Sei sulla pagina 1di 6

Mert Ozbayram math219

Assignment week12 due 08/23/2016 at 11:59pm EEST

This problem set covers material from Chapter 10.


Boundary value problems.
Eigenfunctions.
Separation of variables.
Fourier series.
Heat equation.

Correct Answers:
1. (1 point) Solve the following differential equation with the [(2*n+1)*pi/8]2
given boundary conditions. sin((2*n+1)*pi/8*x)
- If there are infinitely many solutions, use c for any undeter-
mined constants. 3. (1 point) In your answers below, for the variable type
d
- If there are no solutions, write No Solution. the word lambda; for the derivative dx X(x) type X ; for the
- Write answers as functions of x (i.e. y = y(x)). d2
double derivative dx2
X(x) type X ; etc.

y00 + 25y = 0 Separate variables in the following partial differential equa-


A) Boundary conditions: y(0) = 1 tion for u(x,t):
 
6
y = 1 xutt tuxt xu = 0
5
y=

B) Boundary conditions: y(0) = 1 = =


 
9 DE for X(x): =0
y = 1
5
y= DE for T (t) : =0
C) Boundary conditions: y(0) = 1 (Simplify your answers so that the highest derivative in each
  equation is positive.)
17 Correct Answers:
y = 1
10 -[X/(x*X)] = -[(T-T)/t*T] = \(\lambda\)
y=
<table border=0 cellspacing=0 style=margin-left:auto;m
Correct Answers: <tr><td>X-lambda*x*X\( = 0\)</td></tr>
DNE <tr><td>T-T-lambda*t*T\( = 0\)</td></tr>
1*cos(5*x)+c*sin(5*x) </table>
1*cos(5*x)+(-1)*sin(5*x)
2. (1 point) Find the eigenvalues and eigenfunctions for the
following boundary value problem (with > 0).

y00 + y = 0 with y(0) = 0, y0 (4) = 0. 4. (1 point) (


Eigenvalues: n = 0 for 0 < x < 5,
Let f (x) =
Eigenfunctions: yn = 7x for 5 x < 7.
Notation: Your answers should involve n and x.
Compute the Fourier cosine coefficients for f (x).

A0 =
If you dont get this in 2 tries, you can get a hint.
Hint: When computing eigenvalues, the following two for- An =
mulas may be useful: A0
Give values for the Fourier cosine series C(x) = +
sin() = 0 when = n.  n  2
(2n + 1) An cos 7 t .
cos() = 0 when = .
2 n=1
1
C(5) = 0 = X(0) =

C(6) = 0 = X(6) =

C(8) = So X(x) = which means u(x,t) =


Correct Answers:
0.571429 We can ignore this case.
(-1)n*[4/(n*pi)*sin(2*n*pi/7)-14/[(n*pi)2]*[1-cos(2*n*pi/7)]]
1 Case 2: = 2 (In your answers below use gamma
1 instead of lambda)
1
X(x) =
5. (1 point) In your answers below, for the variable type
the word lambda, for type the word gamma; otherwise treat
Plugging the boundary values into this formula gives
these as you would any other variable.
0 = X(0) =
We will solve the heat equation
0 = X(6) =
ut = 3 uxx , 0 < x < 6, t 0
with boundary/initial conditions: So X(x) = which means u(x,t) =
(
u(0,t) = 0, 2, 0 < x 3 We can ingore this case.
and u(x, 0) =
u(6,t) = 0, 0, 3 < x < 6
Case 3: = 2 (In your answers below use gamma
This models temperature in a thin rod of length L = 6 with instead of lambda)
thermal diffusivity = 3 where the temperature at the ends is
fixed at 0 and the initial temperature distribution is u(x, 0). X(x) =
For extra practice we will solve this problem from scratch. Plugging in the boundary values into this formula
Separate Variables. gives
Assume u(x,t) = X(x) T (t) and split the PDE into two differ-
ential equations, one with X and one with T . 0 = X(0) =
= =
(Notation: Write X and T for derivatives. Place all constants 0 = X(6) =
in the differential equation with T).
DE for X(x): =0
Boundary conditions for X(x): Which leads us to the eigenvalues n =
and eigenfunctions Xn (x) =

(Notation: Eigenfunctions should not include any


constants a or b.)

(Enter boundary equations: e.g. X 0 (0) = 10) Solve for T (t).


Plug the eigenvalues n = 2n from Case 3 into the differential
DE for T (t): =0 equation for T (t) and solve:
Find Eigenfunctions for X(x).
The problem splits into cases based on the sign of . Tn (t) =
(Notation: For the cases below, use constants a and b)
Case 1: = 0 (Notation: use c for the unknown constant.)
Combining all of the Xn and Tn we get that
X(x) =
u(x,t) = Bn
Plugging the boundary values into this formula gives n=1
where Bn are unknown constants.
2
Fourier Coefficients. (
We compute Bn by plugging t = 0 into the formula for u(x,t) ux (0,t) = 0, 0, 0 < x < 1
and setting equal to the initial heat distribution given in the and u(x, 0) =
ux (2,t) = 0, 4, 1 x < 2
problem.
This models the temperature in a thin rod of length L = 2
with thermal diffusivity = 2 where the no heat is gained or
(

2, 0 < x 3
u(x, 0) = Bn = lost through the ends of the rod (insulated ends) and the initial
n=1 0, 3 < x < 6
temperature distribution is u(x, 0).
So the Bn are Fourier coefficients.

2
Z 3 For extra practice we will solve this problem from scratch.
Bn = dx Separate Variables.
6 0
Assume u(x,t) = X(x) T (t) and split the PDE into two differ-
= ential equations, one with X and one with T .
Remember that cos(n) = (1)n . = =
Correct Answers: (Notation: Write X and T for derivatives. Place all constants
X/X = T/(3*T) = \(-\lambda\) in the differential equation with T).
X+lambda*X DE for X(x): =0
X(0)=0 Boundary conditions for X(x):
X(6)=0
T+3*lambda*T
<table border=0 cellspacing=0 style=margin-left:auto;margin-right:auto;>
<tr><td>\(X(x) = \) a+b*x</td></tr>
<tr><td>\(X(0) = \) a</td></tr>
<tr><td>\(X(6) = \) a+b*6</td></tr>
</table> (Enter boundary equations: e.g. X 0 (0) = 10)
0
0 DE for T (t): =0
<table border=0 cellspacing=0 style=margin-left:auto;margin-right:auto;>
Find Eigenfunctions for X(x).
<tr><td>\(X(x) = \) a*e(gamma*x)+b*e(-gamma*x)</td></tr>
The problem splits into cases based on the sign of .
<tr><td>\(X(0) = \) a+b</td></tr>
(Notation: For the cases below, use constants a and b)
<tr><td>\(X(6) = \) a*e(gamma*6)+b*e(-gamma*6)</td></tr>
</table> Case 1: = 0
0
0 X(x) =
<table border=0 cellspacing=0 style=margin-left:auto;margin-right:auto;>
<tr><td>\(X(x) = \) a*cos(gamma*x)+b*sin(gamma*x)</td></tr>
Plugging the boundary values into this formula gives
<tr><td>\(X(0) = \) a</td></tr>
<tr><td>\(X(6) = \) a*cos(gamma*6)+b*sin(gamma*6)</td></tr>0 = X 0 (0) =
</table>
n*pi/6
sin(n*pi*x/6)
0 = X 0 (2) =
c*e[-3*(n*pi/6)2*t]
e[-3*(n*pi/6)2*t]*sin(n*pi*x/6) So X(x) = .
sin(n*pi*x/6)
2*sin(n*pi*x/6) We will deal with this case using A0 later....
-4/(n*pi)*[cos(n*pi/2)-1]
Case 2: = 2 (In your answers below use gamma
instead of lambda)
6. (1 point) In your answers below, for the variable type
the word lambda, for type the word gamma; otherwise treat X(x) =
these as you would any other variable.
Plugging the boundary values into this formula gives
We will solve the heat equation
0 = X 0 (0) =
ut = 2 uxx , 0 < x < 2, t 0
0 = X 0 (2) =
with boundary/initial conditions:
3
So X(x) = <table border=0 cellspacing=0 style=margin-left:auto;m
<tr><td>\(X(x) = \) a+b*x</td></tr>
We can ingore this case. <tr><td>\(X(0) = \) b</td></tr>
<tr><td>\(X(2) = \) b</td></tr>
Case 3: = 2 (In your answers below use gamma </table>
a
instead of lambda)
<table border=0 cellspacing=0 style=margin-left:auto;m
<tr><td>\(X(x) = \) a*e(gamma*x)+b*e(-gamma*x)</td><
X(x) = <tr><td>\(X(0) = \) gamma*(a-b)</td></tr>
<tr><td>\(X(2) = \) gamma*[a*e(gamma*2)-b*e(-gamma*
Plugging the boundary values into this formula gives </table>
0
0 = X 0 (0) = <table border=0 cellspacing=0 style=margin-left:auto;m
<tr><td>\(X(x) = \) a*cos(gamma*x)+b*sin(gamma*x)</td>
0 = X 0 (2) = <tr><td>\(X(0) = \) gamma*b</td></tr>
<tr><td>\(X(2) = \) gamma*[b*cos(gamma*2)-a*sin(gamma
</table>
n*pi/2
Which leads us to the eigenvalues n =
cos(n*pi*x/2)
and eigenfunctions Xn (x) = c*e[-2*(n*pi/2)2*t]
e[-2*(n*pi/2)2*t]*cos(n*pi*x/2)
(Notation: Eigenfunctions should not include any cos(n*pi*x/2)
constants a or b.) 4*cos(n*pi*x/2)
-8/(n*pi)*sin(n*pi/2)
Solve for T (t).
2
Plug the eigenvalues n = 2n from Case 3 into the differential
equation for T (t) and solve: 7. (0 points) In your answers below, for the variable type
the word lambda, for type the word gamma; otherwise treat
Tn (t) = these as you would any other variable.

(Notation: use c for the unknown constant.) We will solve the heat equation
Combining all of the Xn and Tn we get that
ut = 4 uxx , 0 < x < 2, t 0

u(x,t) = An with boundary/initial conditions:
n=0 (
where An are unknown constants. ux (0,t) = 0, 2, 0 < x 1
and u(x, 0) =
Fourier Coefficients. u(2,t) = 0, 0, 1 < x < 2
We compute An by plugging t = 0 into the formula for u(x,t) This models temperature in a thin rod of length L = 2 with
and setting equal to the initial heat distribution given in the thermal diffusivity = 4 where one end is insulated and the
problem. other end has fixed temperature and the initial temperature dis-

( tribution is u(x, 0).
0, 0 < x < 1
u(x, 0) = An =
n=0 4, 1 x < 2 For extra practice we will solve this problem from scratch.
So the An are Fourier coefficients. Separate Variables.
Z 2 Separate variables.
2 Assume u(x,t) = X(x) T (t) and split the PDE into two differen-
An = dx (for n 6= 0)
2 1 tial equations, one with X and one with T .
= = =
A0 = (Notation: Write X and T for derivatives. Place all constants
Remember that sin(n) = 0. in the differential equation with T).
Correct Answers: DE for X(x): =0
X/X = T/(2*T) = \(-\lambda\) Boundary conditions for X(x):
X+lambda*X
X(0)=0
X(2)=0
T+2*lambda*T
4
Solve for T (t).
(Enter boundary equations: e.g. X 0 (0) = 10) Plug the eigenvalues n = 2n from Case 3 into the differential
equation for T (t) and solve:
DE for T (t): =0
Tn (t) =
Find Eigenfunctions for X(x).
The problem splits into cases based on the sign of . (Notation: use c for the unknown constant.)
(Notation: For the cases below, use constants a and b) Combining all of the Xn and Tn we get that
Case 1: = 0
u(x,t) = An
X(x) = n=0
where An are unknown constants.
Plugging the boundary values into this formula gives Fourier Coefficients.
We compute An by plugging t = 0 into the formula for u(x,t)
0 = X 0 (0) = and setting equal to the initial heat distribution given in the
problem.
0 = X(2) = (

2, 0 < x 1
u(x, 0) = An =
So X(x) = n=0 0, 1 < x < 2
So the An are Fourier coefficients.
We can ignore this case.
Z 1
2
An = dx (for n 6= 0)
Case 2: = 2 (In your answers below use gamma 2 0
instead of lambda)
=
A0 =
X(x) =
Correct Answers:
X/X = T/(4*T) = \(-\lambda\)
Plugging the boundary values into this formula gives
X+lambda*X
X(0)=0
0 = X 0 (0) = X(2)=0
T+4*lambda*T
0 = X(2) = <table border=0 cellspacing=0 style=margin-left:auto;m
<tr><td>\(X(x) = \) a+b*x</td></tr>
So X(x) = <tr><td>\(X(0) = \) b</td></tr>
<tr><td>\(X(2) = \) a+b*2</td></tr>
We can ingore this case. </table>
a
<table border=0 cellspacing=0 style=margin-left:auto;m
Case 3: = 2 (In your answers below use gamma
<tr><td>\(X(x) = \) a*e(gamma*x)+b*e(-gamma*x)</td><
instead of lambda)
<tr><td>\(X(0) = \) gamma*(a-b)</td></tr>
<tr><td>\(X(2) = \) a*e(gamma*2)+b*e(-gamma*2)</td><
X(x) = </table>
0
Plugging the boundary values into this formula gives <table border=0 cellspacing=0 style=margin-left:auto;m
<tr><td>\(X(x) = \) a*cos(gamma*x)+b*sin(gamma*x)</td>
0 = X 0 (0) = <tr><td>\(X(0) = \) gamma*b</td></tr>
<tr><td>\(X(2) = \) a*cos(gamma*2)+b*sin(gamma*2)</td>
0 = X(2) = </table>
(2*n+1)*pi/4
cos((2*n+1)*pi*x/4)
c*e(-4*[(2*n+1)*pi/4]2*t)
Which leads us to the eigenvalues n =
e(-4*[(2*n+1)*pi/4]2*t)*cos((2*n+1)*pi*x/4)
and eigenfunctions Xn (x) = cos((2*n+1)*pi*x/4)
2*cos((2*n+1)*pi*x/4)
(Notation: Eigenfunctions should not include any 8/[(2*n+1)*pi]*sin((2*n+1)*pi/4)
constants a or b.) 1
5
Generated by WeBWorK,
c http://webwork.maa.org, Mathematical Association of America

Potrebbero piacerti anche