HW9 Solution
description
Transcript of HW9 Solution
![Page 1: HW9 Solution](https://reader035.fdocuments.in/reader035/viewer/2022081908/553f5f234a7959e10e8b4734/html5/thumbnails/1.jpg)
HW 9 Solution
Due: November 29 (Thursday), 2012
1. Find the type of PDE, transform to normal form, and solve. Please show your work in detail.
uxx-16uyy=0.
Solution: Hyperbolic, wave equation. Characteristic equation
π¦β²2 β 16 = (π¦β² + 4)(π¦β² β 4) = 0
New variables are
v = Ο = y + 4x, w = Ο = y β 4x
By the chain rule,
π’π₯ = 4π’π£ β 4π’π€
π’π₯π₯ = 16π’π£π£ β 16π’π£π€ β 16π’π€π£ + 16π’π€π€
And
β16π’π¦π¦ = β16π’π£π£ β 16π’π£π€ β 16π’π€π£ β 16π’π€π€
Assuming π’π£π€ = π’π€π£, as usual , we have
π’π£π€ = 0
Solvable by two integrations,
u = π1(π¦ + 4π₯) + π2(π¦ β 4π₯)
2. Find the type of PDE, transform to normal form, and solve. Please show your work in detail.
uxx-6uxy+9uyy=0.
Solution: Parabolic. Characteristic equation
π¦β²2 + 6π¦β² + 9 = (π¦β² + 3)2
New variables v = Ο = x, w = Ο = y + 3x. By the chain rule,
π’π₯ = π’π£ + 3π’π€
π’π₯π₯ = π’π£π£ + 6π’π£π€ + 9π’π€π€
π’π₯π¦ = π’π£π€ + 3π’π€π€
π’π¦π¦ = π’π€π€
Substitution into the PDE gives the expected normal form
π’π£π£ + 6π’π£π€ + 9π’π€π€ β 6π’π£π€ β 18π’π€π€ + 9π’π€π€ = π’π£π£ = 0
Solution
u = π1(π₯) + π2(π¦ + 3π₯)
where f1 and f2 are any twice differentiable functions of the respective variables.
3. Solve the heat transfer equation
βu
βt= π2
β2u
βx2
0β€xβ€L, x>0; L=1
Subject to the boundary conditions: u(0, t)=0, and βu
βt|π₯=πΏ = βππ’(1, π‘), k>0 and the
initial condition u(x, 0)=1, 0<x<1.
![Page 2: HW9 Solution](https://reader035.fdocuments.in/reader035/viewer/2022081908/553f5f234a7959e10e8b4734/html5/thumbnails/2.jpg)
Solution:
Assuming u(x, t)=X(x) T(t) and βΞ» as the separation constant, we find the separated
ODEs and boundary conditions to be, respectively,
Xβ²β² + Ξ»X = 0 (1)
Tβ² + π2ππ = 0 (2)
X(0) = 0 and Xβ²(1) = βkX(1) (3)
Equation (1) along with the homogeneous boundary conditions (3) comprise the regular
Sturm-Liouville problem:
Xβ²β² + Ξ»X = 0, X(0) = 0, Xβ²(1) + kX(1) = 0 (4)
Except for the presence of the symbol k, the BVP in (4) is essentially the problem solved
in Example 2 of Section 12. 5. As in that example, (4) possesses nontrivial solutions only
in the case Ξ»=Ξ±2>0, Ξ±>0. The general solution of the DE in (4) is X(x)=c1 cos Ξ±x+c2 sin Ξ±x.
The first boundary condition in (4) immediately gives c1=0. Applying the second
boundary condition in (4) to X(x)=c2 sinΞ±x yields
Ξ± cos πΌ + ππ πππΌ = 0 ππ tan πΌ = βπΌ
π (5)
Because the graphs of y=tan x and y=-x/k, k>0, have an infinite number of points of
intersection for x>0, the last equation in (5) has an infinite number of roots. Of course,
these roots depend on the value of k. If the consecutive positive roots are denoted Ξ±n,
n=1, 2, 3, β¦ , then the eigenvalues of the problem are Ξ»n=Ξ±n2, and the corresponding
eigenfunctions are X(x)=c2 sin Ξ±nx, n=1, 2, 3, β¦ , The solution of the first-order DE (2) is
T(t) = π3πβπ2πΌπ2 π‘ and so
π’π = ππ = π΄ππβπ2πΌπ2 π‘ sin πΌππ₯ πππ π’(π₯, π‘) = β π΄ππβπ2πΌπ
2 π‘ sin πΌππ₯
β
π=1
Now at t=0, u(x, 0)=1, 0<x<1, so that,
1 = β π΄π sin πΌππ₯βπ=1 (6)
The series in (6) is not a Fourier sine series; rather, it is an expansion of u(x, 0)=1 in terms
of the orthogonal functions arising from the Sturm-Liouville problem (4). It follows that
the set of eigenfunctions {sin Ξ±nx}, n=1, 2, 3, β¦ , where the Ξ±βs are defined by tan Ξ±=-Ξ±/k
is orthogonal with respect to the weight function p(x)=1 on the interval [0, 1]. With
f(x)=1 and Οn(x)=sin Ξ±nx, it follows from (8) of Section 12.1, that the coefficients An are (6)
are
π΄π =β« sin πΌππ₯
1
0ππ₯
β« sin2 πΌππ₯1
0ππ₯
(7)
Evaluate the square norm of each of the eigenfunctions we use a trigonometric identity:
β« sin2 πΌππ₯1
0ππ₯ =
1
2β« (1 β cos 2πΌππ₯
1
0)ππ₯ =
1
2(1 β
1
2πΌπsin 2πΌπ) (8)
With the aid of the double angle formula sin 2Ξ±n =2sin in the form Ξ±n cos Ξ±n and the first
equation in (5) in the form Ξ±ncosΞ±n=-k sinΞ±n, we can simplify (8) to
β« sin2 πΌππ₯1
0
ππ₯ =1
2π(π + cos2 πΌπ)
Also
![Page 3: HW9 Solution](https://reader035.fdocuments.in/reader035/viewer/2022081908/553f5f234a7959e10e8b4734/html5/thumbnails/3.jpg)
β« sin πΌππ₯1
0
ππ₯ = β1
πΌπcos πΌππ₯ |0
1 =1
πΌπ(1 β cos πΌπ)
Consequently (7) becomes
π΄π =2π(1 β cos πΌπ)
πΌπ(π + cos2 πΌπ)
Finally , a solution of the boundary-value problem is
π’(π₯, π‘ ) = 2π β1 β cos πΌπ
πΌπ(π + cos2 πΌπ)πβπ2πΌπ
2 π‘ sin πΌππ₯
β
π=1
4. For a steady-state heat flow problem in the rectangular sheet of size L*M to which Laplaceβs
equation applies. Suppose that the upper horizontal edge is kept at 100oC, while the other
three edge are kept at 0 oC. Please solve Laplaceβs equation to find the temperature
distribution function.
Solution: We may write these boundary conditions for T(x, y) as follows:
T(0, y) = T(L, y) = 0, 0 < y < M
T(x, 0) = 0, T(x, M) = 100, 0 < x < L
Since we are seeking the steady-state solution, time is not a factor in this problem, and
therefore no initial condition makes sense.
Setting T(x, y)= X(x)Y(y), we find that (6) becomes
Xβ²β²Y + XYβ²β² = 0
Which may be written in separated form as
Xβ²β²
X= β
Yβ²β²
Y
The left-hand side is constant for fixed x and arbitrary y. Hence Yββ/Y must be constant.
Letting k be this constant, we obtain the pair of ordinary differential equations
Xβ²β² β ππ = 0
Yβ²β² + ππ = 0
From the first boundary condition we have
X(0)Y(y) = X(L)Y(y) = 0
And Y(y)β 0 (since otherwise T(x, y)β‘0, contradicting the second boundary condition), so that
X(0)=X(L)=0. Thus again we have the situation encountered in Section 12. 7, and the
eigenvalues are all negative. Setting k=-r2, we find that rL must be a multiple of Ο in order
that we have a nonzero solution of the boundary value problem for X. Hence r=nΟ/L, and the
functions
ππ(π₯) = π΄ sinπππ₯
πΏ, π = 1, 2, 3, β¦,
are the eigenfunctions of the problem. Setting k=-(nΟ/L)2 in Yββ+kY=0 yields the general
solution
ππ(π¦) = π΅1ππππ¦/πΏ + π΅2πβπππ¦/πΏ
The second boundary condition implies that Y(0)=0, thus B1=-B2, and we can rewrite Yn as
ππ(π¦) = π΅ sinhπππ¦
πΏ
![Page 4: HW9 Solution](https://reader035.fdocuments.in/reader035/viewer/2022081908/553f5f234a7959e10e8b4734/html5/thumbnails/4.jpg)
As in Section 12.7, to enlarge the class of possible solutions, we consider and infinite sum of
products of Xn(x) and Yn(y):
T(π₯, π¦) = β ππ sinπππ₯
πΏsinh
πππ¦
πΏ
β
π=1
Evaluating at any point (x, M), we obtain by the second boundary condition,
100 = β(ππ sinhπππ
πΏ) sin
πππ₯
πΏ
β
π=1
, 0 < x < L
Which again is a Fourier sine series, requiring that we extend the boundary condition T(x, M)
to the interval -Lβ€xβ€L as an odd pricewise continuous periodic function of period 2L with a
piecewise continuous derivative. The function
f(x) = {100, 0 < π₯ < πΏ,
β100, βπΏ < π₯ < 0
with f(x+2L+=f(x) satisfies these conditions. Hence cn sinh(nΟM/L) must equal the nth Fourier
(sine) coefficient of f(x):
ππ sinhπππ
πΏ=
1
πΏβ« π(π₯) sin
πππ₯
πΏ
πΏ
βπΏ
ππ₯ =200
πΏβ« sin
πππ₯
πΏ
πΏ
0
ππ₯ =200
ππ(1 β cos ππ)
=200
ππ[1 β (β1)π]
Hence c2k=0, while
π2π+1 =400
π(2π + 1) sinh[(2π + 1)ππ/πΏ]
And the solution is given by
T(π₯, π¦) =400
πβ
sin(2π β 1)ππ₯
πΏ sinh(2π β 1)ππ¦
πΏ(2π β 1) sinh[(2π β 1)ππ/πΏ]
β
π=1
5. A rectangular metal plate 0β€xβ€Ο, 0β€yβ€2, in which the steady state temperature
distribution u(x, y) is required subject to the temperature on the side y=0 being
u(x, 0)=10, and the temperature on the other three sides being maintained at u=0. Find
the temperature distribution across the metal plate.
π’π‘ = π2(π’π₯π₯ + π’π¦π¦) = 0
Solution:
Let u(x, y)=X(y)Y(y)
So
Xβ²β²
X= β
Yβ²β²
Y= π = βπ2
πβ²β² + π2π = 0, Yβ²β² β π2π = 0
General solution of X(x):
X(x) = π΄ cos ππ₯ + π΅ sin ππ₯
B. C.: X(0)=0, so A=0, X(x)=B sin Ξ»x
![Page 5: HW9 Solution](https://reader035.fdocuments.in/reader035/viewer/2022081908/553f5f234a7959e10e8b4734/html5/thumbnails/5.jpg)
X(Ο)=0, since Bβ 0, Ξ»n=n, for n=1, 2, β¦,
So Xπ(x) = π΅ sin ππ₯ for n=1, 2, β¦,
ππ(π¦) = πΆ cosh ππ¦ + π· sinh ππ¦
B. C.: u(x, 2)=0, y(2)=0
So ππ(π¦) = πΆ cosh 2π + π· sinh 2π = 0
Setting D=1, and C=β sinh(2π)/ cosh(2π)
ππ(π¦) =πΆ
cosh 2π(sinh ππ¦ cosh 2π β cosh ππ¦ sinh 2π¦) =
πΆ
cosh 2πsinh π(π¦ β 2)
π’π(π₯, π¦) = ππ(π₯)ππ(π¦) = πΆπ sin ππ₯ sinh π(π¦ β 2)
u(π₯, π¦) = β πΆπ sin ππ₯ sinh π(π¦ β 2)
β
π=1
B. C.: u(x, 0)=10, so y(0)=10
10 = β πΆπ sin ππ₯ sinh(β2π)
β
π=1
πΆπ = β2
π sinh 2πβ« 10 sin ππ₯
π
0
ππ₯ =20
ππ sinh 2π(1 β cos ππ)