1 4 SPLINE FORMULA TO SOLVE HYPERBOLIC
PARTIAL DIFFERENTIAL EQUATION WITH TWO
SPACE VARIABLES :

The general form of hyperbolic PDE with two space


variables is a s follows:
a^u/a^t = c^{d\ldx^+d\ldy'^) ...(3.14.1)
with Dirichilet b o u n d a r y conditions namely,
u(0,y,t) = 0 ;0<y<a
u(x,0,t) = 0 ;0<x<b ...(3.14.2)
a n d initial conditions at t = 0 (Cauchy conditions)
u(x, y, 0) = f(x, y)
a i / a = (x,y,0) = g(x,y) ...(3.14.3)

In e q u a t i o n (3.14.1), c is a c o n s t a n t term, it d e p e n d s
u p o n some physical quantities in case of different types of
problems.
Divide t h e region 0 < x < a into say n s u b - intervals
e a c h of width Ax s u c h t h a t nAx = a , 0 < y < b into m s u b -
intervals e a c h of width Ay s u c h t h a t nAx = b .
The s u b s c r i p t k d e n o t e s t h e time a n d (i, j) denote t h e
position. For simplicity consider s q u a r e region i.e. a = b = L
t h e length of the region, m = n a n d Ax = Ay = h (say). The
p o i n t s of subdivisions are(xj,yj), i, j = 0(l)n.

For explicit s c h e m e , the formula is obtained in similar


m a n n e r a s d i s c u s s e d in section (3.9), by replacing the right
side by twice t h e second derivative of cubic spline S(x) a t a t
the (i, j ) * m e s h point i.e. 2S"(Xjj) Let Uj j , we get

168
(Uwj,k+i+4u,j,k.i+Ui.,,j,k+,)

= (2 + 12r^)U;,,.,+ (8-24r^)u,,^,,+ (2 + 12r^)u,,,,,


- [ U w j , k - i + 4 U i j , . , + Uj.,j,., ]

...(3.14.4)
where r = cAt/h i, j = 1(1) n - 1
The values of u^j .j's i j = l ( l ) n - l are obtaiined from

following relation
U i . , = u . j , - 2g(x.,yj)At ...(3.14.5)

The equation (3.14.4) is k n o w n a s cubic spline explicit


formula to solve hyperbolic type PDE of the form given by
t h e equation (3.14.1). The values obtained from equation
(3.14.5) u s i n g initial a n d b o u n d a r y conditions, the equation
(3.14.4) gives system of (n-1) x (n-1) simultaneous
e q u a t i o n s in (n-1) x (n-1) u n k n o w n s . After calculating the
r e s u l t s for k = 0; t h e r e s u l t s for k = 1, 2, are obtained

in similar m a n n e r . r<- is t h e required condition for

convergence a n d stability of t h i s explicit m e t h o d .


Likewise implicit s c h e m e for obtaining solution of
parabolic PDE with one space variable, we have implicit
s c h e m e to solve hyperbolic differential equation with two
space variables. This implicit s c h e m e is unconditionally
stable. As similar t e c h n i q u e d i s c u s s e d in section (3.9) t h e
cubic spline implicit formula to solve equation (3.14.1) is a s
follows, by replacing t h e right side of equation (3.14.1) i.e.
space derivatives by (S|'j., +S^j^,)

169
,2^„ , /^ , io,„2>

+ 2[Ui^,j, + 4Uij,+Uj.,j.J

...(3.14.6)
where r = cAt/h ; i,j = l ( l ) n - l

D u e to similar d i s c u s s i o n a s above also equation


(3.14.6) give (n - 1) x (n - 1) s i m u l t a n e o u s e q u a t i o n s in
(n - 1) X (n - 1) u n k n o w n s . After calculating r e s u l t s for
(k + 1)* time level, the r e s u l t s for (k + 2)''' time level are
obtained in similar m a n n e r by repeating the process.

3 . 1 5 THE PROBLEM OF VIBRATING MEMBRANE :

The problem of vibrating m e m b r a n e is well-known to


r e s e a r c h e r s . A s s u m e t h a t m e m b r a n e is tightly stretched a n d
h o m o g e n e o u s i.e. its m a s s per u n i t a r e a is c o n s t a n t , it is
perfectly flexible a n d is so t h i n t h a t it offers no resistance
for bending. The m e m b r a n e is stretched a n d t h e n fixed
along its entire b o u n d a r y in t h e xy - plane a n d the tension T
c a u s e d by stretching the m e m b r a n e is the s a m e at every
point in all directions a n d does not c h a n g e d u r i n g t h e
motion. The deflection u(x, y, t) of the m e m b r a n e d u r i n g t h e
motion is small compared to the size of t h e m e m b r a n e .

Consider t h e forces acting on a small portion of t h e


m e m b r a n e a s s h o w n in figure (3.15.1). Since t h e deflection
of t h e m e m b r a n e a n d the angles of inclination are small the

170
sides of t h e portion m a y be t a k e n approximately equal to
Ax and Ay.

TAX

0
x +Ax

Figure (3.15.1)
Vibrating m e m b r a n e
The tension T is the force per u n i t length. Forces
acting on the edges are TAx a n d TAy which could be t a k e n
t a n g e n t to the m e m b r a n e since the m e m b r a n e is perfectly
elastic. Let the force TAy m a k e angles aand^ with t h e
horizontal on t h e opposite edges of the m e m b r a n e since the
m e m b r a n e is perfectly elastic.
The r e s u l t a n t verticad c o m p o n e n t of force d u e to TAy is
therefore,
" (TAy) sin/? - (TAy) sin a
(TAy) tan y^ - (TAy) tan a
= TAy {(du/dx)^,,^ -(du/dx\}
= TAy ( a u ' / a x ' ) A x
u p to a first order approximation. Note t h a t sine have
replaced by t a n g e n t b e c a u s e t h e angles a and /9 are small.

171
Similarly, t h e forces TAx acting on t h e edges of length
Ax c a n be s h o w n to have the vertical component.
TAx {^\l^y^)^y
If m be t h e m a s s per u n i t a r e a of the m e m b r a n e , by
Newton's second law of motion
(mAxAy) d\ I di^ = 1 (a^u / 5x^ + d'^xx I ay^)AxAy
=> a ^ u / a ^ = c^(a^u/ax^ + a^u/ay^) where c^= T / m
...(3.15.1)
The above equation (3.15.1) is hyperbolic PDE in two
space variables x a n d y, a n d time variable t. The solution fro
t h e case of a r e c t a n g u l a r m e m b r a n e is given below by cubic
spline explicit a s well a s implicit m e t h o d u s i n g following
initial a n d b o u n d a r y conditions.
Dirichilet b o u n d a r y conditions :

"<°-^''' = ° °-'^"' ...(3.15.2)


u(x,0,t) = 0 0<x<t
Initial conditions at t = 0 (Cauchy condition)
u(x, y, 0) = f(x, y) = sin;r x sin;r y ; 0 < x, y < 1
u(x,y,0) = 0 ^ ^

3.16(A) SPLINE SOLUTIONS WITH EXPLICIT METHOD :


Using initial a n d b o u n d a r y conditions are described by
e q u a t i o n s (3.15.2) a n d (3.15.3), t h e solution of equation
(3.15.1) is obtained t h r o u g h explicit forraiula given by the
equation (3.14.4) a s follows.
Let t h e region 0 < x < l and 0 < y < l be divided into 10
sub-intervals, e a c h of length h = 0 . 1 . Let c = 0.1 and
At = 0.01, t h e s e gives r = 0 . 0 1 . After calculating the values

172
of Uj j.j, i, j = l(l)n form equation (3.14.5), u s i n g initial

condition, s u b s t i t u t e the values of u^ j . j , 2 + 12r^ a n d 8-24r'^

into equation (3.14,4) one gets.


For k = 0
For j = 1
i= 1 Uo,i,i+4u,i,+U2,,,i = 0.563599

Since Uo,i,i = 0

4uu,i+U2,i,i =0.563599

i =2 Ui,i,i+4u2,i,i+U3,i^i =1.072029

i =3 U2,i,i+4u3ii+U4ii = 1.475521

i=4 U3,,,i+4u4,i^,+U5ii =1.734579

i =5 U4,i^i+4u5ii+U6ii =1.873844

Since ^4j,i~ "6,i,i (Due to symmetry of the


problem)
2u4,i,i+4u5,i,i =1.823844

In similar m a n n e r calculate t h e equation j = 1(1)9 one


gets 9 x 9 simultaneous linear equations in 9 x 9
u n k n o w n s . This system of e q u a t i o n s c a n be solved by a n y
well-known m e t h o d . Once the values of u are k n o w n at the
first level of time, t h e process c a n be repeated for second
time level a n d so on. The r e s u l t s obtained by explicit
m e t h o d are given a s follows. D u e to the symmetry of the
problem, r e s u l t s are given for 0 < y < 0 . 5 , 0 < x < 0 . 5 in t h e table
(3.16.1(a)) a n d (3.16.1b). The r e s u l t s obtained at y = 0.1 are
plotted in t h e figures (3.16.1),

173
Table (3.16.1(a))

Velocity distribution in rectangular vibrating membrane


through cubic spline explicit method u-> at t=0.01

0.0 0.1 0.2 0.3 0.4 0.5

0.0 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000

0.1 0.000000 0.095491 0.181634 0.249998 0.293890 0.309014

0.2 0.000000 0.181634 0.345488 0.475523 0.559012 0.587779

0.3 0.000000 0.249998 0.475523 0.654502 0.769413 0.809009

0.4 0.000000 0.293890 0.559012 0.769413 0.904499 0.951047

0.5 0.000000 0.309014 0.587779 0.809009 0.951047 0.999990

Table (3.16.1(b))

Velocity distribution in rectangular vibrating membrane


through cubic spline explicit method u -^ at t=0.03

0.0 0.1 0.2 0.3 0.4 0.5

0.0 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000

0.1 0.000000 0.095483 0.181619 0.249978 0.293867 0.308989

0.2 0.000000 0.181619 0.345461 0.475485 0.558967 0.587732

0.3 0.000000 0.249978 0.475485 0.654450 0.769352 0.808944

0.4 0.000000 0.293867 0.558967 0.769352 0.904427 0.950917

0.5 0.000000 0.308989 0.587732 0.808944 0.950971 0.999910

174
Figure (3.16.1)
V e l o c i t y distribution in rectangular m e m b r a n e through
cubic spline explicit m e t h o d (at y = 0.1)

3.16(B) SPLINE SOLUTIONS WITH IMPLICIT METHOD :

In similar mginner d i s c u s s e d a s earlier, u s i n g initial


a n d b o u n d a r y condition, at At = 0.01, r = 0 . 0 1 , t h e solution
of equation (3.14.1) is obtained by u s i n g cubic spline
implicit formula given by equation (3.14.6) a s follows.
For k = 0
For j = 1
i = 1 (4.0012)u, , , + (0.9994)U2,J,, = 0.563602

175
i = 2 {0.9994)ui 1 i+(4.0012)u2,i,i+(0.9994)u3, 1 =1.072034

i = 3 (0.9994)u2jj +(4.0012)u3ii +(0.9994)U411 = 1.475528

i=4 (0.9994)u3i 1+(4.0012) U4 11+(0.9994) U51 J =1.734587

i = 5 (1.9988)U4 1 1 +(4.0012)u5 1 1 = 1.823853

Like explicit s c h e m e d i s c u s s e d in section 3.17(A), one


gets system of 9 x 9 s i m u l t a n e o u s linear e q u a t i o n s in 9 x 9
u n k n o w n s . Once t h e values of u are obtained at t h e first
time level, the values for second time level are obtained in
similar m a n n e r by repeating t h e process. Results obtained
by cubic spline implicit m e t h o d are given at different times
in the tables (3.16.2(a)) a n d (3.16.2(b)). The r e s u l t s are
plotted in t h e figures (3.17.2) a t y = 0 . 1 .

Table (3.16.2(a))

V e l o c i t y distribution in rectangular vibrating m e m b r a n e


through cubic spline implicit m e t h o d u ^^ at t = 0 . 0 1

X 0.0 0.1 0.2 0.3 0.4 0.5

0.0 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000

0.1 0.000000 0.095491 0.181634 0.249998 0.293890 0.309014

0.2 0.000000 0.181634 0.345488 0.475523 0.559011 0.587779

0.3 0.000000 0.249998 0.475523 0.654502 0.769413 0.809009

0.4 0.000000 0.293890 0.559011 0.769413 0.904499 0.951047

0.5 0.000000 0.309014 0.587779 0.809009 0.951047 0.999990

176
Table (3.16.2(b))

Velocity distribution in rectangular vibrating membrane


through cubic spline implicit method u ->at t = 0.03

0.0 0.10 0.20 0.30 0.40 0.50

0.0 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000

0.1 0.000000 0.095483 0.181619 0.249978 0.293866 0.308989

0.2 0.000000 0.181619 0.345461 0.475485 0.558967 0.587732

0.3 0.000000 0.249978 0.475485 0.654450 0.769352 0.808944

0.4 0.000000 0.293866 0.558967 0.769352 0.904427 0.950917

0.5 0.000000 0.308989 0.587732 0.808944 0.950971 0.999910

Figure (3.16.2)
Velocity distribution in rectangular membrane through
cubic spline implicit method (at y = 0.1)

177
3.16 (C) DISCUSSIONS OF RESULTS :
The velocity distribution obtained by the solution of
equation (3.14.1) by cubic spline explicit as well as implicit
methods are compared with exact solutions at y = 0.1 for
t = 0.03 in the table (3.16.3). From the table it is clear that
the solutions obtained by both these methods are accurate
u p to five digits of decimal places. Figure (3.16.3(a))
indicates the error analysis which compares the exact
solutions with spline solutions obtained by both the
methods explicit as well as implicit at t = 0.03 clearly from
the figure (3.16.3(b)), the spline solutions are quite accurate
and reliable.
Table (3.16.3)
Error Analysis
Velocity distribution u -> y = 0.1 & t = 0.03

X USB USI UEXT USE-UEXT USI-UEXT

0.0 0.000000 0.000000 0.000000 0.000000 0.000000


0.1 0.095483 0.095483 0.095483 0.000000 0.000000

0.2 0.181619 0.181619 0.181619 0.000000 0.000000


0.3 0.249978 0.249978 0.249978 0.000000 0.000000
0.4 0.293867 0.293866 0.293866 0.000001 0.000000
0.5 0.308989 0.308989 0.308989 0.000000 0.000000

[USE, USI and UEXT denotes cubic spline explicit, cubic


spline implicit and exact solutions]

178
0.0000012

0.0000010

0.0000008

O -•— USE-UEXT
0.0000006
•m— USI-UEXT
m
0.0000004

0.0000002

0.0000000
0.0 0.1 0.2 0.3 0.4 0.5

Figure (3.16.3(a))
Error Analysis (at t = 0.03, y = 0.1)

179
0.000000 ri
QO 0.1 0.2 0.3 0.4 05
X—•

Figure (3.16.3(b))
Velocity distribution in rectangular membrane through cubic spline implicit method
(aty = 0.1, t = 0.03)

180
3.17 CONCLUSION:
In this chapter the method of spline collocation is
applied to linear PDE of parabolic and hyperbolic type with
one and two space variable. The problem describing flow of
electricity in a cable of transmission line, heat conduction
in a rod ,heat flow in thin rectangular plate, vibrating string,
vibrating membrane are discussed briefly. The governing
equations are parabolic PDE in one and two space variables
as well as hyperbolic PDE in one and two space variables.
All the equations are solved by spline explicit and spline
im^plicit scheme. The results are given in tabular form.
Comparison of the results shows that spline solutions are
quite reliable. The figure provides a pictorial evidence of a
good agreement of the curves presenting actual as well as
approximate solutions. Results are accurate u p to five digits
of decimal place in most of the case.
In all problems, reducing the length of sub-interval
gives more closed results. An ultimate conclusion is drawn
from this work is that, both the method of spline collocation
gives accurate results with compact computations. From the
compairison of results and figure it is conclude that, imiplicit
scheme provide better convergence than explicit scheme.

181

You might also like