AM364

Week 5
Chapter 12
Superposition of Boundary Value Problems
The WAVE equation revisited

2
 Simple Example: The 1D WAVE equation
Seek solutions of the form of a product of a function of 𝑥 and a function of 𝑡.

𝑢𝑡𝑡 = 𝑎2 𝑢𝑥𝑥 ,
The FINAL SOLUTION:
∞
𝑘𝜋𝑥 𝑎𝑘𝜋 𝑏ത𝑘 𝐿 𝑎𝑘𝜋
𝑢 𝑥, 𝑡 = ෍ sin 𝑏𝑘 cos 𝑡 + sin 𝑡
𝐿 𝐿 𝑎𝑘𝜋 𝐿
𝑘=1

2 𝐿 𝑘𝜋𝑥
𝑏𝑘 = න 𝑓(𝑥) sin 𝑑𝑥
𝐿 0 𝐿
2 𝐿 𝑘𝜋𝑥
ത
𝑏𝑘 = න 𝑔(𝑥) sin 𝑑𝑥
𝐿 0 𝐿
 Simpler Example 1: The 1D WAVE equation
Seek solutions of the form of a product of a function of 𝑥 and a function of 𝑡.

𝑢𝑡𝑡 = 𝑎2 𝑢𝑥𝑥 ,

0,
The FINAL SOLUTION:
∞
𝑘𝜋𝑥 𝑎𝑘𝜋 𝑏ത𝑘 𝐿 𝑎𝑘𝜋
𝑢1 𝑥, 𝑡 = ෍ sin 𝑏𝑘 cos 𝑡 + sin 𝑡
𝐿 𝐿 𝑎𝑘𝜋 𝐿
𝑘=1

2 𝐿 𝑘𝜋𝑥
𝑏𝑘 = න 𝑓(𝑥) sin 𝑑𝑥
𝐿 0 𝐿
2 𝐿 𝑘𝜋𝑥
ത
𝑏𝑘 = න 𝑔(𝑥) sin 𝑑𝑥
𝐿 0 𝐿
 Simpler Example 2: The 1D WAVE equation
Seek solutions of the form of a product of a function of 𝑥 and a function of 𝑡.

𝑢𝑡𝑡 = 𝑎2 𝑢𝑥𝑥 ,

0,
The FINAL SOLUTION:
∞
𝑘𝜋𝑥 𝑎𝑘𝜋 𝑏ത𝑘 𝐿 𝑎𝑘𝜋
𝑢2 𝑥, 𝑡 = ෍ sin 𝑏𝑘 cos 𝑡 + sin 𝑡
𝐿 𝐿 𝑎𝑘𝜋 𝐿
𝑘=1

2 𝐿 𝑘𝜋𝑥
𝑏𝑘 = න 𝑓(𝑥) sin 𝑑𝑥
𝐿 0 𝐿
2 𝐿 𝑘𝜋𝑥
ത
𝑏𝑘 = න 𝑔(𝑥) sin 𝑑𝑥
𝐿 0 𝐿
 Simple Example: The 1D WAVE equation
Seek solutions of the form of a product of a function of 𝑥 and a function of 𝑡.
𝑢(𝑥, 𝑡)

𝑢𝑡𝑡 = 𝑎2 𝑢𝑥𝑥 ,
 Sum of the two Simpler Examples
Seek solutions of the form of a product of a function of 𝑥 and a function of 𝑡.
𝑢 𝑥, 𝑡 = 𝑢1 𝑥, 𝑡 + 𝑢2 𝑥, 𝑡

𝑢𝑡𝑡 = 𝑎2 𝑢𝑥𝑥 ,

+
Fully developed, time independent velocity profile: flow through square duct
orientated along the 𝑥-axis
Laplace
Background

12
Solve Laplace’s equation on a rectangular domain
Solve 𝑢𝑥𝑥 + 𝑢𝑦𝑦 = 0 where
2𝑥 𝑥 ∈ 0,0.5
• 𝑢 𝑥, 0 = ቊ
2(1 − 𝑥) 𝑥 ∈ (0.5, 1ሿ
• 𝑢 𝑥, 1 = sin(𝜋𝑥)
• 𝑢 0, 𝑦 = 0
• 𝑢 1, 𝑦 = 0
𝑦 on the square domain numerically.

1 1
0 1 0
2 2 Analytical solution:
7 8 9 sinh 𝜋𝑦
∞
𝑒 𝑘𝜋𝑦 − 𝑒 2𝑘𝜋 𝑒 −𝑘𝜋𝑦
0 0 𝑢= sin 𝜋𝑥 + ෍ 𝑏𝑘 sin(𝑘𝜋𝑥)
sinh 𝜋 1 − 𝑒 2𝑘𝜋
𝑘=1
4 5 6 𝑘𝜋
0 0 8 sin
2
𝑏𝑘 = 2
𝑘𝜋
1 2 3
0 0

0 0.5 1 0.5 0
𝑥
Solve Laplace’s equation on a rectangular domain
𝑢𝑥𝑥 + 𝑢𝑦𝑦 = 0 → In general, considering point (𝑥𝑗,𝑛 , 𝑦𝑗,𝑛 ),

𝑈𝑗+1,𝑛 − 2𝑈𝑗,𝑛 + 𝑈𝑗−1,𝑛 𝑈𝑗,𝑛+1 − 2𝑈𝑗,𝑛 + 𝑈𝑗,𝑛−1

0= 2
+
ℎ ℎ2
0 = 𝑈𝑗+1,𝑛 + 𝑈𝑗−1,𝑛 − 4𝑈𝑗,𝑛 + 𝑈𝑗,𝑛+1 + 𝑈𝑗,𝑛−1
𝑦

𝑈𝑗+1,𝑛 + 𝑈𝑗−1,𝑛 + 𝑈𝑗,𝑛+1 + 𝑈𝑗,𝑛−1

𝑈𝑗,𝑛 =
4

Stencil for solving Laplace

ℎ

𝑥
Solve Laplace’s equation on a rectangular domain
𝑦

1 1
0 1 0
2 2

7 8 9
0 0

4 5 6
0 0

1 2 3
0 0

0 0.5 1 0.5 0
𝑥
[XN,YN]=meshgrid(0:0.25:1);
UM=0*XN;
UM(1,2:4)=[0.5 1 0.5];
UM(2,2:4)=UVector(1:3);
UM(3,2:4)=UVector(4:6);
UM(4,2:4)=UVector(7:9);
UM(5,2:4)=[1/sqrt(2) 1 1/sqrt(2)];
Analytical solution Numerical solution
Solve Laplace’s equation on a rectangular domain
Iterative formula will converge 0 = 𝑢𝑥𝑥 + 𝑢𝑦𝑦

𝑈𝑗+1,𝑛 − 2𝑈𝑗,𝑛 + 𝑈𝑗−1,𝑛 𝑈𝑗,𝑛+1 − 2𝑈𝑗,𝑛 + 𝑈𝑗,𝑛−1

0= 2
+
ℎ 𝑘2
𝑘 2 𝑈𝑗+1,𝑛 − 2𝑈𝑗,𝑛 + 𝑈𝑗−1,𝑛 + ℎ2 (𝑈𝑗,𝑛+1 − 2𝑈𝑗,𝑛 + 𝑈𝑗,𝑛−1 )
0=
ℎ2 𝑘 2
𝑘 2 (𝑈𝑗+1,𝑛 + 𝑈𝑗−1,𝑛 ) + ℎ2 (𝑈𝑗,𝑛+1 + 𝑈𝑗,𝑛−1 )
𝑈𝑗,𝑛 =
2(𝑘 2 + ℎ2 )

for n=2:1:N-1
for j=2:1:J-1
UUM(n,j)=(k^2*(UUM(n,j+1)+UUM(n,j-1))+
h^2*(UUM(n+1,j)+UUM(n-1,j)))/
2*(k^2+h^2));
end
end
Laplace
The 2D Heat equation
𝑢𝑡 = 𝜎(𝑢𝑥𝑥 + 𝑢𝑦𝑦 )

• Steady state: 𝑢𝑥𝑥 + 𝑢𝑦𝑦 = 0 → Laplace’s equation

The 2D Wave equation

𝑢𝑡𝑡 = 𝑎2 𝑢𝑥𝑥 + 𝑢𝑦𝑦 − 𝑔

• Steady state and no source term: 𝑢𝑥𝑥 + 𝑢𝑦𝑦 = 0 → Laplace’s equation

Quote from the textbook:

There is a maximum principle that states a solution 𝑢 of Laplace’s
equation within a bounded region 𝑅 with boundary 𝐵 (such as a rectangle,
circle, sphere, and so on) takes on its maximum and minimum values on 𝐵.
Laplace
Example 1: From textbook

19
 𝑦
𝑢𝑥𝑥 + 𝑢𝑦𝑦 = 0
𝐷
on
𝑥 ∈ 0, 𝑊
𝑦 ∈ 0, 𝐷
𝑥
𝑊
 𝑦
𝑢𝑥𝑥 + 𝑢𝑦𝑦 = 0 𝑢(𝑥, 𝐷) = 𝑓(𝑥)
𝐷
on

insulated
insulated
𝑥 ∈ 0, 𝑊
𝑦 ∈ 0, 𝐷
subject to 𝑥
𝑢(𝑥, 0) = 0
𝑊
1 𝑢𝑥 (0, 𝑦) = 0
2 𝑢𝑥 (𝑊, 𝑦) = 0
3 𝑢 𝑥, 0 = 0
4 𝑢 𝑥, 𝐷 = 𝑓(𝑥)
 Separation of variables:
Seek solutions of the form of a product of a function of 𝑥 and a function of 𝑦.

𝑢𝑥𝑥 + 𝑢𝑦𝑦 = 0 …(1)

Let 𝑢 𝑥, 𝑦 = 𝑋 𝑥 𝑌(𝑦) and substitute into (1):

𝑋 ′′ (𝑥)𝑌 𝑦 + 𝑋 𝑥 𝑌′′(𝑦) = 0 …(2)

Now we separate the terms containing 𝑥’s and 𝑦 ′ 𝑠 in (2):

𝑋 ′′ 𝑥 𝑌 ′′ (𝑦)
=− …(3)
𝑋 𝑥 𝑌(𝑦)

Sturm-Liouville?
 𝑦
𝑢𝑥𝑥 + 𝑢𝑦𝑦 = 0 𝑢(𝑥, 𝐷) = 𝑓(𝑥)
𝐷
on

insulated
insulated
𝑥 ∈ 0, 𝑊
𝑦 ∈ 0, 𝐷
subject to 𝑥
𝑢(𝑥, 0) = 0
𝑊
1 𝑢𝑥 (0, 𝑦) = 0
2 𝑢𝑥 (𝑊, 𝑦) = 0
3 𝑢 𝑥, 0 = 0
4 𝑢 𝑥, 𝐷 = 𝑓(𝑥)

1 𝑋′(0) = 0
2 𝑋′(𝑊) = 0
3 𝑌 0 =0
 𝑋 ′′ (𝑥)
= −𝜆 …(4)
𝑋 𝑥

Sturm-Liouville problem
𝑋 ′′ (𝑥) + 𝜆𝑋 𝑥 = 0
1 𝑋′ 0 = 0
2 𝑋′ 𝑊 = 0
 𝑋 ′′ (𝑥)
= −𝜆 …(4)
𝑋 𝑥

Sturm-Liouville problem
Eigenvalue: 𝜆 = 0
𝑋 ′′ 𝑥 =0
1 𝑋′ 0 = 0 Eigenfunction: 𝑋0 (𝑥) = 𝐵
2 𝑋′ 𝑊 = 0

Setting 𝜆 = 0: 𝑋 𝑥 = 𝐴𝑥 + 𝐵
𝑋′ 𝑥 = 𝐴
Imposing either [1] or [2]: 𝐴 = 0.
𝑋0 𝑥 = 𝐵
 𝑋 ′′ (𝑥)
= −𝜆 …(4)
𝑋 𝑥

Sturm-Liouville problem
𝑋 ′′ 𝑥 − 𝛼 2 𝑋 𝑥 = 0
1 𝑋′ 0 = 0
2 𝑋′ 𝑊 = 0

Setting 𝜆 = −𝛼 2 : 𝑋 𝑥 = 𝐴 sinh(𝛼𝑥) + 𝐵 cosh(𝛼𝑥)

𝑋′ 𝑥 = 𝐴𝛼 cosh(𝛼𝑥) + 𝐵𝛼 sinh(𝛼𝑥)
Imposing [1]: 𝑋 ′ 0 = 0 = 𝐴𝛼, therefore 𝐴 = 0

𝑋′ 𝑥 = 𝐵𝛼 sinh(𝛼𝑥)
Imposing [2]: 𝑋 ′ 𝑊 = 0 = 𝐵𝛼 sinh(𝛼𝑊), therefore 𝐵 = 0, since 𝛼 ≠ 0
𝑋 𝑥 =0 Trivial solution
 𝑋 ′′ (𝑥)
= −𝜆 …(4)
𝑋 𝑥

Sturm-Liouville problem 𝑘𝜋
2
2
Eigenvalues: 𝜆𝑘 = 𝛼𝑘 =
𝑋 ′′ 𝑥 + 𝛼 2 𝑋 𝑥 = 0 𝐿

1 𝑋′ 0 = 0 𝑘𝜋
Eigenfunctions: 𝑋𝑘 (𝑥) = 𝐵𝑘 cos 𝑥
2 𝑋′ 𝑊 = 0 𝑊

Setting 𝜆 = 𝛼 2 : 𝑋 𝑥 = 𝐴 sin(𝛼𝑥) + 𝐵 cos(𝛼𝑥)

𝑋 ′ (𝑥) = 𝐴𝛼 cos 𝛼𝑥 − 𝐵𝛼 sin(𝛼𝑥)
Imposing [1]: 𝑋 ′ 0 = 0 = 𝐴𝛼, therefore 𝐴 = 0

𝑋 ′ (𝑥) = −𝐵𝛼 sin(𝛼𝑥)

Imposing [2]: 𝑋 ′ 𝑊 = 0 = −𝐵𝛼 sin(𝛼𝑊), therefore sin(𝛼𝑊) = 0,
yielding 𝛼 = 𝑘𝜋/𝑊 where 𝑘 = 1,2,3, …
𝑘𝜋𝑥
𝑋𝑘 𝑥 = 𝐵𝑘 cos , 𝑘 = 1,2,3, …
𝑊
 Separation of variables:
Seek solutions of the form of a product of a function of 𝑥 and a function of 𝑡.

𝑢𝑥𝑥 + 𝑢𝑦𝑦 = 0 …(1)

Let 𝑢 𝑥, 𝑦 = 𝑋 𝑥 𝑌(𝑦) and substitute into (1):

𝑋 ′′ (𝑥)𝑌 𝑦 + 𝑋 𝑥 𝑌′′(𝑦) = 0 …(2)

Now we separate the terms containing 𝑥’s and 𝑦 ′ 𝑠 in (2):

𝑋 ′′ 𝑥 𝑌 ′′ (𝑦)
=− …(3)
𝑋 𝑥 𝑌(𝑦)
𝑋 ′′ 𝑥 𝑌 ′′ 𝑦
=− = −𝜆
𝑋 𝑥 𝑌 𝑦
 𝑌 ′′ (𝑦)
=𝜆 …(4)
𝑌(𝑦)

𝑌 ′′ 𝑦 − 𝜆𝑌 𝑦 = 0 Eigenvalue: 𝜆 = 0
3 𝑌 0 =0 𝑘𝜋
2
2
Eigenvalues: 𝜆𝑘 = 𝛼𝑘 =
𝑊
 𝑌 ′′ (𝑦)
=𝜆 …(4)
𝑌(𝑦)

𝑌 ′′ 𝑦 = 0 Eigenvalue: 𝜆 = 0
3 𝑌 0 =0 𝑘𝜋
2
2
Eigenvalues: 𝜆𝑘 = 𝛼𝑘 =
𝑊

Setting 𝜆 = 0: 𝑌 𝑦 = 𝐴𝑦 + 𝐵
Imposing [3]: 𝐵 = 0. Eigenfunction: 𝑋0 (𝑥) = 𝐵
𝑌0 𝑦 = 𝐴𝑦

𝑢1 𝑥, 𝑦 = 𝐴𝐵𝑦 = 𝐴𝑦 (𝐵 redundant, set it equal to 1)

 𝑌 ′′ (𝑦)
=𝜆 …(4)
𝑌(𝑦)

𝑌 ′′ 𝑦 − 𝛼 2 𝑌 𝑦 = 0 Eigenvalue: 𝜆 = 0
3 𝑌 0 =0 𝑘𝜋
2
2
Eigenvalues: 𝜆𝑘 = 𝛼𝑘 =
𝑊

Setting 𝜆 = 𝛼 2 : 𝑌 𝑦 = 𝐴 sinh(𝛼𝑦) + 𝐵 cosh(𝛼𝑦)

Imposing [3]: 𝑌 0 = 0 = 𝐵, therefore 𝐵 = 0
𝑘𝜋𝑦
𝑌 𝑦 = 𝐴 sinh
𝑊 𝑘𝜋
Eigenfunctions: 𝑋𝑘 (𝑥) = 𝐵𝑘 cos 𝑥
𝑊

𝑘𝜋𝑦 𝑘𝜋𝑥 𝑘𝜋𝑦 𝑘𝜋𝑥

𝑢2 𝑥, 𝑦 = 𝐴 sinh 𝐵𝑘 cos = 𝐵𝑘 sinh cos
𝑊 𝑊 𝑊 𝑊
(𝐴 redundant, set it equal to 1)
𝑢1 𝑥, 𝑦 = 𝐴𝑦
𝑘𝜋𝑦 𝑘𝜋𝑥
𝑢2 𝑥, 𝑦 = 𝐵𝑘 sinh cos
𝑊 𝑊

From the superposition principle, the more specific solution, after imposing
boundary conditions [1], [2] and [3] (i.e. the homogeneous (or, zero)
boundary conditions), is:
∞
𝑘𝜋𝑦 𝑘𝜋𝑥
𝑢 𝑥, 𝑦 = 𝐴𝑦 + ෍ 𝐵𝑘 sinh cos
𝑊 𝑊
𝑘=1

Impose boundary condition [4]: 𝑢 𝑥, 𝐷 = 𝑓(𝑥)

∞
𝑘𝜋𝐷 𝑘𝜋𝑥
𝑢 𝑥, 𝐷 = 𝑓(𝑥) = 𝐴𝐷 + ෍ 𝐵𝑘 sinh cos
𝑊 𝑊
𝑘=1
𝑢1 𝑥, 𝑦 = 𝐴𝑦
𝑘𝜋𝑦 𝑘𝜋𝑥
𝑢2 𝑥, 𝑦 = 𝐵𝑘 sinh cos
𝑊 𝑊

From the superposition principle, the more specific solution, after imposing
boundary conditions [1], [2] and [3] (i.e. the homogeneous (or, zero)
boundary conditions), is:
∞
𝑘𝜋𝑦 𝑘𝜋𝑥
𝑢 𝑥, 𝑦 = 𝐴𝑦 + ෍ 𝐵𝑘 sinh cos
𝑊 𝑊
𝑘=1

Impose boundary condition [4]: 𝑢 𝑥, 𝐷 = 𝑓(𝑥)

∞
𝑘𝜋𝐷 𝑘𝜋𝑥
𝑢 𝑥, 𝐷 = 𝑓(𝑥) = 𝐴𝐷(1) + ෍ 𝐵𝑘 sinh cos
𝑊 𝑊
𝑘=1
𝑎0
2 𝑎𝑘
𝑢1 𝑥, 𝑦 = 𝐴𝑦
𝑘𝜋𝑦 𝑘𝜋𝑥
𝑢2 𝑥, 𝑦 = 𝐵𝑘 sinh cos
𝑊 𝑊

Impose boundary condition [4]: 𝑢 𝑥, 𝐷 = 𝑓(𝑥)

∞
𝑘𝜋𝐷 𝑘𝜋𝑥
𝑢 𝑥, 𝐷 = 𝑓(𝑥) = 𝐴𝐷(1) + ෍ 𝐵𝑘 sinh cos
𝑊 𝑊
𝑘=1
𝑎0
2 𝑎𝑘

𝑎0 2 𝑊
𝐴= where 𝑎0 = න 𝑓(𝑥) 𝑑𝑥
2𝐷 𝑊 0

𝑎𝑘 2 𝑊 𝑘𝜋𝑥
𝐵𝑘 = where 𝑎𝑘 = න 𝑓 𝑥 cos 𝑑𝑥
𝑘𝜋𝐷 𝑊 0 𝑊
sinh
𝑊
Final solution: Laplace’s equation, Example 1
∞
𝑎0 𝑦 𝑎𝑘 𝑘𝜋𝑦 𝑘𝜋𝑥
𝑢 𝑥, 𝑦 = +෍ sinh cos
2𝐷 𝑘𝜋𝐷 𝑊 𝑊
𝑘=1 sinh
𝑊
with

2 𝑊
𝑎0 = න 𝑓(𝑥) 𝑑𝑥
𝑊 0

2 𝑊 𝑘𝜋𝑥
𝑎𝑘 = න 𝑓 𝑥 cos 𝑑𝑥
𝑊 0 𝑊
Laplace
Example 2: Dirichlet nonhomogeneous

36
 𝑦
𝑢𝑥𝑥 + 𝑢𝑦𝑦 = 0 𝑢(𝑥, 𝐷) = 𝑓(𝑥)
𝐷

𝑢(𝑊, 𝑦) = ℎ(𝑦)
𝑢(0, 𝑦) = ℓ(𝑦)
on
𝑥 ∈ 0, 𝑊
𝑦 ∈ 0, 𝐷
subject to 𝑥
𝑢(𝑥, 0) = 𝑔(𝑥) 𝑊
1 𝑢(0, 𝑦) = ℓ(𝑦)
2 𝑢(𝑊, 𝑦) = ℎ(𝑦)
3 𝑢 𝑥, 0 = 𝑔(𝑥)
4 𝑢 𝑥, 𝐷 = 𝑓(𝑥)
𝑢𝑥𝑥 + 𝑢𝑦𝑦 = 0
on
𝑥 ∈ 0, 𝑊
𝑦 ∈ 0, 𝐷
Problem A Problem B
subject to
1 𝑢(0, 𝑦) = 0 1 𝑢(0, 𝑦) = ℓ(𝑦)
2 𝑢(𝑊, 𝑦) = 0 2 𝑢(𝑊, 𝑦) = ℎ(𝑦)
3 𝑢 𝑥, 0 = 𝑔(𝑥) 3 𝑢 𝑥, 0 = 0
4 𝑢 𝑥, 𝐷 = 𝑓(𝑥) 4 𝑢 𝑥, 𝐷 = 0
𝑢𝑥𝑥 + 𝑢𝑦𝑦 = 0
on
𝑥 ∈ 0, 𝑊
𝑦 ∈ 0, 𝐷
subject to
Problem A + Problem B
1 𝑢(0, 𝑦) = 0 1 𝑢(0, 𝑦) = ℓ(𝑦)
2 𝑢(𝑊, 𝑦) = 0 2 𝑢(𝑊, 𝑦) = ℎ(𝑦)
3 𝑢 𝑥, 0 = 𝑔(𝑥) 3 𝑢 𝑥, 0 = 0
4 𝑢 𝑥, 𝐷 = 𝑓(𝑥) 4 𝑢 𝑥, 𝐷 = 0
 Separation of variables:
Seek solutions of the form of a product of a function of 𝑥 and a function of 𝑡.

𝑢𝑥𝑥 + 𝑢𝑦𝑦 = 0 …(1)

Let 𝑢 𝑥, 𝑦 = 𝑋 𝑥 𝑌(𝑦) and substitute into (1):

𝑋 ′′ (𝑥)𝑌 𝑦 + 𝑋 𝑥 𝑌′′(𝑦) = 0 …(2)

Now we separate the terms containing 𝑥’s and 𝑦 ′ 𝑠 in (2):

𝑋 ′′ 𝑥 𝑌 ′′ (𝑦)
=− …(3)
𝑋 𝑥 𝑌(𝑦)

Sturm-Liouville?
𝑢𝑥𝑥 + 𝑢𝑦𝑦 = 0
on
𝑥 ∈ 0, 𝑊
𝑦 ∈ 0, 𝐷
Problem A Problem B
subject to
1 𝑢(0, 𝑦) = 0 1 𝑢(0, 𝑦) = ℓ(𝑦)
2 𝑢(𝑊, 𝑦) = 0 2 𝑢(𝑊, 𝑦) = ℎ(𝑦)
3 𝑢 𝑥, 0 = 𝑔(𝑥) 3 𝑢 𝑥, 0 = 0
4 𝑢 𝑥, 𝐷 = 𝑓(𝑥) 4 𝑢 𝑥, 𝐷 = 0
 𝑋 ′′ (𝑥)
= −𝜆 …(4)
𝑋 𝑥

Sturm-Liouville problem
𝑋 ′′ (𝑥) + 𝜆𝑋 𝑥 = 0
1 𝑋 0 =0
2 𝑋 𝑊 =0

Setting 𝜆 = 0: 𝑋 𝑥 = 𝐴𝑥 + 𝐵 Trivial solution

Setting 𝜆 = −𝛼 2 : 𝑋 𝑥 = 𝐴 sinh(𝛼𝑥) + 𝐵 cosh(𝛼𝑥) Trivial solution

𝑘𝜋
Setting 𝜆 = 𝛼 2: 𝑋 𝑥 = 𝐴 sin(𝛼𝑥) + 𝐵 cos(𝛼𝑥) 𝐵 = 0; 𝛼 =
𝑊
, 𝑘 = 1,2,3

2
2 𝑘𝜋 𝑘𝜋
Eigenvalues: 𝜆𝑘 = 𝛼𝑘 = Eigenfunctions: 𝑋𝑘 (𝑥) = 𝐴𝑘 sin 𝑥
𝑊 𝑊
 𝑌 ′′ 𝑦
=𝜆 …(5)
𝑌(𝑦)
2
2 𝑘𝜋
Eigenvalues: 𝜆𝑘 = 𝛼𝑘 =
𝑊
2
𝑘𝜋
𝑌 ′′ 𝑦 − 𝑌(𝑦) = 0
𝑊

𝑘𝜋 𝑘𝜋
𝑌𝑘 𝑦 = 𝐶𝑘 cosh 𝑦 + 𝐷𝑘 sinh 𝑦
𝑊 𝑊

More specific solution after imposing boundary conditions [1] and [2]:
𝑘𝜋𝑥 𝑘𝜋 𝑘𝜋
𝑢 𝑥, 𝑦 = 𝑋𝑘 (𝑦)𝑌𝑘 𝑦 = 𝐴𝑘 sin 𝐶𝑘 cosh 𝑦 + 𝐷𝑘 sinh 𝑦
𝑊 𝑊 𝑊
 𝑌 ′′ 𝑦
=𝜆 …(5)
𝑌(𝑦)
2
2 𝑘𝜋
Eigenvalues: 𝜆𝑘 = 𝛼𝑘 =
𝑊
2
𝑘𝜋
𝑌 ′′ 𝑦 − 𝑌(𝑦) = 0
𝑊

𝑘𝜋 𝑘𝜋
𝑌𝑘 𝑦 = 𝐶𝑘 cosh 𝑦 + 𝐷𝑘 sinh 𝑦
𝑊 𝑊

From the superposition principle (and setting 𝐴𝑘 = 1 ∀ 𝑘):

∞
𝑘𝜋𝑥 𝑘𝜋 𝑘𝜋
𝑢 𝑥, 𝑦 = ෍ sin 𝐶𝑘 cosh 𝑦 + 𝐷𝑘 sinh 𝑦
𝑊 𝑊 𝑊
𝑘=1
From the superposition principle (and setting 𝐴𝑘 = 1 ∀ 𝑘):
∞
𝑘𝜋𝑥 𝑘𝜋 𝑘𝜋
𝑢 𝑥, 𝑦 = ෍ sin 𝐶𝑘 cosh 𝑦 + 𝐷𝑘 sinh 𝑦
𝑊 𝑊 𝑊
𝑘=1

Impose [3]: 𝑢 𝑥, 0 = 𝑔(𝑥)

∞
𝑘𝜋𝑥
𝑢 𝑥, 0 = 𝑔(𝑥) = ෍ 𝐶𝑘 sin
𝑊
𝑘=1

2 𝑊 𝑘𝜋𝑥
𝐶𝑘 = 𝑏𝑘 where 𝑏𝑘 = න 𝑔 𝑥 sin 𝑑𝑥
𝑊 0 𝑊

More specific solution:

∞
𝑘𝜋𝑥 𝑘𝜋 𝑘𝜋
𝑢 𝑥, 𝑦 = ෍ sin 𝑏𝑘 cosh 𝑦 + 𝐷𝑘 sinh 𝑦
𝑊 𝑊 𝑊
𝑘=1
More specific solution:
∞
𝑘𝜋𝑥 𝑘𝜋 𝑘𝜋
𝑢 𝑥, 𝑦 = ෍ sin 𝑏𝑘 cosh 𝑦 + 𝐷𝑘 sinh 𝑦
𝑊 𝑊 𝑊
𝑘=1

Impose [4]: 𝑢 𝑥, 𝐷 = 𝑓(𝑥)

∞
𝑘𝜋𝑥 𝑘𝜋 𝑘𝜋
𝑢 𝑥, 𝐷 = 𝑓(𝑥) = ෍ sin 𝑏𝑘 cosh 𝐷 + 𝐷𝑘 sinh 𝐷
𝑊 𝑊 𝑊
𝑘=1

𝑘𝜋 𝑘𝜋
𝑏𝑘 cosh 𝐷 + 𝐷𝑘 sinh 𝐷 = 𝑏ത𝑘
𝑊 𝑊
𝑏ത𝑘 𝑏𝑘
𝐷𝑘 = −
𝑘𝜋 𝑘𝜋
sinh 𝐷 tanh 𝐷
𝑊 𝑊
2 𝑊 𝑘𝜋𝑥
where ത
𝑏𝑘 = න 𝑓 𝑥 sin 𝑑𝑥
𝑊 0 𝑊
Final solution: Laplace’s equation, Example 2 Problem A

∞
𝑘𝜋𝑥 𝑘𝜋 𝑏ത𝑘 𝑏𝑘 𝑘𝜋
𝑢𝐴 𝑥, 𝑦 = ෍ sin 𝑏𝑘 cosh 𝑦 + − sinh 𝑦
𝑊 𝑊 𝑘𝜋 𝑘𝜋 𝑊
𝑘=1 sinh 𝐷 tanh 𝐷
𝑊 𝑊

with

2 𝑊 𝑘𝜋𝑥
𝑏𝑘 = න 𝑔 𝑥 sin 𝑑𝑥
𝑊 0 𝑊

2 𝑊 𝑘𝜋𝑥
ത
𝑏𝑘 = න 𝑓 𝑥 sin 𝑑𝑥
𝑊 0 𝑊
 Similarly, final solution: Laplace’s equation, Example 2 Problem B

∞
𝑘𝜋𝑦 𝑘𝜋 𝑏ത𝑘 𝑏𝑘 𝑘𝜋
𝑢𝐵 𝑥, 𝑦 = ෍ sin 𝑏𝑘 cosh 𝑥 + − sinh 𝑥
𝐷 𝐷 𝑘𝜋 𝑘𝜋 𝐷
𝑘=1 sinh 𝑊 tanh 𝑊
𝐷 𝐷

with

2 𝐷 𝑘𝜋𝑦
𝑏𝑘 = න ℓ 𝑦 sin 𝑑𝑦
𝐷 0 𝐷

2 𝐷 𝑘𝜋𝑦
ത
𝑏𝑘 = න ℎ 𝑦 sin 𝑑𝑦
𝐷 0 𝐷
 1 𝑢(0, 𝑦) = ℓ(𝑦)
2 𝑢(𝑊, 𝑦) = ℎ(𝑦)
Final solution: Laplace’s equation, Example 2
3 𝑢 𝑥, 0 = 𝑔(𝑥)
4 𝑢 𝑥, 𝐷 = 𝑓(𝑥)

𝑢 𝑥, 𝑦 = 𝑢𝐴 𝑥, 𝑦 + 𝑢𝐵 𝑥, 𝑦
∞ 𝑓 𝑥 𝑔 𝑥
𝑘𝜋𝑥 𝑔 𝑥 𝑘𝜋 𝑏𝑘 𝑏𝑘 𝑘𝜋
= ෍ sin 𝑏𝑘 cosh 𝑦 + − sinh 𝑦
𝑊 𝑊 𝑘𝜋 𝑘𝜋 𝑊
𝑘=1 sinh 𝐷 tanh 𝐷
𝑊 𝑊
∞ ℎ 𝑦 ℓ 𝑦
𝑘𝜋𝑦 ℓ 𝑦 𝑘𝜋 𝑏𝑘 𝑏𝑘 𝑘𝜋
+ ෍ sin 𝑏𝑘 cosh 𝑥 + − sinh 𝑥
𝐷 𝐷 𝑘𝜋 𝑘𝜋 𝐷
𝑘=1 sinh 𝐷 𝑊 tanh 𝐷 𝑊

with

𝑔 𝑥 2 𝑊 𝑘𝜋𝑥 ℓ 𝑦 2 𝐷 𝑘𝜋𝑦
𝑏𝑘 = න 𝑔 𝑥 sin 𝑑𝑥 𝑏𝑘 = න ℓ 𝑦 sin 𝑑𝑦
𝑊 0 𝑊 𝐷 0 𝐷

𝑓 𝑥 2 𝑊 𝑘𝜋𝑥 ℎ 𝑦 2 𝐷 𝑘𝜋𝑦
𝑏𝑘 = න 𝑓 𝑥 sin 𝑑𝑥 𝑏𝑘 = න ℎ 𝑦 sin 𝑑𝑦
𝑊 0 𝑊 𝐷 0 𝐷
Alternatively in exponential form…
 𝑌 ′′ 𝑦
=𝜆 …(5)
𝑌(𝑦)
2
2 𝑘𝜋
Eigenvalues: 𝜆𝑘 = 𝛼𝑘 =
𝑊
2
𝑘𝜋
𝑌 ′′ 𝑦 − 𝑌(𝑦) = 0
𝑊

𝑘𝜋 𝑘𝜋
𝑌𝑘 𝑦 = 𝐶𝑘 𝑒 𝑊 𝑦 + 𝐷𝑘 𝑒 −𝑊𝑦

More specific solution after imposing boundary conditions [1] and [2]:
𝑘𝜋𝑥 𝑘𝜋
𝑦 −
𝑘𝜋
𝑦
𝑢 𝑥, 𝑦 = 𝑋𝑘 (𝑦)𝑌𝑘 𝑦 = 𝐴𝑘 sin 𝐶𝑘 𝑒 𝑊 + 𝐷𝑘 𝑒 𝑊
𝑊
 𝑌 ′′ 𝑦
=𝜆 …(5)
𝑌(𝑦)
2
2 𝑘𝜋
Eigenvalues: 𝜆𝑘 = 𝛼𝑘 =
𝑊
2
𝑘𝜋
𝑌 ′′ 𝑦 − 𝑌(𝑦) = 0
𝑊

𝑘𝜋 𝑘𝜋
𝑌𝑘 𝑦 = 𝐶𝑘 𝑒 𝑊 𝑦 + 𝐷𝑘 𝑒 −𝑊𝑦

From the superposition principle (and setting 𝐴𝑘 = 1 ∀ 𝑘):

∞
𝑘𝜋𝑥 𝑘𝜋
𝑦
𝑘𝜋
− 𝑦
𝑢 𝑥, 𝑦 = ෍ sin 𝐶𝑘 𝑒 𝑊 + 𝐷𝑘 𝑒 𝑊
𝑊
𝑘=1
From the superposition principle (and setting 𝐴𝑘 = 1 ∀ 𝑘):
∞
𝑘𝜋𝑥 𝑘𝜋
𝑦
𝑘𝜋
− 𝑦
𝑢 𝑥, 𝑦 = ෍ sin 𝐶𝑘 𝑒 𝑊 + 𝐷𝑘 𝑒 𝑊
𝑊
𝑘=1

Impose [3]: 𝑢 𝑥, 0 = 𝑔(𝑥)

∞
𝑘𝜋𝑥
𝑢 𝑥, 0 = 𝑔(𝑥) = ෍ (𝐶𝑘 + 𝐷𝑘 ) sin
𝑊
𝑘=1

2 𝑊 𝑘𝜋𝑥
𝐶𝑘 + 𝐷𝑘 = 𝑏𝑘 where 𝑏𝑘 = න 𝑔 𝑥 sin 𝑑𝑥
𝑊 0 𝑊

More specific solution:

∞
𝑘𝜋𝑥 𝑘𝜋
𝑦 −
𝑘𝜋
𝑦
𝑢 𝑥, 𝑦 = ෍ sin 𝑏𝑘 − 𝐷𝑘 𝑒 𝑊 + 𝐷𝑘 𝑒 𝑊
𝑊
𝑘=1
More specific solution:
∞
𝑘𝜋𝑥 𝑘𝜋
𝑦 −
𝑘𝜋
𝑦
𝑢 𝑥, 𝑦 = ෍ sin 𝑏𝑘 − 𝐷𝑘 𝑒 𝑊 + 𝐷𝑘 𝑒 𝑊
𝑊
𝑘=1

Impose [4]: 𝑢 𝑥, 𝐷 = 𝑓(𝑥)

∞
𝑘𝜋𝑥 𝑘𝜋 𝑘𝜋
𝑢 𝑥, 𝐷 = 𝑓(𝑥) = ෍ sin 𝑏𝑘 − 𝐷𝑘 𝑒 𝑊 𝐷 + 𝐷𝑘 𝑒 −𝑊𝐷
𝑊
𝑘=1

𝑘𝜋 𝑘𝜋
𝐷 − 𝐷
𝑏𝑘 − 𝐷𝑘 𝑒𝑊 + 𝐷𝑘 𝑒 𝑊 = 𝑏𝑘ҧ

𝑘𝜋
𝑏ത𝑘 − 𝑏𝑘 𝑒 𝑊 𝐷
𝐷𝑘 = 𝑘𝜋 𝑘𝜋
−𝑊𝐷
𝑒 − 𝑒𝑊𝐷
2 𝑊 𝑘𝜋𝑥
where ത
𝑏𝑘 = න 𝑓 𝑥 sin 𝑑𝑥
𝑊 0 𝑊
Final solution: Laplace’s equation, Example 2 Problem A

∞ 𝑘𝜋 𝑘𝜋
− 𝑦 𝑦
𝑘𝜋𝑥 𝑘𝜋
𝑦
𝑘𝜋
𝐷 𝑒 𝑊 − 𝑒𝑊
𝑢𝐴 𝑥, 𝑦 = ෍ sin 𝑏𝑘 𝑒 𝑊 + 𝑏𝑘ҧ − 𝑏𝑘 𝑒 𝑊 𝑘𝜋 𝑘𝜋
𝑊 − 𝐷 𝐷
𝑘=1 𝑒 𝑊 − 𝑒𝑊

with

2 𝑊 𝑘𝜋𝑥
𝑏𝑘 = න 𝑔 𝑥 sin 𝑑𝑥
𝑊 0 𝑊

2 𝑊 𝑘𝜋𝑥
ത
𝑏𝑘 = න 𝑓 𝑥 sin 𝑑𝑥
𝑊 0 𝑊
Similarly, final solution: Laplace’s equation, Example 2 Problem B

∞ 𝑘𝜋 𝑘𝜋
− 𝑥 𝑥
𝑘𝜋𝑦 𝑘𝜋
𝑥
𝑘𝜋
𝑊 𝑒 𝐷 − 𝑒𝐷
𝑢𝐵 𝑥, 𝑦 = ෍ sin 𝑏𝑘 𝑒 𝐷 + 𝑏𝑘ҧ − 𝑏𝑘 𝑒 𝐷 𝑘𝜋 𝑘𝜋
𝐷 − 𝑊 𝑊
𝑘=1 𝑒 𝐷 − 𝑒𝐷

with

2 𝐷 𝑘𝜋𝑦
𝑏𝑘 = න 𝑙 𝑦 sin 𝑑𝑦
𝐷 0 𝐷

2 𝐷 𝑘𝜋𝑦
ത
𝑏𝑘 = න ℎ 𝑦 sin 𝑑𝑦
𝐷 0 𝑊
Final solution: Laplace’s equation, Example 2

𝑢 𝑥, 𝑦 = 𝑢𝐴 𝑥, 𝑦 + 𝑢𝐵 𝑥, 𝑦
∞ 𝑘𝜋 𝑘𝜋
− 𝑦
𝑘𝜋𝑥 𝑔 𝑥
𝑘𝜋
𝑓 𝑥 𝑔 𝑥
𝑘𝜋 𝑒 𝑊 − 𝑒 𝑊𝑦
= ෍ sin 𝑏𝑘 𝑒𝑊𝑦 + 𝑏𝑘 − 𝑏𝑘 𝑒𝑊𝐷 𝑘𝜋 𝑘𝜋
𝑊 −𝑊𝐷
𝑘=1 𝑒 − 𝑒 𝑊𝐷
∞ 𝑘𝜋 𝑘𝜋
−𝐷𝑥
𝑘𝜋𝑦 ℓ 𝑦
𝑘𝜋
ℎ 𝑦 ℓ 𝑦
𝑘𝜋 𝑒 −𝑒𝐷𝑥
+ ෍ sin 𝑏𝑘 𝑒𝐷𝑥 + 𝑏𝑘 − 𝑏𝑘 𝑒𝐷𝑊 𝑘𝜋 𝑘𝜋
𝐷 − 𝑊
𝑘=1 𝑒 𝐷 − 𝑒𝐷𝑊

with

𝑔 𝑥 2 𝑊 𝑘𝜋𝑥 ℓ 𝑦 2 𝐷 𝑘𝜋𝑦
𝑏𝑘 = න 𝑔 𝑥 sin 𝑑𝑥 𝑏𝑘 = න ℓ 𝑦 sin 𝑑𝑦
𝑊 0 𝑊 𝐷 0 𝐷

𝑓 𝑥 2 𝑊 𝑘𝜋𝑥 ℎ 𝑦 2 𝐷 𝑘𝜋𝑦
𝑏𝑘 = න 𝑓 𝑥 sin 𝑑𝑥 𝑏𝑘 = න ℎ 𝑦 sin 𝑑𝑦
𝑊 0 𝑊 𝐷 0 𝐷
Laplace
Example 3: Dirichlet nonhomogeneous
Note that this example is done incorrectly on purpose here in the slides. We follow “an
obvious” mistake. You must solve the problem correctly in the tutorial. Still, do not
underestimate the solving skills developed in working through this incorrect solution.

58
𝑢𝑥𝑥 + 𝑢𝑦𝑦 = 0
𝑢
on 𝑦
𝑥 ∈ 0, 𝐿
𝑦 ∈ 0, 𝐿

subject to 𝑥
2𝜋𝑦
1 𝑢(0, 𝑦) = sin
𝐿
2 𝑢(𝐿, 𝑦) = 1
𝜋𝑥
3 𝑢 𝑥, 0 = sin
2𝐿
𝑥
4 𝑢 𝑥, 𝐿 =
𝐿
𝑢𝑥𝑥 + 𝑢𝑦𝑦 = 0
𝑢
on 𝑦
𝑥 ∈ 0, 𝐿
𝑦 ∈ 0, 𝐿

subject to 𝑥
2𝜋𝑦
1 𝑢(0, 𝑦) = sin
𝐿
2 𝑢(𝐿, 𝑦) = 1
𝜋𝑥
3 𝑢 𝑥, 0 = sin
2𝐿
𝑥
4 𝑢 𝑥, 𝐿 =
𝐿
𝑢𝑥𝑥 + 𝑢𝑦𝑦 = 0
𝑢
on 𝑦
𝑥 ∈ 0, 𝐿
𝑦 ∈ 0, 𝐿

subject to 𝑥
1 𝑢(0, 𝑦) = 0

2 𝑢(𝐿, 𝑦) = 0
𝜋𝑥
3 𝑢 𝑥, 0 = sin
2𝐿
𝑥
4 𝑢 𝑥, 𝐿 =
𝐿
Final solution (from Example 2 Problem A): Laplace’s equation, Example 3
Problem A

∞
𝑘𝜋𝑥 𝑘𝜋 𝑏ത𝑘 𝑏𝑘 𝑘𝜋
𝑢𝐴 𝑥, 𝑦 = ෍ sin 𝑏𝑘 cosh 𝑦 + − sinh 𝑦
𝑊 𝑊 𝑘𝜋 𝑘𝜋 𝑊
𝑘=1 sinh 𝐷 tanh 𝐷
𝑊 𝑊
∞
𝑘𝜋𝑥 𝑘𝜋 𝑏ത𝑘 𝑏𝑘 𝑘𝜋
𝑢𝐴 𝑥, 𝑦 = ෍ sin 𝑏𝑘 cosh 𝑦 + − sinh 𝑦
𝐿 𝐿 sinh 𝑘𝜋 tanh 𝑘𝜋 𝐿
𝑘=1

with

2 𝐿 𝜋𝑥 𝑘𝜋𝑥 2 𝑊 𝑘𝜋𝑥
𝑏𝑘 = න sin sin 𝑑𝑥 𝑏𝑘 = න 𝑔 𝑥 sin 𝑑𝑥
𝐿 0 2𝐿 𝐿 𝑊 0 𝑊

2 𝐿𝑥 𝑘𝜋𝑥 2 𝑊 𝑘𝜋𝑥
ത
𝑏𝑘 = න sin 𝑑𝑥 ത
𝑏𝑘 = න 𝑓 𝑥 sin 𝑑𝑥
𝐿 0 𝐿 𝑊 𝑊 0 𝑊
𝑢𝑥𝑥 + 𝑢𝑦𝑦 = 0
𝑢
on 𝑦
𝑥 ∈ 0, 𝐿
𝑦 ∈ 0, 𝐿

subject to 𝑥
2𝜋𝑦
1 𝑢(0, 𝑦) = sin
𝐿
2 𝑢(𝐿, 𝑦) = 1
3 𝑢 𝑥, 0 = 0
4 𝑢 𝑥, 𝐿 = 0
From the superposition principle (and setting 𝐴𝑘 = 1 ∀ 𝑘):
∞
𝑘𝜋𝑦 𝑘𝜋 𝑘𝜋
𝑢 𝑥, 𝑦 = ෍ sin 𝐶𝑘 cosh 𝑥 + 𝐷𝑘 sinh 𝑥
𝐿 𝐿 𝐿
𝑘=1

2𝜋𝑦
Impose [1]: 𝑢 0, 𝑦 = sin
𝐿
∞
2𝜋𝑦 𝑘𝜋𝑦
𝑢 0, 𝑦 = sin = ෍ 𝐶𝑘 sin
𝐿 𝐿
𝑘=1

𝐶2 = 1 and 𝐶𝑘 = 0 ∀ 𝑘 ≠ 2.

More specific solution:

∞
2𝜋𝑦 2𝜋 2𝜋 𝑘𝜋𝑦 𝑘𝜋
𝑢 𝑥, 𝑦 = sin cosh 𝑥 + 𝐷2 sinh 𝑥 + ෍ sin 𝐷𝑘 sinh 𝑥
𝐿 𝐿 𝐿 𝐿 𝐿
𝑘=1
𝑘≠2
More specific solution:
∞
2𝜋𝑦 2𝜋 2𝜋 𝑘𝜋𝑦 𝑘𝜋
𝑢 𝑥, 𝑦 = sin cosh 𝑥 + 𝐷2 sinh 𝑥 + ෍ sin 𝐷𝑘 sinh 𝑥
𝐿 𝐿 𝐿 𝐿 𝐿
𝑘=1
𝑘≠2

Impose [2]: 𝑢(𝐿, 𝑦) = 1

∞
2𝜋𝑦 𝑘𝜋𝑦
𝑢 𝐿, 𝑦 = 1 = sin cosh 2𝜋 + 𝐷2 sinh 2𝜋 + ෍ sin 𝐷𝑘 sinh 𝑘𝜋
𝐿 𝐿
𝑘=1
𝑘≠2

∞
ധ
𝑘𝜋𝑦 2 𝐿 𝑘𝜋𝑦
1 = ෍ 𝑏𝑘 sin where ധ
𝑏𝑘 = න sin 𝑑𝑦
𝐿 𝐿 0 𝐿
𝑘=1

If 𝑘 = 2:
cosh 2𝜋 + 𝐷2 sinh 2𝜋 = 𝑏ധ2
𝑏ധ2 1
𝐷2 = −
sinh 2𝜋 tanh 2𝜋
else (∀ 𝑘 ≠ 2):
𝑏ധ𝑘
𝐷𝑘 =
sinh 𝑘𝜋
Final solution: Laplace’s equation, Example 3 Problem B

𝑢𝐵 𝑥, 𝑦
2𝜋𝑦 2𝜋 𝑏Ӗ2 1 2𝜋
= sin cosh 𝑥 + − sinh 𝑥
𝐿 𝐿 sinh 2𝜋 tanh 2𝜋 𝐿
∞
𝑘𝜋𝑦 𝑏Ӗ𝑘 𝑘𝜋
+ ෍ sin sinh 𝑥
𝐿 sinh 𝑘𝜋 𝐿
𝑘=1
𝑘≠2

with

2 𝐿 𝑘𝜋𝑦
𝑏ധ𝑘 = න sin 𝑑𝑦
𝐿 0 𝐿
Final solution: Laplace’s equation, Example 3

𝑢 𝑥, 𝑦 = 𝑢𝐴 𝑥, 𝑦 + 𝑢𝐵 𝑥, 𝑦
∞
𝑘𝜋𝑥 𝑘𝜋 𝑏ത𝑘 𝑏𝑘 𝑘𝜋
= ෍ sin 𝑏𝑘 cosh 𝑦 + − sinh 𝑦
𝐿 𝐿 sinh 𝑘𝜋 tanh 𝑘𝜋 𝐿
𝑘=1
2𝜋𝑦 2𝜋 𝑏Ӗ2 1 2𝜋
+ sin cosh 𝑥 + − sinh 𝑥
𝐿 𝐿 sinh 2𝜋 tanh 2𝜋 𝐿
∞
𝑘𝜋𝑦 𝑏Ӗ𝑘 𝑘𝜋
+ ෍ sin sinh 𝑥
𝐿 sinh 2𝜋 𝐿
𝑘=1
𝑘≠2

with
2 𝐿 𝜋𝑥 𝑘𝜋𝑥
𝑏𝑘 = න sin sin 𝑑𝑥
𝐿 0 2𝐿 𝐿
2 𝐿𝑥 𝑘𝜋𝑥
𝑏ത𝑘 = න sin 𝑑𝑥
𝐿 0 𝐿 𝑊
2 𝐿 𝑘𝜋𝑦
ധ
𝑏𝑘 = න sin 𝑑𝑦
𝐿 0 𝐿
 To think of an idea of
how to get this solution,
let us first find the
general solution of
Laplace’s equation.
Laplace
General solution (separation of variables, 𝑢 𝑥, 𝑦 = 𝑋 𝑥 𝑌(𝑦))

69
 Separation of variables:
Seek solutions of the form of a product of a function of 𝑥 and a function of 𝑡.

𝑢𝑥𝑥 + 𝑢𝑦𝑦 = 0 …(1)

Let 𝑢 𝑥, 𝑦 = 𝑋 𝑥 𝑌(𝑦) and substitute into (1):

𝑋 ′′ (𝑥)𝑌 𝑦 + 𝑋 𝑥 𝑌′′(𝑦) = 0 …(2)

Now we separate the terms containing 𝑥’s and 𝑦 ′ 𝑠 in (2):

𝑋 ′′ 𝑥 𝑌 ′′ (𝑦)
=− …(3)
𝑋 𝑥 𝑌(𝑦)
𝑋 ′′ 𝑥 𝑌 ′′ 𝑦
=−
𝑋 𝑥 𝑌 𝑦
Separation constant positive: 𝜇2

𝑋 ′′ 𝑥 − 𝜇2 𝑋 𝑥 = 0 𝑌 ′′ 𝑦 + 𝜇2 𝑌 𝑦 = 0
𝑋 𝑥 = 𝐴 sinh(𝜇𝑥) + 𝐵 cosh(𝜇𝑥) 𝑌(𝑦) = 𝐶 sin(𝜇𝑦) + 𝐷 cos(𝜇𝑦)

∞
𝑢1 = ෍ 𝐴𝑘 sinh 𝜇𝑘 𝑥 + 𝐵𝑘 cosh 𝜇𝑥 𝐶𝑘 sin 𝜇𝑘 𝑦 + 𝐷𝑘 cos 𝜇𝑘 𝑦
𝑘=1

Separation constant negative: −𝛽 2

𝑋 ′′ 𝑥 + 𝛽2 𝑋 𝑥 = 0 𝑌 ′′ 𝑦 − 𝛽2 𝑌 𝑦 = 0
𝑋 𝑥 = 𝐸 sin(𝛽𝑥) + 𝐹 cos(𝛽𝑥) 𝑌(𝑦) = 𝐺 sinh(𝛽𝑦) + 𝐻 cosh(𝛽𝑦)

∞
𝑢2 = ෍ 𝐸𝑘 sin 𝛽𝑘 𝑥 + 𝐹𝑘 cos 𝛽𝑘 𝑥 𝐺𝑘 sinh 𝛽𝑘 𝑦 + 𝐻𝑘 cosh 𝛽𝑘 𝑦
𝑘=1

Separation constant zero

𝑋 ′′ 𝑥 = 0 ⇒ 𝑋 𝑥 = 𝑀𝑥 + 𝑁 𝑌 ′′ 𝑦 = 0 ⇒ 𝑌 𝑦 = 𝑂𝑦 + 𝑃

𝑢3 = 𝑄𝑥𝑦 + 𝑅𝑥 + 𝑆𝑦 + 𝑇
𝑢
𝑦
𝑢3 = 𝑄𝑥𝑦 + 𝑅𝑥 + 𝑆𝑦 + 𝑇
𝑢
𝑦
𝜋𝑥
𝑢 𝑥, 0 = sin
2𝐿
𝑥
𝑥
𝑢 𝑥, 0 =
𝐿
𝑥
𝑢
𝑦𝑢 ∞
=෍ 𝐴𝑘 sinh 𝜇𝑘 𝑥 + 𝐵𝑘 cosh 𝜇𝑥 𝐶𝑘 sin 𝜇𝑘 𝑦 + 𝐷𝑘 cos 𝜇𝑘 𝑦
1
𝑘=1
𝑢
𝑦

𝑥 𝜋𝑥 𝑥
𝑢 𝑥, 0 = sin −
2𝐿 𝐿
𝑥
∞
𝑢2 = ෍ 𝐸𝑘 sin 𝛽𝑘 𝑥 + 𝐹𝑘 cos 𝛽𝑘 𝑥 𝐺𝑘 sinh 𝛽𝑘 𝑦 + 𝐻𝑘 cosh 𝛽𝑘 𝑦
𝑘=1
Laplace
Polar coordinates

73
Cartesian - Polar
y
𝑟2 = 𝑥2 + 𝑦2
𝑦
tan 𝜃 =
𝑥

𝑥 = 𝑟 cos 𝜃 x

𝑦 = 𝑟 sin 𝜃

𝑢 𝑟 𝑥, 𝑦 , 𝜃 𝑥, 𝑦 𝑢𝑟 𝑟 𝑥, 𝑦 , 𝜃 𝑥, 𝑦 𝑢𝑟𝑟 𝑟 𝑥, 𝑦 , 𝜃 𝑥, 𝑦
𝑢𝜃 𝑟 𝑥, 𝑦 , 𝜃 𝑥, 𝑦 𝑢𝑟𝜃 𝑟 𝑥, 𝑦 , 𝜃 𝑥, 𝑦
𝑢𝜃𝜃 𝑟 𝑥, 𝑦 , 𝜃 𝑥, 𝑦
Cartesian - Polar 𝑢𝑥𝑥 + 𝑢𝑦𝑦 = 0
𝑢𝑥 = 𝑢𝑟 𝑟𝑥 + 𝑢𝜃 𝜃𝑥 𝑢 𝑟 𝑥 ,𝜃 𝑥 𝑢 𝑟 𝑥, 𝑦 , 𝜃 𝑥, 𝑦
𝑢𝑟 𝑟 𝑥, 𝑦 , 𝜃 𝑥, 𝑦
𝑢𝜃 𝑟 𝑥, 𝑦 , 𝜃 𝑥, 𝑦
𝑢𝑟𝑟 𝑟 𝑥, 𝑦 , 𝜃 𝑥, 𝑦
𝑢𝑟𝜃 𝑟 𝑥, 𝑦 , 𝜃 𝑥, 𝑦
𝑢𝜃𝜃 𝑟 𝑥, 𝑦 , 𝜃 𝑥, 𝑦

𝑟2 = 𝑥2 + 𝑦2
𝑦
tan 𝜃 =
𝑥
𝑥 = 𝑟 cos 𝜃
𝑦 = 𝑟 sin 𝜃
Cartesian - Polar 𝑢𝑥𝑥 + 𝑢𝑦𝑦 = 0
𝑢𝑥 = 𝑢𝑟 𝑟𝑥 + 𝑢𝜃 𝜃𝑥 𝑢 𝑟 𝑥, 𝑦 , 𝜃 𝑥, 𝑦
𝜕𝑢𝑟 𝜕𝑢𝜃 𝑢𝑟 𝑟 𝑥, 𝑦 , 𝜃 𝑥, 𝑦
𝑢𝑥𝑥 = 𝑢𝑟 𝑟𝑥𝑥 + 𝑟𝑥 + 𝑢𝜃 𝜃𝑥𝑥 + 𝜃𝑥
𝜕𝑥 𝜕𝑥
𝑢𝜃 𝑟 𝑥, 𝑦 , 𝜃 𝑥, 𝑦
𝑢𝑟𝑟 𝑟 𝑥, 𝑦 , 𝜃 𝑥, 𝑦
𝑢𝑟𝜃 𝑟 𝑥, 𝑦 , 𝜃 𝑥, 𝑦
𝑢𝜃𝜃 𝑟 𝑥, 𝑦 , 𝜃 𝑥, 𝑦

𝑟2 = 𝑥2 + 𝑦2
𝑦
tan 𝜃 =
𝑥
𝑥 = 𝑟 cos 𝜃
𝑦 = 𝑟 sin 𝜃
Cartesian - Polar
𝑢𝑥 = 𝑢𝑟 𝑟𝑥 + 𝑢𝜃 𝜃𝑥
𝜕𝑢𝑟 𝜕𝑢𝜃
𝑢𝑥𝑥 = 𝑢𝑟 𝑟𝑥𝑥 + 𝑟𝑥 + 𝑢𝜃 𝜃𝑥𝑥 + 𝜃𝑥
𝜕𝑥 𝜕𝑥
= 𝑢𝑟 𝑟𝑥𝑥 + 𝑟𝑥 𝑢𝑟𝑟 𝑟𝑥 + 𝑢𝑟𝜃 𝜃𝑥 + 𝑢𝜃 𝜃𝑥𝑥 + 𝜃𝑥 𝑢𝜃𝑟 𝑟𝑥 + 𝑢𝜃𝜃 𝜃𝑥

𝑢 𝑟 𝑥, 𝑦 , 𝜃 𝑥, 𝑦
𝑢𝑟 𝑟 𝑥, 𝑦 , 𝜃 𝑥, 𝑦
𝑢𝜃 𝑟 𝑥, 𝑦 , 𝜃 𝑥, 𝑦
𝑢𝑟𝑟 𝑟 𝑥, 𝑦 , 𝜃 𝑥, 𝑦
𝑢𝑟𝜃 𝑟 𝑥, 𝑦 , 𝜃 𝑥, 𝑦
𝑢𝜃𝜃 𝑟 𝑥, 𝑦 , 𝜃 𝑥, 𝑦
Cartesian - Polar
𝑢𝑥 = 𝑢𝑟 𝑟𝑥 + 𝑢𝜃 𝜃𝑥
𝜕𝑢𝑟 𝜕𝑢𝜃
𝑢𝑥𝑥 = 𝑢𝑟 𝑟𝑥𝑥 + 𝑟𝑥 + 𝑢𝜃 𝜃𝑥𝑥 + 𝜃𝑥
𝜕𝑥 𝜕𝑥
= 𝑢𝑟 𝑟𝑥𝑥 + 𝑟𝑥 𝑢𝑟𝑟 𝑟𝑥 + 𝑢𝑟𝜃 𝜃𝑥 + 𝑢𝜃 𝜃𝑥𝑥 + 𝜃𝑥 𝑢𝜃𝑟 𝑟𝑥 + 𝑢𝜃𝜃 𝜃𝑥

= 𝑢𝑟𝑟 𝑟𝑥2 + 2𝑢𝑟𝜃 𝑟𝑥 𝜃𝑥 + 𝑢𝜃𝜃 𝜃𝑥2 + 𝑢𝑟 𝑟𝑥𝑥 + 𝑢𝜃 𝜃𝑥𝑥

𝑢 𝑟 𝑥, 𝑦 , 𝜃 𝑥, 𝑦
𝑢𝑟 𝑟 𝑥, 𝑦 , 𝜃 𝑥, 𝑦
𝑢𝜃 𝑟 𝑥, 𝑦 , 𝜃 𝑥, 𝑦
𝑢𝑟𝑟 𝑟 𝑥, 𝑦 , 𝜃 𝑥, 𝑦
𝑢𝑟𝜃 𝑟 𝑥, 𝑦 , 𝜃 𝑥, 𝑦
𝑢𝜃𝜃 𝑟 𝑥, 𝑦 , 𝜃 𝑥, 𝑦
Cartesian - Polar
𝑢𝑥 = 𝑢𝑟 𝑟𝑥 + 𝑢𝜃 𝜃𝑥
𝜕𝑢𝑟 𝜕𝑢𝜃
𝑢𝑥𝑥 = 𝑢𝑟 𝑟𝑥𝑥 + 𝑟𝑥 + 𝑢𝜃 𝜃𝑥𝑥 + 𝜃𝑥
𝜕𝑥 𝜕𝑥
= 𝑢𝑟 𝑟𝑥𝑥 + 𝑟𝑥 𝑢𝑟𝑟 𝑟𝑥 + 𝑢𝑟𝜃 𝜃𝑥 + 𝑢𝜃 𝜃𝑥𝑥 + 𝜃𝑥 𝑢𝜃𝑟 𝑟𝑥 + 𝑢𝜃𝜃 𝜃𝑥

= 𝑢𝑟𝑟 𝑟𝑥2 + 2𝑢𝑟𝜃 𝑟𝑥 𝜃𝑥 + 𝑢𝜃𝜃 𝜃𝑥2 + 𝑢𝑟 𝑟𝑥𝑥 + 𝑢𝜃 𝜃𝑥𝑥

Similarly
𝑢𝑦𝑦 = 𝑢𝑟𝑟 𝑟𝑦2 + 2𝑢𝑟𝜃 𝑟𝑦 𝜃𝑦 + 𝑢𝜃𝜃 𝜃𝑦2 + 𝑢𝑟 𝑟𝑦𝑦 + 𝑢𝜃 𝜃𝑦𝑦
Cartesian - Polar
𝑢𝑥𝑥 = 𝑢𝑟𝑟 𝑟𝑥2 + 2𝑢𝑟𝜃 𝑟𝑥 𝜃𝑥 + 𝑢𝜃𝜃 𝜃𝑥2 + 𝑢𝑟 𝑟𝑥𝑥 + 𝑢𝜃 𝜃𝑥𝑥
𝑢𝑦𝑦 = 𝑢𝑟𝑟 𝑟𝑦2 + 2𝑢𝑟𝜃 𝑟𝑦 𝜃𝑦 + 𝑢𝜃𝜃 𝜃𝑦2 + 𝑢𝑟 𝑟𝑦𝑦 + 𝑢𝜃 𝜃𝑦𝑦

𝑟2 = 𝑥2 + 𝑦2
𝜕
2𝑟𝑟𝑥 = 2𝑥
𝜕𝑥 𝑥
𝑟𝑥 =
𝑟 𝑢𝑥𝑥 + 𝑢𝑦𝑦 = 0
𝜕2
2
2𝑟[𝑟𝑥𝑥 ሿ + 𝑟𝑥 [2𝑟𝑥 ሿ = 2 𝑟2 = 𝑥2 + 𝑦2
𝜕𝑥
1 𝑟𝑥2 1 𝑥 2 𝑦
𝑟𝑥𝑥 = − = − 3 tan 𝜃 =
𝑟 𝑟 𝑟 𝑟 𝑥
Similarly 𝑥 = 𝑟 cos 𝜃
𝑦 1 𝑦2
𝑟𝑦 = 𝑟𝑦𝑦 = − 3 𝑦 = 𝑟 sin 𝜃
𝑟 𝑟 𝑟
 𝑥
Cartesian - Polar 𝑟𝑥 =
𝑟
𝑦
𝑢𝑥𝑥 = 𝑢𝑟𝑟 𝑟𝑥2 + 2𝑢𝑟𝜃 𝑟𝑥 𝜃𝑥 + 𝑢𝜃𝜃 𝜃𝑥2 + 𝑢𝑟 𝑟𝑥𝑥 + 𝑢𝜃 𝜃𝑥𝑥 𝑟𝑦 =
𝑟
𝑢𝑦𝑦 = 𝑢𝑟𝑟 𝑟𝑦2 + 2𝑢𝑟𝜃 𝑟𝑦 𝜃𝑦 + 𝑢𝜃𝜃 𝜃𝑦2 + 𝑢𝑟 𝑟𝑦𝑦 + 𝑢𝜃 𝜃𝑦𝑦

𝑦
tan 𝜃 =
𝑥
𝜕 𝑦
𝜕𝑥 sec 2 𝜃 𝜃𝑥=− 2 𝑢𝑥𝑥 + 𝑢𝑦𝑦 = 0
𝑥
𝑦 2
𝑦
𝜃𝑥 = − 2 cos 𝜃 = − 2
𝑥 𝑟 𝑟2 = 𝑥2 + 𝑦2
𝑦
tan 𝜃 =
𝜕2 2𝑦 2𝑥𝑦 𝑥
𝜕𝑥 2 𝜃𝑥𝑥 = 3 𝑟𝑥 = 4 𝑥 = 𝑟 cos 𝜃
𝑟 𝑟
𝑦 = 𝑟 sin 𝜃
 𝑥
Cartesian - Polar 𝑟𝑥 =
𝑟
𝑦
𝑢𝑥𝑥 = 𝑢𝑟𝑟 𝑟𝑥2 + 2𝑢𝑟𝜃 𝑟𝑥 𝜃𝑥 + 𝑢𝜃𝜃 𝜃𝑥2 + 𝑢𝑟 𝑟𝑥𝑥 + 𝑢𝜃 𝜃𝑥𝑥 𝑟𝑦 =
𝑟
𝑢𝑦𝑦 = 𝑢𝑟𝑟 𝑟𝑦2 + 2𝑢𝑟𝜃 𝑟𝑦 𝜃𝑦 + 𝑢𝜃𝜃 𝜃𝑦2 + 𝑢𝑟 𝑟𝑦𝑦 + 𝑢𝜃 𝜃𝑦𝑦

𝑦
tan 𝜃 =
𝑥
𝜕 2
1
sec 𝜃 𝜃𝑦 =
𝜕𝑦 𝑥 𝑢𝑥𝑥 + 𝑢𝑦𝑦 = 0
1 2
𝑥
𝜃𝑦 = cos 𝜃 = 2 𝑟2 = 𝑥2 + 𝑦2
𝑥 𝑟
𝑦
tan 𝜃 =
𝜕2 2𝑥 2𝑥𝑦 𝑥
𝜕𝑦 2 𝜃𝑦𝑦 = − 3 𝑟𝑦 = − 4 𝑥 = 𝑟 cos 𝜃
𝑟 𝑟
𝑦 = 𝑟 sin 𝜃
Cartesian - Polar
𝑢𝑥𝑥 = 𝑢𝑟𝑟 𝑟𝑥2 + 2𝑢𝑟𝜃 𝑟𝑥 𝜃𝑥 + 𝑢𝜃𝜃 𝜃𝑥2 + 𝑢𝑟 𝑟𝑥𝑥 + 𝑢𝜃 𝜃𝑥𝑥

𝑢𝑦𝑦 = 𝑢𝑟𝑟 𝑟𝑦2 + 2𝑢𝑟𝜃 𝑟𝑦 𝜃𝑦 + 𝑢𝜃𝜃 𝜃𝑦2 + 𝑢𝑟 𝑟𝑦𝑦 + 𝑢𝜃 𝜃𝑦𝑦

𝑥 2 𝑥 𝑦 𝑦 2 1 𝑥2 2𝑥𝑦
𝑢𝑥𝑥 = 𝑢𝑟𝑟 + 2𝑢𝑟𝜃 − 2 + 𝑢𝜃𝜃 − 2 + 𝑢𝑟 − + 𝑢𝜃
𝑟 𝑟 𝑟 𝑟 𝑟 𝑟3 𝑟4
𝑦 2 𝑦 𝑥 𝑥 2 1 𝑦2 2𝑥𝑦
𝑢𝑦𝑦 = 𝑢𝑟𝑟 + 2𝑢𝑟𝜃 + 𝑢𝜃𝜃 2 + 𝑢𝑟 − + 𝑢𝜃 − 4
𝑟 𝑟 𝑟2 𝑟 𝑟 𝑟3 𝑟

𝑦2 + 𝑥2 𝑦2 + 𝑥2 2 1 2
𝑢𝑥𝑥 + 𝑢𝑦𝑦 = 𝑢𝑟𝑟 2
+ 2𝑢𝑟𝜃 0 + 𝑢𝜃𝜃 4
+ 𝑢𝑟 − 3 (𝑥 + 𝑦 2 ) + 𝑢𝜃 0
𝑟 𝑟 𝑟 𝑟
𝑟2 𝑟2 2 𝑟2
= 𝑢𝑟𝑟 + 𝑢𝜃𝜃 4 + 𝑢𝑟 −
𝑟2 𝑟 𝑟 𝑟3

𝑢𝜃𝜃 𝑢𝑟
𝑢𝑥𝑥 + 𝑢𝑦𝑦 = 𝑢𝑟𝑟 + 2 + 𝑥 𝑦 1 𝑥 2 𝜃 = 2𝑥𝑦
𝑟 𝑟 𝑟𝑥 = 𝜃𝑥 = − 2 𝑟𝑥𝑥 = − 3 𝑥𝑥 𝑟4
𝑟 𝑟 𝑟 𝑟
𝑦 𝑥 1 𝑦2 2𝑥𝑦
𝑟𝑦 = 𝜃𝑦 = 2 𝑟𝑦𝑦 = − 3 𝜃𝑦𝑦 = − 4
𝑟 𝑟 𝑟 𝑟 𝑟
Laplace in Polar coordinates:
1 1
𝑢𝑟𝑟 + 𝑢𝑟 + 2 𝑢𝜃𝜃 = 0
𝑟 𝑟
Boundary value problem:
1 1
𝑢𝑟𝑟 + 𝑢𝑟 + 2 𝑢𝜃𝜃 = 0
𝑟 𝑟
𝑅
on 𝑟 ∈ 0, 𝑅 and 𝜃 ∈ [0,2𝜋)
subject to 𝑓 𝜃
[1] 𝑢 𝑅, 𝜃 = 𝑓 𝜃

Separation of variables
Searching for solutions of the form 𝑢 𝑟, 𝜃 = 𝑅 𝑟 Θ(𝜃)
1 ′ 1
𝑅′′ Θ + 𝑅 Θ + 2 𝑅Θ′′ = 0
𝑟 𝑟 𝑟2
2 ′′ ′ ′′
×
𝑟 𝑅 𝑟𝑅 Θ 𝑅Θ 𝑟 2 𝑅′′ 𝑟𝑅′ Θ′′
+ + =0 − + = = −𝜆
𝑅 𝑅 Θ 𝑅 𝑅 Θ
 Although this is not a regular Sturm–Liouville problem,
Θ′′ nonetheless the problem generates eigenvalues and
= −𝜆 eigenfunctions. Our physical intuition leads to implicit
Θ conditions we can impose.

Setting 𝜆 = 0: Θ′′ = 0

Θ 𝜃 = 𝐴𝜃 + 𝐵

Solution must be periodic in 𝜃.

Therefore, 𝐴 = 0.
𝑢 𝑟, 𝜃
𝑢(𝑟, 𝜃 + 2𝜋)
Eigenvalue: 𝜆 = 0

Eigenfunction: Θ(𝜃) = 𝐵
 Although this is not a regular Sturm–Liouville problem,
Θ′′ nonetheless the problem generates eigenvalues and
= −𝜆 eigenfunctions. Our physical intuition leads to implicit
Θ conditions we can impose.

Setting 𝜆 = −𝛼 2 : Θ′′ − 𝛼 2 Θ = 0

Θ 𝜃 = 𝐴 sinh 𝛼𝜃 + 𝐵 cosh 𝛼𝜃

Solution must be periodic in 𝜃.

Therefore, 𝐴 = 0 and 𝐵 = 0.
𝑢 𝑟, 𝜃
𝑢(𝑟, 𝜃 + 2𝜋)

Θ 𝜃 =0 Trivial solution
 Although this is not a regular Sturm–Liouville problem,
Θ′′ nonetheless the problem generates eigenvalues and
= −𝜆 eigenfunctions. Our physical intuition leads to implicit
Θ conditions we can impose.

Setting 𝜆 = 𝛼 2 : Θ′′ + 𝛼 2 Θ = 0

Θ 𝜃 = 𝐴 sin 𝛼𝜃 + 𝐵 cos 𝛼𝜃

Solution must be 2𝜋-periodic.

Therefore, 𝛼 = 𝑘 where 𝑘 = 1,2,3, …
𝑢 𝑟, 𝜃
𝑢(𝑟, 𝜃 + 2𝜋)

Eigenvalues: 𝜆𝑘 = 𝑘 2

Eigenfunctions: Θ𝑘 𝜃 = 𝐴𝑘 sin(𝑘𝜃) + 𝐵𝑘 cos(𝑘𝜃)

 𝑟 2 𝑅′′ 𝑟𝑅′ Eigenvalue: 𝜆 = 0
+ =𝜆
𝑅 𝑅 Eigenvalues: 𝜆𝑘 = 𝑘 2

Setting 𝜆𝑘 = 𝑘 2 : 𝑟 2 𝑅′′ + 𝑟𝑅′ − 𝑘 2 𝑅 = 0

 𝑟 2 𝑅′′ + 𝑟𝑅′ − 𝑘 2 𝑅 = 0 Eigenvalues: 𝜆𝑘 = 𝑘 2

Cauchy-Euler equation
𝑑 𝑛𝑦 𝑑 𝑛−1 𝑦
𝑎𝑛 𝑥 𝑛 𝑛 + 𝑎𝑛−1 𝑥 𝑛−1 𝑛−1 + ⋯ + 𝑎0 𝑦 = 0
𝑑𝑥 𝑑𝑥
Trial solution 𝑦 = 𝑥 𝑚

Auxiliary equation
Solve 𝑚.
Find general solution.
 𝑟 2 𝑅′′ + 𝑟𝑅′ − 𝑘 2 𝑅 = 0 Eigenvalues: 𝜆𝑘 = 𝑘 2

Cauchy-Euler equation
𝑎𝑥 2 𝑦 ′′ + 𝑏𝑥𝑦 ′ + 𝑐𝑦 = 0

Trial solution 𝑦 = 𝑥 𝑚
𝑎𝑥 2 𝑚(𝑚 − 1)𝑥 𝑚−2 + 𝑏𝑥 𝑚𝑥 𝑚−1 + 𝑐𝑥 𝑚 = 0

𝑥𝑚 𝑎 𝑚 𝑚 − 1 + 𝑏𝑚 + 𝑐 = 0

Auxiliary equation
𝑎 𝑚 𝑚−1 + 𝑏𝑚 + 𝑐 = 0
𝑚2 𝑎 + 𝑚(𝑏 − 𝑎) + 𝑐 = 0
 𝑟 2 𝑅′′ + 𝑟𝑅′ − 𝑘 2 𝑅 = 0 Eigenvalues: 𝜆𝑘 = 𝑘 2

Cauchy-Euler equation
Auxiliary equation
𝑚2 𝑎 + 𝑚(𝑏 − 𝑎) + 𝑐 = 0
Solve 𝑚
𝑎 − 𝑏 ± 𝑏 2 − 2𝑎𝑏 + 𝑎2 − 4𝑎𝑐
𝑚=
2𝑎

2
𝑎−𝑏 𝑏 𝑏 1 𝑐
𝑚= ± − + −
2𝑎 2𝑎 2𝑎 4 𝑎

2
𝑏 𝑏 1 𝑐
𝛿= − + −
2𝑎 2𝑎 4 𝑎
 𝑟 2 𝑅′′ + 𝑟𝑅′ − 𝑘 2 𝑅 = 0 Eigenvalues: 𝜆𝑘 = 𝑘 2

Cauchy-Euler equation
𝑎𝑥 2 𝑦 ′′ + 𝑏𝑥𝑦 ′ + 𝑐𝑦 = 0 (1)𝑟 2 𝑅′′ + 1 𝑟𝑅′ + −𝑘 2 𝑅 = 0
2
𝑏 𝑏 1 𝑐 1 1 1
𝛿= − + − 𝛿 = − + + 𝑘2 = 𝑘2
2𝑎 2𝑎 4 𝑎 4 2 4

2 𝑚 = ±𝑘
𝑎−𝑏 𝑏 𝑏 1 𝑐
𝑚= ± − + −
2𝑎 2𝑎 2𝑎 4 𝑎

Trial solution 𝑦 = 𝑥 𝑚 𝑅(𝑟) = 𝐶𝑘 𝑟 𝑘 + 𝐷𝑘 𝑟 −𝑘

 𝑟 2 𝑅′′ + 𝑟𝑅′ = 0 Eigenvalue: 𝜆 = 0

Cauchy-Euler equation
First solve 𝑟 2 𝑊′ + 𝑟𝑊 = 0 where 𝑊 = 𝑅′.
1
‫𝑟𝑑𝑟׬‬
Integration Factor: 𝑒 = 𝑒 ln(𝑟) = 𝑟

Multiply both sides of DE by integration factor:

𝑑
𝑊𝑟 = 0 ⇒ 𝑊𝑟 = 𝑐1
𝑑𝑟

𝑑𝑅 𝑐1
= ⇒ 𝑅 = 𝑐1 ln(𝑟) + 𝑐2
𝑑𝑟 𝑟
Laplace: General solution (annulus or disk)
∞

𝑢 𝑟, 𝜃 = ෍ (𝐴𝑘 sin 𝑘𝜃 + 𝐵𝑘 cos(𝑘𝜃)) 𝐶𝑘 𝑟 𝑘 + 𝐷𝑘 𝑟 −𝑘 + 𝐵 𝑐1 ln(𝑟) + 𝑐2

𝑘=1
Laplace: General solution (annulus or disk)
∞

𝑢 𝑟, 𝜃 = ෍ (𝐴𝑘 sin 𝑘𝜃 + 𝐵𝑘 cos(𝑘𝜃)) 𝐶𝑘 𝑟 𝑘 + 𝐷𝑘 𝑟 −𝑘 + 𝐸 ln 𝑟 + 𝐹

𝑘=1

Laplace: General solution: Disk

Another implicit condition: 𝑢(0, 𝜃) is finite
Therefore 𝐷𝑘 = 0 ∀ 𝑘 and 𝐸 = 0.

Setting 𝐶𝑘 = 1 ∀ 𝑘:
∞

𝑢 𝑟, 𝜃 = ෍ 𝑟 𝑘 (𝐴𝑘 sin 𝑘𝜃 + 𝐵𝑘 cos(𝑘𝜃)) + 𝐹

𝑘=1
Laplace: General solution: Disk
∞

𝑢 𝑟, 𝜃 = ෍ 𝑟 𝑘 (𝐴𝑘 sin 𝑘𝜃 + 𝐵𝑘 cos(𝑘𝜃)) + 𝐹

𝑘=1

Impose: [1] 𝑢 𝑅, 𝜃 = 𝑓 𝜃
∞

𝑢 𝑅, 𝜃 = 𝑓(𝜃) = ෍ 𝑅𝑘 (𝐴𝑘 sin 𝑘𝜃 + 𝐵𝑘 cos(𝑘𝜃)) + 𝐹

𝑘=1

Fourier series
∞
𝑎0
𝑓 𝜃 = + ෍ 𝑏𝑘 sin 𝑘𝜃 + 𝑎𝑘 cos 𝑘𝜃
2
𝑘=1
Laplace: General solution: Disk
∞

𝑢 𝑟, 𝜃 = ෍ 𝑟 𝑘 (𝐴𝑘 sin 𝑘𝜃 + 𝐵𝑘 cos(𝑘𝜃)) + 𝐹

𝑘=1
∞

𝑢 𝑅, 𝜃 = 𝑓(𝜃) = ෍ 𝑅𝑘 (𝐴𝑘 sin 𝑘𝜃 + 𝐵𝑘 cos(𝑘𝜃)) + 𝐹

𝑘=1
∞
𝑎0
= + ෍ 𝑏𝑘 sin 𝑘𝜃 + 𝑎𝑘 cos 𝑘𝜃
2
𝑘=1
Therefore:
𝑎0 1 𝜋
𝐹= 𝑎0 = න 𝑓 𝜃 𝑑𝜃
2 𝜋 −𝜋
𝑏𝑘 1 𝜋
𝐴𝑘 = 𝑘 𝑏𝑘 = න 𝑓 𝜃 sin(𝑘𝜃) 𝑑𝜃
𝑅 𝜋 −𝜋
𝑎𝑘 1 𝜋
𝐵𝑘 = 𝑘 𝑎𝑘 = න 𝑓 𝜃 cos(𝑘𝜃) 𝑑𝜃
𝑅 𝜋 −𝜋
Laplace: General solution: Disk
∞

𝑢 𝑟, 𝜃 = ෍ 𝑟 𝑘 (𝐴𝑘 sin 𝑘𝜃 + 𝐵𝑘 cos(𝑘𝜃)) + 𝐹

𝑘=1
∞

𝑢 𝑅, 𝜃 = 𝑓(𝜃) = ෍ 𝑅𝑘 (𝐴𝑘 sin 𝑘𝜃 + 𝐵𝑘 cos(𝑘𝜃)) + 𝐹

𝑘=1
∞
𝑎0
= + ෍ 𝑏𝑘 sin 𝑘𝜃 + 𝑎𝑘 cos 𝑘𝜃
2
𝑘=1
Therefore:
𝑎0 1 2𝜋
𝐹= 𝑎0 = න 𝑓 𝜃 𝑑𝜃
2 𝜋 0
𝑏𝑘 1 2𝜋
𝐴𝑘 = 𝑘 𝑏𝑘 = න 𝑓 𝜃 sin(𝑘𝜃) 𝑑𝜃
𝑅 𝜋 0
𝑎𝑘 1 2𝜋
𝐵𝑘 = 𝑘 𝑎𝑘 = න 𝑓 𝜃 cos(𝑘𝜃) 𝑑𝜃
𝑅 𝜋 0
Laplace: Final solution: Disk
∞
𝑎0 𝑟 𝑘
𝑢 𝑟, 𝜃 = +෍ (𝑎𝑘 cos 𝑘𝜃 + 𝑏𝑘 sin 𝑘𝜃 )
2 𝑅
𝑘=1

where
1 2𝜋
𝑎0 = න 𝑓 𝜃 𝑑𝜃
𝜋 0
1 2𝜋
𝑎𝑘 = න 𝑓 𝜃 cos(𝑘𝜃) 𝑑𝜃
𝜋 0
1 2𝜋
𝑏𝑘 = න 𝑓 𝜃 sin(𝑘𝜃) 𝑑𝜃
𝜋 0