Docsity
Docsity

Prepare for your exams
Prepare for your exams

Study with the several resources on Docsity


Earn points to download
Earn points to download

Earn points by helping other students or get them with a premium plan


Guidelines and tips
Guidelines and tips

Numerical Methods: Solving Diff. Equations & Eigenvalue Problems with Finite Differences, Assignments of Aerospace Engineering

A problem set for the university course ame 521: numerical methods, taught by j. M. Powers in the fall of 1999. The problem set includes various tasks related to solving differential equations and eigenvalue problems using finite difference schemes. Topics covered include finding the thermal resistance of an ice slab, analyzing the necessary and sufficient conditions for bounded solutions of a difference equation, and calculating eigenvalues and eigenfunctions of a finite difference analog of an eigenvalue problem.

Typology: Assignments

Pre 2010

Uploaded on 09/17/2009

koofers-user-wbh
koofers-user-wbh 🇺🇸

10 documents

1 / 2

Toggle sidebar

This page cannot be seen from the preview

Don't miss anything!

bg1
AME 521: Numerical Methods
FALL 1999, J. M. Powers. Prepared by S. V. Shepel
Problem Set 2
Due 7 October 1999
1. The differential equation describing the ice build-up (on a lake for example)
is given by
yd
2
y
dt
2
+2
dy
dt
2
+4
L
1+
y
2+
y
dy
dt
0
4
L
2+
y
=0
where
y
is the thermal resistance of the ice slab (thickness divided by conductiv-
ity),
L
is a dimensionless latent heat for which 0.75 is a reasonable value, and
t
is the nondimensional time. Appropriate initial conditions are
y
(0)=0
:
02
and
y
0
(0) = 0
:
5
. Solve the problem for all values of
t
until
y
=2
by using finite
difference schemes of first and second order.
In the case of a problem when the analytical solution does not exist, one can
use a very fine mesh and/or a higher order method to obtain a solution of high
accuracy. Obtain such a solution, and use it to verify convergence properties of
both the 1st and 2nd schemes by measuring the norm of the error. Plot the error
norm vs. discretization time interval on a logarithmic scale.
2. (a) Find the necessary and sufficient conditions on the coefficients
a
,
b
,
and
c
of the difference equation
ay
n
0
1
+
by
n
+
cy
n
+1
=0
;n
=0
;
6
1
;
6
2
;
...
so that the solutions of it are bounded. Find the explicit form of all the possible
solutions.
(b) Consider the eigenvalue problem
y
00
=
y ; y
(0) =
y
(1)=0
(1)
Find the eigenvalues
and eigenfunctions analytically.
Now consider a finite difference analog of the problem given by
1
h
2
(
y
n
0
1
0
2
y
n
+
y
n
+1
)=
y
n
;
y
0
=
y
N
=0
;Nh
=1
(2)
where
h
is the discretization interval and
N
is the number of the intervals.
Find analytically all the eigenfunctions
and the eigenvalues
of the
difference equation (2). How many eigenvalues can you find? For
N
=10
,
calculate the first 10 eigenvalues and eigenfunctions of Eq. (2) and compare them
1
pf2

Partial preview of the text

Download Numerical Methods: Solving Diff. Equations & Eigenvalue Problems with Finite Differences and more Assignments Aerospace Engineering in PDF only on Docsity!

AME 521: Numerical Methods FALL 1999, J. M. Powers. Prepared by S. V. Shepel Problem Set 2 Due 7 October 1999

1. The differential equation describing the ice build-up (on a lake for example) is given by

y

d^2 y dt^2

dy dt

+ 4L

1 + y 2 + y

dy dt

4 L

2 + y

where y is the thermal resistance of the ice slab (thickness divided by conductiv- ity), L is a dimensionless latent heat for which 0.75 is a reasonable value, and t is the nondimensional time. Appropriate initial conditions are y(0) = 0: 02 and y^0 (0) = 0: 5. Solve the problem for all values of t until y = 2 by using finite difference schemes of first and second order. In the case of a problem when the analytical solution does not exist, one can use a very fine mesh and/or a higher order method to obtain a solution of high accuracy. Obtain such a solution, and use it to verify convergence properties of both the 1st and 2nd schemes by measuring the norm of the error. Plot the error norm vs. discretization time interval on a logarithmic scale.

2. (a) Find the necessary and sufficient conditions on the coefficients a, b, and c of the difference equation

ayn 1 + byn + cyn+1 = 0 ; n = 0;  1 ;  2 ;...

so that the solutions of it are bounded. Find the explicit form of all the possible solutions.

(b) Consider the eigenvalue problem

y^00 = y ; y(0) = y(1) = 0 (1)

Find the eigenvalues  and eigenfunctions analytically.

Now consider a finite difference analog of the problem given by 1 h^2

(yn 1 2 yn + yn+1) = yn ; y 0 = yN = 0 ; Nh = 1

where h is the discretization interval and N is the number of the intervals.

Find analytically all the eigenfunctions and the eigenvalues  of the difference equation (2). How many eigenvalues can you find? For N = 10, calculate the first 10 eigenvalues and eigenfunctions of Eq. (2) and compare them

1

with the exact solution. Explain the results. Suggest a method to distinguish the correct eigenvalues.

3. (a) Check the following finite difference approximations to y^0 = f(t; y) for consistency and stability. For the consistent approximations, give the truncation error.

i. yn+1 = 3yn 2 yn 1 + h 2 (fn+1 + 2fn + fn 1 ) ii. yn+1 = 12 (yn + yn 1 ) + 34 h (3fn fn 1 ) iii yn+1 = yn 3 + 43 h (2fn fn 1 + 2fn 2 ) (b) Divide the segment x 2 [0; 1] over N parts x 0 = 0; x 1 ; x 2 ; :::; xN = 1, so that

xn+1 xn xn xn 1

= q

where q=const. Show whether it is possible to solve the following problem

u^0 = f(x; u) ; u(0) = 1

by using the following numerical scheme

u(xn+1) u(xn) xn+1 xn

f(xn; un) = 0 ; u(x 0 ) = 1

4. Integrate the system of equations

y 10 = 0 : 015 y 1 1000 y 1 y 3 y 2 y 3 y 20 = 3000 y 2 y 3 y 30 = 0 : 015 y 1 1000 y 1 y 3 3000 y 2 y 3

with initial conditions y 1 (0) = 1, y 2 (0) = 1, y 3 (0) = 0 up to t = 50 so that the local error in any variable is at most 10 ^4.

5. The problem of laminar convection over a hot vertical plate can be reduced by introducing a self-similar variable to the following form

f^000 2(f^0 )^2 + 3ff^00 +  = 0 ^00 + 3 Pr f^0 = 0

where f(0) = 0 , f^0 (0) = 0, (0) = 1, f^0 ( 1 ) = 0, ( 1 ) = 0. Here  1 is the effective boundary layer thickness, which is not known a priori and has to be found. Solve the problem, and plot velocity f^0 () and temperature () profiles. Consider two cases: air Pr = 0: 72 and liquid lead Pr = 0: 024.

2