BMCG 1013

DIFFERENTIAL EQUATIONS

APPLICATION OF
PARTIAL DIFFERENTIAL EQUATIONS

Ser Lee Loh

[email protected]
Lesson Outcomes

Upon completion of this lesson, students should be able to:

• solve heat equation on the heated rod problems.
• solve wave equation on the vibrating spring problems.

2
5.3 Applications of PDE
5.3.1 The Heat Equation

In this section, we will study the heat flow problem in a long uniform rod. In
general, this problem depends on the initial distribution of temperature and
the physical properties of rod. The physical properties of rod refers to the
thermal diffusivity of the material where it measures the rate of transfer of
heat of the material from the hot end to the cold end.

By using the method of separation of variables, the solution function 𝑢𝑢 𝑥𝑥, 𝑡𝑡

that represents temperature at a point 𝑥𝑥 along the rod at time 𝑡𝑡 can be
determined.
3
Heat Conduction

A uniform homogeneous rod or metal bar is made up of infinite numbers of

molecules that are interconnected by cohesive force. When heat is applied to
one end of rod, the heat tends to transfer due to heat flows from molecule to
molecule. Heat moves rapidly in a substance with high thermal diffusivity.

In the following section, the heat transfer of the heated rod with zero
temperature at the endpoints will be discussed. Graphical interpretation will
be illustrated in the next slide.

4
 Graphical Interpretation
Given a rod with length 𝐿𝐿 with the setup and condition as shown in Figure 5.1,
we aim to find the temperature of the rod at position point 𝑥𝑥 after some time 𝑡𝑡.

Boundary condition: Boundary condition:

𝑢𝑢 0, 𝑡𝑡 = 0 𝑢𝑢 𝐿𝐿, 𝑡𝑡 = 0
Always zero 𝑥𝑥 Always zero
0 Length of rod 𝐿𝐿
temperature temperature
at this end point Initial condition: at this end point
𝑢𝑢 𝑥𝑥, 0 = 𝑓𝑓(𝑥𝑥)
At 𝑡𝑡 = 0, temperature of rod
at position 𝑥𝑥 is set up as 𝑓𝑓(𝑥𝑥)

Figure 5.1: Graphical interpretation of experimental setup for heat conduction

5
For heat equation, the function 𝑢𝑢 𝑥𝑥, 𝑡𝑡 represents temperature at a point
𝑥𝑥 along the rod at time 𝑡𝑡.

For example:

𝑢𝑢 0, 5 = 10 means at the time 𝑡𝑡 = 5, the temperature at the point 𝑥𝑥 = 0

is 10℃. (The unit of measurements of length and time depend on the
unit of measurement of thermal diffusivity)

6
Theorem: The heated rod with zero temperature at the endpoints
The partial differential equation
𝑢𝑢𝑡𝑡 = 𝑘𝑘𝑢𝑢𝑥𝑥𝑥𝑥 , 0 < 𝑥𝑥 < 𝐿𝐿, 𝑡𝑡 > 0
for a rod of length 𝐿𝐿 and thermal diffusivity of 𝑘𝑘, with boundary conditions
𝑢𝑢 0, 𝑡𝑡 = 𝑢𝑢 𝐿𝐿, 𝑡𝑡 = 0, 𝑡𝑡 > 0
and initial condition
𝑢𝑢 𝑥𝑥, 0 = 𝑓𝑓(𝑥𝑥), 0 < 𝑥𝑥 < 𝐿𝐿
has the series solution
∞
𝑛𝑛𝑛𝑛𝑛𝑛 − 𝑛𝑛2𝜋𝜋22𝑘𝑘𝑘𝑘
𝑢𝑢 𝑥𝑥, 𝑡𝑡 = � 𝑏𝑏𝑛𝑛 sin 𝑒𝑒 𝐿𝐿 (5.1)
𝐿𝐿
𝑛𝑛=1

where 𝑏𝑏𝑛𝑛 is the Fourier series coefficients obtained by

2 𝐿𝐿 𝑛𝑛𝑛𝑛𝑛𝑛
𝑏𝑏𝑛𝑛 = � 𝑓𝑓(𝑥𝑥) sin 𝑑𝑑𝑑𝑑, 𝑛𝑛 = 1,2,3, … (5.2)
𝐿𝐿 0 𝐿𝐿 7
5.3.1 The Heat Equation

Example 5.13:

Solve the following PDE for the heat transfer of a metal bar with zero
temperature at both ends by method of separation of variables:

1
𝑢𝑢𝑡𝑡 = 𝑢𝑢𝑥𝑥𝑥𝑥 , 0 < 𝑥𝑥 < 2𝜋𝜋, 𝑡𝑡 > 0.
9

𝑢𝑢 0, 𝑡𝑡 = 𝑢𝑢 2𝜋𝜋, 𝑡𝑡 = 0; 𝑢𝑢 𝑥𝑥, 0 = 𝑥𝑥(2𝜋𝜋 − 𝑥𝑥)

Next, by using the first three non-zero terms of the series, compute the
temperature of the metal bar at the point 𝑥𝑥 = 𝜋𝜋 when 𝑡𝑡 = 1.
(Give your answer correct to 4 decimal places)
8
Solution:
Let 𝑢𝑢 𝑥𝑥, 𝑡𝑡 = 𝑋𝑋 𝑥𝑥 𝑇𝑇(𝑡𝑡), the heat equation becomes
1
𝑋𝑋𝑇𝑇 ′ = 𝑋𝑋 ′′ 𝑇𝑇 (5.3)
9

Separate the variables, we have

𝑋𝑋𝑋𝑋 9𝑇𝑇𝑇
= = 𝑝𝑝 (5.4)
𝑋𝑋 𝑇𝑇

where constant 𝑝𝑝 is known as separation constant.

From Eqn. (5.4),
𝑋𝑋𝑋𝑋 9𝑇𝑇𝑇
= 𝑝𝑝 and = 𝑝𝑝 (5.5)
𝑋𝑋 𝑇𝑇

Hence 𝑋𝑋 and 𝑇𝑇 must satisfy

1
𝑋𝑋 ′′ − 𝑝𝑝𝑝𝑝 = 0 and 𝑇𝑇 ′ − 𝑝𝑝𝑝𝑝 = 0 (5.6)
9

Since 𝑢𝑢 𝑥𝑥, 𝑡𝑡 = 𝑋𝑋 𝑥𝑥 𝑇𝑇(𝑡𝑡) and 𝑇𝑇 𝑡𝑡 ≠ 0, the boundary conditions become

𝑢𝑢 0, 𝑡𝑡 = 𝑋𝑋 0 𝑇𝑇 𝑡𝑡 = 0 ⇒ 𝑋𝑋 0 = 0 (5.7)
𝑢𝑢 2𝜋𝜋, 𝑡𝑡 = 𝑋𝑋 2𝜋𝜋 𝑇𝑇(𝑡𝑡) = 0 ⇒ 𝑋𝑋 2𝜋𝜋 = 0 9

Now, we need to consider three possible cases for 𝑝𝑝.

Case (1): 𝑝𝑝 = 0.
From Eqn (5.6), (substitute 𝑝𝑝 = 0 into 𝑋𝑋 ′′ − 𝑝𝑝𝑝𝑝 = 0)
𝑋𝑋 ′′ = 0
𝑋𝑋 ′ = 𝐴𝐴
∴ 𝑋𝑋 𝑥𝑥 = 𝐴𝐴𝑥𝑥 + 𝐵𝐵 (5.8)
By substituting boundary conditions (5.7) into Eqn. (5.8),
𝑋𝑋 0 = 0, 𝑋𝑋 0 = 𝐴𝐴 0 + 𝐵𝐵 𝑋𝑋 2𝜋𝜋 = 0, 𝑋𝑋 2𝜋𝜋 = 𝐴𝐴 2𝜋𝜋 + 𝐵𝐵
∴ 𝐵𝐵 = 0 0 = 𝐴𝐴 2𝜋𝜋 + 0
𝐴𝐴 = 0
From Eqn. (5.8), we found that 𝑋𝑋 𝑥𝑥 = 0 and this implies
𝑢𝑢 𝑥𝑥, 𝑡𝑡 = 𝑋𝑋 𝑥𝑥 𝑇𝑇 𝑡𝑡 = 0.
This case gives a trivial solution. Hence, we omit this case.
10
Case (2): 𝑝𝑝 > 0.
Let 𝑝𝑝 = 𝜆𝜆2 and 𝜆𝜆 > 0. Substitute 𝑝𝑝 = 𝜆𝜆2 into Eqn. (5.6) leads to ODE
𝑋𝑋 ′′ − 𝜆𝜆2 𝑋𝑋 = 0
Characteristic equation:
𝑚𝑚2 − 𝜆𝜆2 = 0
𝑚𝑚 = ±𝜆𝜆
∴ 𝑋𝑋 𝑥𝑥 = 𝐴𝐴𝑒𝑒 𝜆𝜆𝜆𝜆 + 𝐵𝐵𝑒𝑒 −𝜆𝜆𝜆𝜆 (5.9)
𝑋𝑋 0 = 0, 𝑋𝑋 0 = 𝐴𝐴𝑒𝑒 𝜆𝜆(0) + 𝐵𝐵𝑒𝑒 −𝜆𝜆(0) 𝑋𝑋 2𝜋𝜋 = 0, 0 = 𝐴𝐴𝑒𝑒 2𝜋𝜋𝜆𝜆 + 𝐵𝐵𝑒𝑒 −2𝜋𝜋𝜆𝜆
∴ 0 = 𝐴𝐴 + 𝐵𝐵 0 = −𝐵𝐵𝑒𝑒 2𝜋𝜋𝜆𝜆 + 𝐵𝐵𝑒𝑒 −2𝜋𝜋𝜆𝜆
𝐴𝐴 = −𝐵𝐵 0 = 𝐵𝐵(−𝑒𝑒 2𝜋𝜋𝜆𝜆 + 𝑒𝑒 −2𝜋𝜋𝜆𝜆 )
Since −𝑒𝑒 2𝜋𝜋𝜆𝜆 + 𝑒𝑒 −2𝜋𝜋𝜆𝜆 ≠ 0, 𝐵𝐵 = 0 and then 𝐴𝐴 = 0. From Eqn. (5.9),
𝑋𝑋 𝑥𝑥 = 0 ⇒ 𝑢𝑢 𝑥𝑥, 𝑡𝑡 = 𝑋𝑋 𝑥𝑥 𝑇𝑇 𝑡𝑡 = 0
This case gives a trivial solution. Hence, we omit this case.
11
Case (3): 𝑝𝑝 < 0.
Let 𝑝𝑝 = −𝜆𝜆2 and 𝜆𝜆 > 0. Substitute 𝑝𝑝 = −𝜆𝜆2 into Eqn (5.6) leads to ODE
1
𝑋𝑋 ′′ + 𝜆𝜆2 𝑋𝑋 = 0and 𝑇𝑇 ′ + 𝜆𝜆2 𝑇𝑇 = 0
9

Characteristic equation:
𝑚𝑚2 + 𝜆𝜆2 = 0 ⇒ 𝑚𝑚 = ±𝜆𝜆𝑖𝑖
∴ 𝑋𝑋 𝑥𝑥 = 𝐴𝐴 cos 𝜆𝜆𝑥𝑥 + 𝐵𝐵 sin 𝜆𝜆𝜆𝜆 (5.10)
𝑋𝑋 0 = 0, 𝑋𝑋 0 = 𝐴𝐴 cos 0 + 𝐵𝐵 sin 0, 𝑋𝑋 2𝜋𝜋 = 0, 𝑋𝑋 2𝜋𝜋 = 𝐴𝐴 cos 2𝜋𝜋𝜆𝜆 + 𝐵𝐵 sin 2𝜋𝜋𝜆𝜆
∴ 0 = 𝐴𝐴 0 = 𝐵𝐵 sin 2𝜋𝜋𝜆𝜆
𝐵𝐵 ≠ 0, ∴ sin 2𝜋𝜋𝜆𝜆 = 0
Given sin 𝑛𝑛𝑛𝑛 = 0, 𝑛𝑛 = 1,2,3, … 2𝜋𝜋𝜆𝜆 = 𝑛𝑛𝜋𝜋, 𝑛𝑛 = 1,2,3, …
So, sin 2𝜋𝜋𝜆𝜆 = 0 implies 2𝜋𝜋𝜆𝜆 = 𝑛𝑛𝜋𝜋
𝑛𝑛
𝜆𝜆𝑛𝑛 = , 𝑛𝑛 = 1,2,3, … (5.11)
2

∴ 𝑋𝑋𝑛𝑛 𝑥𝑥 = 𝐵𝐵𝑛𝑛 sin 𝜆𝜆𝑛𝑛 𝑥𝑥

𝑛𝑛𝑥𝑥
= 𝐵𝐵𝑛𝑛 sin , 𝑛𝑛 = 1,2,3, … (5.12)
12
2
 1 2
𝑇𝑇 ′ + 𝜆𝜆 𝑇𝑇 = 0
9
𝑑𝑑𝑑𝑑 1
= − 𝜆𝜆2 𝑇𝑇
𝑑𝑑𝑑𝑑 9
1 1
∫ 𝑇𝑇 𝑑𝑑𝑇𝑇 = ∫ − 9 𝜆𝜆2 𝑑𝑑𝑑𝑑
1 2
ln 𝑇𝑇 = − 𝜆𝜆 𝑡𝑡 + 𝑘𝑘
9
From Eqn. (5.11), 1
− 𝜆𝜆2 𝑡𝑡+𝑘𝑘
𝑛𝑛 𝑇𝑇 = 𝑒𝑒 9
𝜆𝜆𝑛𝑛 = Let 𝐶𝐶 = 𝑒𝑒 𝑘𝑘
2 1
−9𝜆𝜆2 𝑡𝑡
= 𝐶𝐶𝑒𝑒
1 𝑛𝑛2
− 𝜆𝜆2𝑛𝑛 𝑡𝑡 − 36𝑡𝑡
𝑇𝑇𝑛𝑛 𝑡𝑡 = 𝐶𝐶𝑛𝑛 𝑒𝑒 9 = 𝐶𝐶𝑛𝑛 𝑒𝑒 , 𝑛𝑛 = 1,2,3, … (5.13)
Hence,
𝑢𝑢𝑛𝑛 𝑥𝑥, 𝑡𝑡 = 𝑋𝑋𝑛𝑛 𝑥𝑥 𝑇𝑇𝑛𝑛 𝑡𝑡
𝑛𝑛2
𝑛𝑛𝑥𝑥 − 36 𝑡𝑡
= 𝐵𝐵𝑛𝑛 sin 𝐶𝐶𝑛𝑛 𝑒𝑒
2
Let 𝐷𝐷𝑛𝑛 = 𝐵𝐵𝑛𝑛 𝐶𝐶𝑛𝑛 13
𝑛𝑛2
𝑛𝑛𝑥𝑥 − 𝑡𝑡
= 𝐷𝐷𝑛𝑛 (sin )𝑒𝑒 36 , 𝑛𝑛 = 1,2,3, … (5.14)
2
 If 𝑢𝑢1 , 𝑢𝑢2 ,…, 𝑢𝑢∞ are linearly independent
solutions to a PDE, then the linear combination
of all 𝑢𝑢1 , 𝑢𝑢2 ,…, 𝑢𝑢∞ are also a solution to the PDE
By applying superposition principle on Eqn. (5.14),
∞
𝑛𝑛𝑥𝑥 − 𝑛𝑛2𝑡𝑡
𝑢𝑢 𝑥𝑥, 𝑡𝑡 = � 𝐷𝐷𝑛𝑛 (sin )𝑒𝑒 36 (5.15)
2
𝑛𝑛=1

Apply initial condition, 𝑢𝑢 𝑥𝑥, 0 = 𝑥𝑥(2𝜋𝜋 − 𝑥𝑥), on Eqn. (5.15),

∞
𝑛𝑛𝑛𝑛
𝑥𝑥(2𝜋𝜋 − 𝑥𝑥) = � 𝐷𝐷𝑛𝑛 sin (5.16)
2
𝑛𝑛=1

Recall from Fourier Sine Series:

∞
𝑛𝑛𝜋𝜋𝑥𝑥
𝑓𝑓(𝑥𝑥) = � 𝐵𝐵𝑛𝑛 sin
𝐿𝐿
𝑛𝑛=1
where
2 𝐿𝐿 𝑛𝑛𝜋𝜋𝑥𝑥
𝐵𝐵𝑛𝑛 = � 𝑓𝑓 𝑥𝑥 sin 𝑑𝑑𝑑𝑑 14
𝐿𝐿 0 𝐿𝐿
 From Eqn. (5.16), by half-range expansion for 𝑓𝑓(𝑥𝑥) on 0 < 𝑥𝑥 < 2𝜋𝜋,
2 𝐿𝐿 𝑛𝑛𝜋𝜋𝑥𝑥
𝐷𝐷𝑛𝑛 = � 𝑓𝑓 𝑥𝑥 sin 𝑑𝑑𝑑𝑑
𝐿𝐿 0 𝐿𝐿 Tabular Method:
2 2𝜋𝜋 𝑛𝑛𝑛𝑛 Diff. Integrate
= � 𝑥𝑥(2𝜋𝜋 − 𝑥𝑥) sin 𝑑𝑑𝑑𝑑
2𝜋𝜋 0 2 + 2𝜋𝜋𝑥𝑥 − 𝑥𝑥 2 𝑛𝑛𝑛𝑛
sin
1 2 2𝜋𝜋𝑥𝑥 − 𝑥𝑥 2 𝑛𝑛𝑛𝑛 2
= �− cos − 2𝜋𝜋 − 2𝑥𝑥 2 𝑛𝑛𝑛𝑛
𝜋𝜋 𝑛𝑛 2 − cos
2𝜋𝜋 𝑛𝑛 2
4 2𝜋𝜋 − 2𝑥𝑥 𝑛𝑛𝑛𝑛 16 𝑛𝑛𝑛𝑛
+ sin − cos � + −2 4 𝑛𝑛𝑛𝑛
𝑛𝑛2 2 𝑛𝑛3 2 0 − 2 sin
𝑛𝑛 2
1 16 16
= − 3 cos 𝑛𝑛𝜋𝜋 − − 3 − 0 8 𝑛𝑛𝑛𝑛
𝜋𝜋 𝑛𝑛 𝑛𝑛 cos
−1 𝑛𝑛 𝑛𝑛3 2
16
= 3 1 − cos 𝑛𝑛𝜋𝜋
𝑛𝑛 𝜋𝜋
16
= 3 1 − −1 𝑛𝑛
𝑛𝑛 𝜋𝜋 15
By substituting 𝐷𝐷𝑛𝑛 into Eqn. (5.15), the temperature function is given by
∞
𝑛𝑛𝑛𝑛 − 𝑛𝑛2𝑡𝑡
𝑢𝑢 𝑥𝑥, 𝑡𝑡 = � 𝐷𝐷𝑛𝑛 (sin )𝑒𝑒 36
2
𝑛𝑛=1
∞
16 𝑛𝑛
𝑛𝑛𝑛𝑛 − 𝑛𝑛2𝑡𝑡
= � 3 1 − −1 (sin )𝑒𝑒 36
𝑛𝑛 𝜋𝜋 2
𝑛𝑛=1
∞
32 1 𝑛𝑛𝑛𝑛 − 𝑛𝑛2𝑡𝑡
= � 3 (sin )𝑒𝑒 36
𝜋𝜋 𝑛𝑛 2
𝑛𝑛 𝑜𝑜𝑜𝑜𝑜𝑜

Finally,
∞
32 1 𝑛𝑛𝜋𝜋 − 𝑛𝑛2(1)
𝑢𝑢 𝜋𝜋, 1 = � 3 (sin )𝑒𝑒 36
𝜋𝜋 𝑛𝑛 2
𝑛𝑛 𝑜𝑜𝑜𝑜𝑜𝑜
32 1 𝜋𝜋 − 1 2 1 3𝜋𝜋 − 3 2 1 5𝜋𝜋 − 5 2
= sin 𝑒𝑒 36 + 3 sin 𝑒𝑒 36 + 3 sin 𝑒𝑒 36 +⋯
𝜋𝜋 13 2 3 2 5 2
32
≈ (0.9726 − 0.0288 + 0.004 … )
𝜋𝜋 16

≈ 9.6542℃
Example 5.14:

Solve the following heat equation by using method of separation of

variables:
𝑢𝑢𝑡𝑡 = 4𝑢𝑢𝑥𝑥𝑥𝑥 , 0 < 𝑥𝑥 < 𝜋𝜋, 𝑡𝑡 > 0.
𝑢𝑢 0, 𝑡𝑡 = 𝑢𝑢 𝜋𝜋, 𝑡𝑡 = 0; 𝑢𝑢 𝑥𝑥, 0 = 2 sin 3𝑥𝑥

17
Solution:
Let 𝑢𝑢 𝑥𝑥, 𝑡𝑡 = 𝑋𝑋 𝑥𝑥 𝑇𝑇(𝑡𝑡), the heat equation becomes
𝑋𝑋𝑇𝑇 ′ = 4𝑋𝑋 ′′ 𝑇𝑇 (5.17)
Separate the variables, we have
𝑋𝑋𝑋𝑋 𝑇𝑇𝑇
= 4𝑇𝑇 = 𝑝𝑝 (5.18)
𝑋𝑋

where constant 𝑝𝑝 is known as separation constant.

From Eqn. (5.18),
𝑋𝑋𝑋𝑋 𝑇𝑇𝑇
= 𝑝𝑝 and = 𝑝𝑝 (5.19)
𝑋𝑋 4𝑇𝑇

Hence 𝑋𝑋 and 𝑇𝑇 must satisfy

𝑋𝑋 ′′ − 𝑝𝑝𝑝𝑝 = 0 and 𝑇𝑇 ′ − 4𝑝𝑝𝑝𝑝 = 0 (5.20)
Since 𝑢𝑢 𝑥𝑥, 𝑡𝑡 = 𝑋𝑋 𝑥𝑥 𝑇𝑇(𝑡𝑡) and 𝑇𝑇 𝑡𝑡 ≠ 0, the boundary conditions become
𝑢𝑢 0, 𝑡𝑡 = 𝑋𝑋 0 𝑇𝑇 𝑡𝑡 = 0 ⇒ 𝑋𝑋 0 = 0
(5.21)
𝑢𝑢 𝜋𝜋, 𝑡𝑡 = 𝑋𝑋 𝜋𝜋 𝑇𝑇(𝑡𝑡) = 0 ⇒ 𝑋𝑋 𝜋𝜋 = 0
18

Now, we need to consider three possible cases for 𝑝𝑝.

Case (1): 𝑝𝑝 = 0.
From Eqn (5.20), (substitute 𝑝𝑝 = 0 into 𝑋𝑋 ′′ − 𝑝𝑝𝑝𝑝 = 0)
𝑋𝑋 ′′ = 0
𝑋𝑋 ′ = 𝐴𝐴
∴ 𝑋𝑋 𝑥𝑥 = 𝐴𝐴𝑥𝑥 + 𝐵𝐵 (5.22)
By substituting boundary conditions from Eqn. (5.21) into Eqn. (5.22),
𝑋𝑋 0 = 0, 𝑋𝑋 0 = 𝐴𝐴 0 + 𝐵𝐵 𝑋𝑋 𝜋𝜋 = 0, 𝑋𝑋 𝜋𝜋 = 𝐴𝐴 𝜋𝜋 + 𝐵𝐵
∴ 𝐵𝐵 = 0 0 = 𝐴𝐴 𝜋𝜋 + 0
𝐴𝐴 = 0
From Eqn. (5.22), we found that 𝑋𝑋 𝑥𝑥 = 0 and this implies
𝑢𝑢 𝑥𝑥, 𝑡𝑡 = 𝑋𝑋 𝑥𝑥 𝑇𝑇 𝑡𝑡 = 0.
This case gives a trivial solution. Hence, we omit this case.
19
Case (2): 𝑝𝑝 > 0.
Let 𝑝𝑝 = 𝜆𝜆2 and 𝜆𝜆 > 0. Substitute 𝑝𝑝 = 𝜆𝜆2 into Eqn (5.20) leads to ODE
𝑋𝑋 ′′ − 𝜆𝜆2 𝑋𝑋 = 0
Characteristic equation:
𝑚𝑚2 − 𝜆𝜆2 = 0
𝑚𝑚 = ±𝜆𝜆
∴ 𝑋𝑋 𝑥𝑥 = 𝐴𝐴𝑒𝑒 𝜆𝜆𝜆𝜆 + 𝐵𝐵𝑒𝑒 −𝜆𝜆𝜆𝜆 (5.23)
𝑋𝑋 0 = 0, 𝑋𝑋 0 = 𝐴𝐴𝑒𝑒 𝜆𝜆(0) + 𝐵𝐵𝑒𝑒 −𝜆𝜆(0) 𝑋𝑋 𝜋𝜋 = 0, 0 = 𝐴𝐴𝑒𝑒 𝜋𝜋𝜆𝜆 + 𝐵𝐵𝑒𝑒 −𝜋𝜋𝜆𝜆
∴ 0 = 𝐴𝐴 + 𝐵𝐵 0 = −𝐵𝐵𝑒𝑒 𝜋𝜋𝜆𝜆 + 𝐵𝐵𝑒𝑒 −𝜋𝜋𝜆𝜆
𝐴𝐴 = −𝐵𝐵 0 = 𝐵𝐵(−𝑒𝑒 𝜋𝜋𝜆𝜆 + 𝑒𝑒 −𝜋𝜋𝜆𝜆 )
Since −𝑒𝑒 𝜋𝜋𝜆𝜆 + 𝑒𝑒 −𝜋𝜋𝜆𝜆 ≠ 0, 𝐵𝐵 = 0 and then 𝐴𝐴 = 0. From Eqn. (5.23),
𝑋𝑋 𝑥𝑥 = 0 ⇒ 𝑢𝑢 𝑥𝑥, 𝑡𝑡 = 𝑋𝑋 𝑥𝑥 𝑇𝑇 𝑡𝑡 = 0
This case gives a trivial solution. Hence, we omit this case.
20
Case (3): 𝑝𝑝 < 0.
Let 𝑝𝑝 = −𝜆𝜆2 and 𝜆𝜆 > 0. Substitute 𝑝𝑝 = −𝜆𝜆2 into Eqn (5.20) leads to ODE
𝑋𝑋 ′′ + 𝜆𝜆2 𝑋𝑋 = 0and 𝑇𝑇 ′ + 4𝜆𝜆2 𝑇𝑇 = 0
Characteristic equation:
𝑚𝑚2 + 𝜆𝜆2 = 0 ⇒ 𝑚𝑚 = ±𝜆𝜆𝑖𝑖
∴ 𝑋𝑋 𝑥𝑥 = 𝐴𝐴 cos 𝜆𝜆𝑥𝑥 + 𝐵𝐵 sin 𝜆𝜆𝜆𝜆 (5.24)
𝑋𝑋 0 = 0, 𝑋𝑋 0 = 𝐴𝐴 cos 0 + 𝐵𝐵 sin 0, 𝑋𝑋 𝜋𝜋 = 0, 𝑋𝑋 𝜋𝜋 = 𝐴𝐴 cos 𝜋𝜋𝜆𝜆 + 𝐵𝐵 sin 𝜋𝜋𝜆𝜆
∴ 0 = 𝐴𝐴 0 = 𝐵𝐵 sin 𝜋𝜋𝜆𝜆
𝐵𝐵 ≠ 0, ∴ sin 𝜋𝜋𝜆𝜆 = 0
Given sin 𝑛𝑛𝑛𝑛 = 0, 𝑛𝑛 = 1,2,3, … 𝜋𝜋𝜆𝜆 = 𝑛𝑛𝜋𝜋, 𝑛𝑛 = 1,2,3, …
So, sin 𝜋𝜋𝜆𝜆 = 0 implies 𝜋𝜋𝜆𝜆 = 𝑛𝑛𝜋𝜋
𝜆𝜆𝑛𝑛 = 𝑛𝑛, 𝑛𝑛 = 1,2,3, … (5.25)
∴ 𝑋𝑋𝑛𝑛 𝑥𝑥 = 𝐵𝐵𝑛𝑛 sin 𝜆𝜆𝑛𝑛 𝑥𝑥
= 𝐵𝐵𝑛𝑛 sin 𝑛𝑛𝑥𝑥 , 𝑛𝑛 = 1,2,3, … (5.26)
21
 𝑇𝑇 ′ + 4𝜆𝜆2 𝑇𝑇 = 0
𝑑𝑑𝑑𝑑
= −4𝜆𝜆2 𝑇𝑇
𝑑𝑑𝑑𝑑
1
� 𝑑𝑑𝑇𝑇 = � −4𝜆𝜆2 𝑑𝑑𝑑𝑑
𝑇𝑇

ln 𝑇𝑇 = −4𝜆𝜆2 𝑡𝑡 + 𝑘𝑘
From Eqn. (5.25),
2 𝑡𝑡+𝑘𝑘
𝜆𝜆𝑛𝑛 = 𝑛𝑛 𝑇𝑇 = 𝑒𝑒 −4𝜆𝜆 Let 𝐶𝐶 = 𝑒𝑒 𝑘𝑘
2 𝑡𝑡
= 𝐶𝐶𝑒𝑒 −4𝜆𝜆
2 2
𝑇𝑇𝑛𝑛 𝑡𝑡 = 𝐶𝐶𝑛𝑛 𝑒𝑒 −4𝜆𝜆𝑛𝑛𝑡𝑡 = 𝐶𝐶𝑛𝑛 𝑒𝑒 −4𝑛𝑛 𝑡𝑡 , 𝑛𝑛 = 1,2,3, … (5.27)
Hence,
𝑢𝑢𝑛𝑛 𝑥𝑥, 𝑡𝑡 = 𝑋𝑋𝑛𝑛 𝑥𝑥 𝑇𝑇𝑛𝑛 𝑡𝑡
2 𝑡𝑡
= 𝐵𝐵𝑛𝑛 sin 𝑛𝑛𝑛𝑛 𝐶𝐶𝑛𝑛 𝑒𝑒 −4𝑛𝑛
2 Let 𝐷𝐷𝑛𝑛 = 𝐵𝐵𝑛𝑛 𝐶𝐶𝑛𝑛
= 𝐷𝐷𝑛𝑛 (sin 𝑛𝑛𝑛𝑛)𝑒𝑒 − 4𝑛𝑛 𝑡𝑡 , 𝑛𝑛 = 1,2,3, … (5.28)
22
 If 𝑢𝑢1 , 𝑢𝑢2 ,…, 𝑢𝑢∞ are linearly independent
solutions to a PDE, then the linear combination
of all 𝑢𝑢1 , 𝑢𝑢2 ,…, 𝑢𝑢∞ are also a solution to the PDE
By applying superposition principle on Eqn. (5.28),
∞
2 𝑡𝑡
𝑢𝑢 𝑥𝑥, 𝑡𝑡 = � 𝐷𝐷𝑛𝑛 (sin 𝑛𝑛𝑛𝑛)𝑒𝑒 − 4𝑛𝑛 (5.29)
𝑛𝑛=1

Apply initial condition, 𝑢𝑢 𝑥𝑥, 0 = 2 sin 3𝑥𝑥, on Eqn. (5.29),

∞

2 sin 3𝑥𝑥 = � 𝐷𝐷𝑛𝑛 sin 𝑛𝑛𝑛𝑛 (5.30)

𝑛𝑛=1

Recall from Fourier Sine Series:

∞
𝑛𝑛𝜋𝜋𝑥𝑥
𝑓𝑓(𝑥𝑥) = � 𝐵𝐵𝑛𝑛 sin
𝐿𝐿
𝑛𝑛=1
where
2 𝐿𝐿 𝑛𝑛𝜋𝜋𝑥𝑥 23
𝐵𝐵𝑛𝑛 = � 𝑓𝑓 𝑥𝑥 sin 𝑑𝑑𝑑𝑑
𝐿𝐿 0 𝐿𝐿
From Eqn. (5.30), the half-range expansion for 𝑓𝑓(𝑥𝑥) on 0 < 𝑥𝑥 < 𝜋𝜋 is

2 𝐿𝐿 𝑛𝑛𝜋𝜋𝑥𝑥
𝐷𝐷𝑛𝑛 = � 𝑓𝑓 𝑥𝑥 sin 𝑑𝑑𝑑𝑑
𝐿𝐿 0 𝐿𝐿
2 𝜋𝜋 1
= � 2 sin 3𝑥𝑥 sin 𝑛𝑛𝑛𝑛 𝑑𝑑𝑑𝑑 sin 𝐴𝐴 sin 𝐵𝐵 = [cos(𝐴𝐴 − 𝐵𝐵) − cos(𝐴𝐴 + 𝐵𝐵)]
𝜋𝜋 0 2
2 𝜋𝜋
= � cos(3𝑥𝑥 − 𝑛𝑛𝑛𝑛) − cos(3𝑥𝑥 + 𝑛𝑛𝑛𝑛) 𝑑𝑑𝑑𝑑
𝜋𝜋 0
𝜋𝜋
2 sin(3 − 𝑛𝑛)𝑥𝑥 sin(3 + 𝑛𝑛)𝑥𝑥
= − , 𝑛𝑛 ≠ 3
𝜋𝜋 3 − 𝑛𝑛 3 + 𝑛𝑛 0

2 sin(3 − 𝑛𝑛) 𝜋𝜋 sin(3 + 𝑛𝑛)𝜋𝜋

= − − (0 − 0) , 𝑛𝑛 ≠ 3 (5.31)
𝜋𝜋 3 − 𝑛𝑛 3 + 𝑛𝑛
= 0, 𝑛𝑛 ≠ 3
24
From Eqn. (5.31),
2 sin(3 − 𝑛𝑛) 𝜋𝜋 sin(3 + 𝑛𝑛)𝜋𝜋
𝐷𝐷𝑛𝑛 = − , 𝑛𝑛 ≠ 3
𝜋𝜋 3 − 𝑛𝑛 3 + 𝑛𝑛
When 𝑛𝑛 = 3,
2 sin(3𝜋𝜋 − 𝑛𝑛𝜋𝜋) sin 6𝜋𝜋
𝐷𝐷3 = lim − Apply
𝜋𝜋 𝑛𝑛→3 3 − 𝑛𝑛 6
L’Hopital’s Rule
2 −𝜋𝜋 cos 3𝜋𝜋 − 𝑛𝑛𝜋𝜋
= lim −0
𝜋𝜋 𝑛𝑛→3 −1
2
= 𝜋𝜋
𝜋𝜋
𝐷𝐷3 = 2

25
Alternative Way to find 𝐷𝐷𝑛𝑛 for this question:
From Eqn. (5.30),
∞

2 sin 3𝑥𝑥 = � 𝐷𝐷𝑛𝑛 sin 𝑛𝑛𝑛𝑛

𝑛𝑛=1

By expanding the summation,

2 sin 3𝑥𝑥 = 𝐷𝐷1 sin 𝑥𝑥 + 𝐷𝐷2 sin 2𝑥𝑥 + 𝐷𝐷3 sin 3𝑥𝑥 + 𝐷𝐷4 sin 4𝑥𝑥 + ⋯
By comparing coefficients of both sides,
𝐷𝐷3 = 2 and 𝐷𝐷1 = 𝐷𝐷2 = 𝐷𝐷4 = ⋯ = 0
Hence, the only possible 𝐷𝐷𝑛𝑛 for 𝑢𝑢(𝑥𝑥, 𝑡𝑡) is 𝐷𝐷3 when 𝑛𝑛 = 3.
By substituting 𝐷𝐷𝑛𝑛 into Eqn. (5.29), the temperature function is given by
∞
2 𝑡𝑡
𝑢𝑢 𝑥𝑥, 𝑡𝑡 = � 𝐷𝐷𝑛𝑛 (sin 𝑛𝑛𝑛𝑛)𝑒𝑒 − 4𝑛𝑛
𝑛𝑛=1
= 𝐷𝐷3 sin 3𝑥𝑥 𝑒𝑒 − 36𝑡𝑡
= 2𝑒𝑒 −36𝑡𝑡 sin 3𝑥𝑥 26
Exercise 5.8:
1) Solve the following heat equation by method of separation of variables.
5𝑢𝑢𝑡𝑡 = 𝑢𝑢𝑥𝑥𝑥𝑥 , 0 < 𝑥𝑥 < 10, 𝑡𝑡 > 0.
𝑢𝑢 0, 𝑡𝑡 = 𝑢𝑢 10, 𝑡𝑡 = 0; 𝑢𝑢 𝑥𝑥, 0 = sin 𝜋𝜋𝑥𝑥
𝜋𝜋2 𝑡𝑡
[Ans: (sin 𝜋𝜋𝜋𝜋)𝑒𝑒 − 5 ]

2) Suppose that a rod of length 30 (cm) with 𝑘𝑘 = 0.09cm2 /s is heated and

its initial temperature is 𝑢𝑢 𝑥𝑥, 0 = 30. At time 𝑡𝑡 = 0, its lateral surface is
insulated and its two ends are imbedded in ice at 0℃ (𝑢𝑢 0, 𝑡𝑡 = 𝑢𝑢 𝐿𝐿, 𝑡𝑡 =
0). By method of separation of variables, calculate the rod’s temperature
at its midpoint after 20 minutes for the following case.
[Ans: 𝑢𝑢 15,1200 = 11.6859℃]

27
5.3.2 The Wave Equation
The wave equation usually describes mechanical waves, such as water
wave, sound wave and seismic waves. Its application can be found in
electromagnetics, fluid dynamics and acoustics.
Consider a string, stretched between two fixed points, is being initially
plucked to the form of 𝑓𝑓(𝑥𝑥) to start vibrate and the string moves in a
direction perpendicular to the 𝑥𝑥-axis as shown in Figure 5.2. The vertical
displacement of a point at 𝑥𝑥 at time 𝑡𝑡 is represented by 𝑢𝑢(𝑥𝑥, 𝑡𝑡).
𝑢𝑢

𝑢𝑢 𝑥𝑥, 0 = 𝑓𝑓(𝑥𝑥)

𝑃𝑃
𝑢𝑢(𝑥𝑥, 𝑡𝑡)
𝑥𝑥
0 𝑥𝑥 𝐿𝐿
28

Figure 5.2: Displacement of vibrating string 𝑢𝑢(𝑥𝑥, 𝑡𝑡) at time 𝑡𝑡

Theorem: The vibrating string with an initial velocity
The partial differential equation
𝑢𝑢𝑡𝑡𝑡𝑡 = 𝑐𝑐 2 𝑢𝑢𝑥𝑥𝑥𝑥 , 0 < 𝑥𝑥 < 𝐿𝐿, 𝑡𝑡 > 0
where 𝑐𝑐 is the physical constant (a ratio of string’s tension to its density) with
boundary conditions
𝑢𝑢 0, 𝑡𝑡 = 𝑢𝑢 𝐿𝐿, 𝑡𝑡 = 0, 𝑡𝑡 > 0
and initial conditions
𝑢𝑢 𝑥𝑥, 0 = 𝑓𝑓(𝑥𝑥), 𝑢𝑢𝑡𝑡 𝑥𝑥, 0 = 𝑔𝑔(𝑥𝑥), 0 < 𝑥𝑥 < 𝐿𝐿
has the series solution
∞
𝑛𝑛𝑛𝑛𝑐𝑐𝑐𝑐 𝑛𝑛𝑛𝑛𝑐𝑐𝑐𝑐 𝑛𝑛𝑛𝑛𝑛𝑛
𝑢𝑢 𝑥𝑥, 𝑡𝑡 = � 𝑎𝑎𝑛𝑛 cos + 𝑏𝑏𝑛𝑛 sin sin (5.32)
𝐿𝐿 𝐿𝐿 𝐿𝐿
𝑛𝑛=1

where 𝑎𝑎𝑛𝑛 and 𝑏𝑏𝑛𝑛 are the Fourier series coefficients obtained by
2 𝐿𝐿 𝑛𝑛𝑛𝑛𝑛𝑛
𝑎𝑎𝑛𝑛 = � 𝑓𝑓(𝑥𝑥) sin 𝑑𝑑𝑑𝑑, 𝑛𝑛 = 1,2,3, …
𝐿𝐿 0 𝐿𝐿 (5.33)
29
𝐿𝐿
2 𝑛𝑛𝑛𝑛𝑛𝑛
𝑏𝑏𝑛𝑛 = � 𝑔𝑔(𝑥𝑥) sin 𝑑𝑑𝑑𝑑, 𝑛𝑛 = 1,2,3, …
𝑛𝑛𝜋𝜋𝜋𝜋 0 𝐿𝐿
Example 5.15:
Solve the following wave equation by using the method of separation of
variables.

𝑢𝑢𝑡𝑡𝑡𝑡 = 4𝑢𝑢𝑥𝑥𝑥𝑥 , 0 < 𝑥𝑥 < 2𝜋𝜋, 𝑡𝑡 > 0.

𝑢𝑢 0, 𝑡𝑡 = 𝑢𝑢 2𝜋𝜋, 𝑡𝑡 = 0; 𝑢𝑢 𝑥𝑥, 0 = 3 sin 𝑥𝑥 , 𝑢𝑢𝑡𝑡 𝑥𝑥, 0 = 1.

Next, by using the first three non-zero terms of the series, find the
displacement of spring at the point 𝑥𝑥 = 𝜋𝜋 at 𝑡𝑡 = 10.

30
Solution:
Let 𝑢𝑢 𝑥𝑥, 𝑡𝑡 = 𝑋𝑋 𝑥𝑥 𝑇𝑇(𝑡𝑡), the wave equation becomes
𝑋𝑋𝑇𝑇 ′′ = 4𝑋𝑋 ′′ 𝑇𝑇 (5.34)
Separate the variables, we have
𝑋𝑋𝑋𝑋 𝑇𝑇′′
= = 𝑝𝑝 (5.35)
𝑋𝑋 4𝑇𝑇

where constant 𝑝𝑝 is known as separation constant.

From Eqn. (5.35),
𝑋𝑋𝑋𝑋 𝑇𝑇′′
= 𝑝𝑝 and = 𝑝𝑝 (5.36)
𝑋𝑋 4𝑇𝑇

Hence 𝑋𝑋 and 𝑇𝑇 must satisfy

𝑋𝑋 ′′ − 𝑝𝑝𝑝𝑝 = 0 and 𝑇𝑇 ′′ − 4𝑝𝑝𝑝𝑝 = 0 (5.37)
Since 𝑢𝑢 𝑥𝑥, 𝑡𝑡 = 𝑋𝑋 𝑥𝑥 𝑇𝑇(𝑡𝑡) and 𝑇𝑇 𝑡𝑡 ≠ 0, the boundary conditions become
𝑢𝑢 0, 𝑡𝑡 = 𝑋𝑋 0 𝑇𝑇 𝑡𝑡 = 0 ⇒ 𝑋𝑋 0 = 0
(5.38)
𝑢𝑢 2𝜋𝜋, 𝑡𝑡 = 𝑋𝑋 2𝜋𝜋 𝑇𝑇(𝑡𝑡) = 0 ⇒ 𝑋𝑋 2𝜋𝜋 = 0
31

Now, we need to consider three possible cases for 𝑝𝑝.

Case (1): 𝑝𝑝 = 0.
From Eqn (5.37), (substitute 𝑝𝑝 = 0 into 𝑋𝑋 ′′ − 𝑝𝑝𝑝𝑝 = 0)
𝑋𝑋 ′′ = 0
𝑋𝑋 ′ = 𝐴𝐴
∴ 𝑋𝑋 𝑥𝑥 = 𝐴𝐴𝑥𝑥 + 𝐵𝐵 (5.39)
By substituting boundary conditions (5.38) into Eqn. (5.39),
𝑋𝑋 0 = 0, 𝑋𝑋 0 = 𝐴𝐴 0 + 𝐵𝐵 𝑋𝑋 2𝜋𝜋 = 0, 𝑋𝑋 2𝜋𝜋 = 𝐴𝐴 2𝜋𝜋 + 𝐵𝐵
∴ 𝐵𝐵 = 0 0 = 𝐴𝐴 2𝜋𝜋 + 0
𝐴𝐴 = 0
From Eqn. (5.39), we found that 𝑋𝑋 𝑥𝑥 = 0 and this implies
𝑢𝑢 𝑥𝑥, 𝑡𝑡 = 𝑋𝑋 𝑥𝑥 𝑇𝑇 𝑡𝑡 = 0.
This case gives a trivial solution. Hence, we omit this case.
32
Case (2): 𝑝𝑝 > 0.
Let 𝑝𝑝 = 𝜆𝜆2 and 𝜆𝜆 > 0. Substitute 𝑝𝑝 = 𝜆𝜆2 into Eqn (5.37) leads to ODE
𝑋𝑋 ′′ − 𝜆𝜆2 𝑋𝑋 = 0
Characteristic equation:
𝑚𝑚2 − 𝜆𝜆2 = 0
𝑚𝑚 = ±𝜆𝜆
∴ 𝑋𝑋 𝑥𝑥 = 𝐴𝐴𝑒𝑒 𝜆𝜆𝜆𝜆 + 𝐵𝐵𝑒𝑒 −𝜆𝜆𝜆𝜆 (5.40)
𝑋𝑋 0 = 0, 𝑋𝑋 0 = 𝐴𝐴𝑒𝑒 𝜆𝜆(0) + 𝐵𝐵𝑒𝑒 −𝜆𝜆(0) 𝑋𝑋 2𝜋𝜋 = 0, 0 = 𝐴𝐴𝑒𝑒 2𝜋𝜋𝜆𝜆 + 𝐵𝐵𝑒𝑒 −2𝜋𝜋𝜆𝜆
∴ 0 = 𝐴𝐴 + 𝐵𝐵 0 = −𝐵𝐵𝑒𝑒 2𝜋𝜋𝜆𝜆 + 𝐵𝐵𝑒𝑒 −2𝜋𝜋𝜆𝜆
𝐴𝐴 = −𝐵𝐵 0 = 𝐵𝐵(−𝑒𝑒 2𝜋𝜋𝜆𝜆 + 𝑒𝑒 −2𝜋𝜋𝜆𝜆 )
Since −𝑒𝑒 2𝜋𝜋𝜆𝜆 + 𝑒𝑒 −2𝜋𝜋𝜆𝜆 ≠ 0, 𝐵𝐵 = 0 and then 𝐴𝐴 = 0. From Eqn. (5.40),
𝑋𝑋 𝑥𝑥 = 0 ⇒ 𝑢𝑢 𝑥𝑥, 𝑡𝑡 = 𝑋𝑋 𝑥𝑥 𝑇𝑇 𝑡𝑡 = 0
This case gives a trivial solution. Hence, we omit this case.
33
Case (3): 𝑝𝑝 < 0.
Let 𝑝𝑝 = −𝜆𝜆2 and 𝜆𝜆 > 0. From Eqn (5.37), it leads to ODE
𝑋𝑋 ′′ + 𝜆𝜆2 𝑋𝑋 = 0 and 𝑇𝑇 ′′ + 4𝜆𝜆2 𝑇𝑇 = 0
Characteristic equation for 𝑋𝑋 ′′ + 𝜆𝜆2 𝑋𝑋 = 0:
𝑚𝑚2 + 𝜆𝜆2 = 0 ⇒ 𝑚𝑚 = ±𝜆𝜆𝑖𝑖
∴ 𝑋𝑋 𝑥𝑥 = 𝐴𝐴 cos 𝜆𝜆𝑥𝑥 + 𝐵𝐵 sin 𝜆𝜆𝜆𝜆 (5.41)
𝑋𝑋 0 = 0, 𝑋𝑋 0 = 𝐴𝐴 cos 0 + 𝐵𝐵 sin 0, 𝑋𝑋 2𝜋𝜋 = 0, 𝑋𝑋 2𝜋𝜋 = 𝐴𝐴 cos 2𝜋𝜋𝜆𝜆 + 𝐵𝐵 sin 2𝜋𝜋𝜆𝜆
∴ 0 = 𝐴𝐴 0 = 𝐵𝐵 sin 2𝜋𝜋𝜆𝜆
𝐵𝐵 ≠ 0, ∴ sin 2𝜋𝜋𝜆𝜆 = 0
Given sin 𝑛𝑛𝑛𝑛 = 0, 𝑛𝑛 = 1,2,3, …
2𝜋𝜋𝜆𝜆 = 𝑛𝑛𝜋𝜋, 𝑛𝑛 = 1,2,3, …
So, sin 2𝜋𝜋𝜆𝜆 = 0 implies 2𝜋𝜋𝜆𝜆 = 𝑛𝑛𝜋𝜋
𝑛𝑛
𝜆𝜆𝑛𝑛 = , 𝑛𝑛 = 1,2,3, … (5.42)
2

∴ 𝑋𝑋𝑛𝑛 𝑥𝑥 = 𝐵𝐵𝑛𝑛 sin 𝜆𝜆𝑛𝑛 𝑥𝑥

𝑛𝑛𝑥𝑥
= 𝐵𝐵𝑛𝑛 sin , 𝑛𝑛 = 1,2,3, … (5.43)
34
2
Characteristic equation for 𝑇𝑇 ′′ + 4𝜆𝜆2 𝑇𝑇 = 0:
𝑚𝑚2 + 4𝜆𝜆2 = 0
𝑚𝑚 = ±2𝜆𝜆𝑖𝑖
∴ 𝑇𝑇 𝑡𝑡 = 𝐶𝐶 cos 2𝜆𝜆𝑡𝑡 + 𝐷𝐷 sin 2𝜆𝜆𝑡𝑡
𝑛𝑛
Since 𝜆𝜆𝑛𝑛 = from Eqn. (5.42),
2
∴ 𝑇𝑇𝑛𝑛 𝑡𝑡 = 𝐶𝐶𝑛𝑛 cos 𝑛𝑛𝑡𝑡 + 𝐷𝐷𝑛𝑛 sin 𝑛𝑛𝑡𝑡 , 𝑛𝑛 = 1,2,3, … (5.44)
Hence,
𝑢𝑢𝑛𝑛 𝑥𝑥, 𝑡𝑡 = 𝑋𝑋𝑛𝑛 𝑥𝑥 𝑇𝑇𝑛𝑛 𝑡𝑡
𝑛𝑛𝑛𝑛
= 𝐵𝐵𝑛𝑛 sin 𝐶𝐶𝑛𝑛 cos 𝑛𝑛𝑡𝑡 + 𝐷𝐷𝑛𝑛 sin 𝑛𝑛𝑡𝑡
2 Let 𝐸𝐸𝑛𝑛 = 𝐵𝐵𝑛𝑛 𝐶𝐶𝑛𝑛
𝑛𝑛𝑛𝑛 and 𝐹𝐹𝑛𝑛 = 𝐵𝐵𝑛𝑛 𝐷𝐷𝑛𝑛
= 𝐸𝐸𝑛𝑛 cos 𝑛𝑛𝑡𝑡 + 𝐹𝐹𝑛𝑛 sin 𝑛𝑛𝑡𝑡 sin , 𝑛𝑛 = 1,2,3, …
2 (5.45)
By superposition principle:
∞
𝑛𝑛𝑛𝑛
𝑢𝑢 𝑥𝑥, 𝑡𝑡 = � 𝐸𝐸𝑛𝑛 cos 𝑛𝑛𝑡𝑡 + 𝐹𝐹𝑛𝑛 sin 𝑛𝑛𝑡𝑡 sin
2 (5.46)
𝑛𝑛=1 35
Apply initial condition 𝑢𝑢 𝑥𝑥, 0 = 3 sin 𝑥𝑥 to the Eqn. (5.46)
∞
𝑛𝑛𝑛𝑛
3 sin 𝑥𝑥 = � 𝐸𝐸𝑛𝑛 sin (5.47)
2
𝑛𝑛=1

By applying Fourier Sine series to Eqn. (5.47):

2 𝐿𝐿 𝑛𝑛𝜋𝜋𝑥𝑥
𝐸𝐸𝑛𝑛 = � 𝑓𝑓 𝑥𝑥 sin 𝑑𝑑𝑑𝑑
𝐿𝐿 0 𝐿𝐿
2 2𝜋𝜋 𝑛𝑛𝑛𝑛 1
= � 3 sin 𝑥𝑥 sin 𝑑𝑑𝑑𝑑 sin 𝐴𝐴 sin 𝐵𝐵 = [cos(𝐴𝐴 − 𝐵𝐵) − cos(𝐴𝐴 + 𝐵𝐵)]
2𝜋𝜋 0 2 2
3 2𝜋𝜋 2 − 𝑛𝑛 2 + 𝑛𝑛
= � cos 𝑥𝑥 − cos 𝑥𝑥 𝑑𝑑𝑑𝑑
2𝜋𝜋 0 2 2
2𝜋𝜋
3 2 2 − 𝑛𝑛 2 2 + 𝑛𝑛
= sin 𝑥𝑥 − sin 𝑥𝑥 , 𝑛𝑛 ≠ 2
2𝜋𝜋 2 − 𝑛𝑛 2 2 + 𝑛𝑛 2 0
3 2 2
= sin 2 − 𝑛𝑛 𝜋𝜋 − sin 2 + 𝑛𝑛 𝜋𝜋 − 0 , 𝑛𝑛 ≠ 2 (5.48)
2𝜋𝜋 2−𝑛𝑛 2+𝑛𝑛 36

= 0, 𝑛𝑛 ≠ 2
From Eqn. (5.48),
3 2 2
𝐸𝐸𝑛𝑛 = sin 2 − 𝑛𝑛 𝜋𝜋 − sin 2 + 𝑛𝑛 𝜋𝜋 , 𝑛𝑛 ≠ 2
2𝜋𝜋 2 − 𝑛𝑛 2 + 𝑛𝑛
When 𝑛𝑛 = 2,
3 sin 2 − 𝑛𝑛 𝜋𝜋 sin 4𝜋𝜋
𝐸𝐸2 = lim − Apply
𝜋𝜋 𝑛𝑛→2 2 − 𝑛𝑛 4
L’Hopital’s Rule
3 −𝜋𝜋 cos 2 − 𝑛𝑛 𝜋𝜋
= lim −0
𝜋𝜋 𝑛𝑛→2 −1
3
= 𝜋𝜋
𝜋𝜋
𝐸𝐸2 = 3

37
Alternative Way to find 𝐸𝐸𝑛𝑛 for this question:
From Eqn. (5.47),
∞
𝑛𝑛𝑛𝑛
3 sin 𝑥𝑥 = � 𝐸𝐸𝑛𝑛 sin
2
𝑛𝑛=1

By expanding the summation,

𝑥𝑥 3𝑥𝑥
3 sin 𝑥𝑥 = 𝐸𝐸1 sin + 𝐸𝐸2 sin 𝑥𝑥 + 𝐸𝐸3 sin + 𝐸𝐸4 sin 2𝑥𝑥 + ⋯
2 2
By comparing coefficients of both sides,
𝐸𝐸2 = 3 and 𝐸𝐸1 = 𝐸𝐸3 = 𝐸𝐸4 = ⋯ = 0

Hence, the only possible 𝐸𝐸𝑛𝑛 for 𝑢𝑢(𝑥𝑥, 𝑡𝑡) is 𝐸𝐸2 when 𝑛𝑛 = 2.
38
Differentiate Eqn. (5.46) with respect to 𝑡𝑡 ,
∞
𝑛𝑛𝑛𝑛
𝑢𝑢𝑡𝑡 𝑥𝑥, 𝑡𝑡 = � −𝑛𝑛𝐸𝐸𝑛𝑛 sin 𝑛𝑛𝑡𝑡 + 𝑛𝑛𝐹𝐹𝑛𝑛 cos 𝑛𝑛𝑡𝑡 sin (5.49)
2
𝑛𝑛=1

Apply initial velocity, 𝑢𝑢𝑡𝑡 𝑥𝑥, 0 = 1 into Eqn. (5.49):

∞
𝑛𝑛𝑛𝑛
1 = � 𝑛𝑛𝐹𝐹𝑛𝑛 sin
2
𝑛𝑛=1

By applying Fourier sine series:

2 𝐿𝐿 𝑛𝑛𝜋𝜋𝑥𝑥 2 2𝜋𝜋 𝑛𝑛𝑛𝑛
𝑛𝑛𝐹𝐹𝑛𝑛 = � 𝑓𝑓 𝑥𝑥 sin 𝑑𝑑𝑑𝑑 = � sin 𝑑𝑑𝑑𝑑
𝐿𝐿 0 𝐿𝐿 2𝜋𝜋 0 2
𝑛𝑛𝑛𝑛 2𝜋𝜋
1 2 cos
= − 2
𝜋𝜋 𝑛𝑛 0
2 2 𝑛𝑛
= − cos 𝑛𝑛𝑛𝑛 + 1 = 1 − −1
𝑛𝑛𝑛𝑛 𝑛𝑛𝑛𝑛
2 𝑛𝑛 39
∴ 𝐹𝐹𝑛𝑛 = 2 1 − −1 , 𝑛𝑛 = 1,2,3, … (5.50)
𝑛𝑛 𝜋𝜋
Finally, substitute 𝐸𝐸𝑛𝑛 and 𝐹𝐹𝑛𝑛 into Eqn. (5.46),
∞
𝑛𝑛𝑛𝑛
𝑢𝑢 𝑥𝑥, 𝑡𝑡 = � 𝐸𝐸𝑛𝑛 cos 𝑛𝑛𝑡𝑡 + 𝐹𝐹𝑛𝑛 sin 𝑛𝑛𝑡𝑡 sin
2
𝑛𝑛=1
∞ ∞
𝑛𝑛𝑛𝑛 𝑛𝑛𝑛𝑛
= � 𝐸𝐸𝑛𝑛 cos 𝑛𝑛𝑡𝑡 sin + � 𝐹𝐹𝑛𝑛 sin 𝑛𝑛𝑡𝑡 sin
2 2
𝑛𝑛=1 𝑛𝑛=1
∞
2 𝑛𝑛
𝑛𝑛𝑛𝑛
= 𝐸𝐸2 cos 2𝑡𝑡 sin 𝑥𝑥 + � 1 − −1 sin 𝑛𝑛𝑡𝑡 sin
𝑛𝑛2 𝜋𝜋 2
𝑛𝑛=1
∞
4 1 𝑛𝑛𝑛𝑛
= 3 cos 2𝑡𝑡 sin 𝑥𝑥 + � 2 sin 𝑛𝑛𝑡𝑡 sin
𝜋𝜋 𝑛𝑛 2
𝑛𝑛 𝑜𝑜𝑜𝑜𝑜𝑜
Next,
∞
4 1 𝑛𝑛𝑛𝑛
𝑢𝑢 𝜋𝜋, 10 = 3 cos 20 sin 𝜋𝜋 + � 2 sin 10𝑛𝑛 sin
𝜋𝜋 𝑛𝑛 2
𝑛𝑛 𝑜𝑜𝑜𝑜𝑜𝑜
4 1 𝜋𝜋 1 3𝜋𝜋 1 5𝜋𝜋
=0+ sin 10 sin + sin 30 sin + sin 50 sin
𝜋𝜋 12 2 32 2 52 2 40

≈ −0.5663
Example 5.16:
A string is stretched between two points and plucked at 𝑡𝑡 = 0 as follows:
𝑢𝑢
𝑢𝑢 𝑥𝑥, 0 = 𝑓𝑓(𝑥𝑥)
0.01
𝑥𝑥
0 2 4

Length in meter
If the string is being released initially with zero velocity, apply the wave
equation
𝜕𝜕 2 𝑢𝑢 𝜕𝜕 2 𝑢𝑢
= 100 2
𝜕𝜕𝑡𝑡 2 𝜕𝜕𝑥𝑥
to determine the subsequent motion of a point P with distance 𝑥𝑥 from the
origin at time 𝑡𝑡.
41
Solution:
From the figure, we need to find the 𝑓𝑓(𝑥𝑥) for the intervals of 0 ≤ 𝑥𝑥 ≤ 2 and
2 ≤ 𝑥𝑥 ≤ 4.
For the interval 0 ≤ 𝑥𝑥 ≤ 2, the line between (0,0) and (2,0.01) is
𝑦𝑦1 − 𝑦𝑦0
𝑦𝑦 − 𝑦𝑦0 = (𝑥𝑥 − 𝑥𝑥0 )
𝑥𝑥1 − 𝑥𝑥0
0.01 − 0
𝑦𝑦 − 0 = (𝑥𝑥 − 0)
2−0
𝑦𝑦 = 0.005𝑥𝑥
For the interval 2 ≤ 𝑥𝑥 ≤ 4, the line between (2,0.01) and (4,0) is
0 − 0.01
𝑦𝑦 − 0 = (𝑥𝑥 − 4)
4−2
𝑦𝑦 = 0.02 − 0.005𝑥𝑥
Hence,
0.005𝑥𝑥, 0 ≤ 𝑥𝑥 ≤ 2
𝑓𝑓 𝑥𝑥 = �
0.02 − 0.005𝑥𝑥, 2 ≤ 𝑥𝑥 ≤ 4 42
The string is being released initially with zero velocity,
𝑢𝑢𝑡𝑡 𝑥𝑥, 0 = 0
Hence, the model of the wave equation is as follow:
𝑢𝑢𝑡𝑡𝑡𝑡 = 100𝑢𝑢𝑥𝑥𝑥𝑥 , 0 < 𝑥𝑥 < 4, 𝑡𝑡 > 0.
𝑢𝑢 0, 𝑡𝑡 = 𝑢𝑢 4, 𝑡𝑡 = 0, 𝑢𝑢𝑡𝑡 𝑥𝑥, 0 = 0,
0.005𝑥𝑥, 0 ≤ 𝑥𝑥 ≤ 2
𝑢𝑢 𝑥𝑥, 0 = 𝑓𝑓 𝑥𝑥 = � . (5.51)
0.02 − 0.005𝑥𝑥, 2 ≤ 𝑥𝑥 ≤ 4

Apply the method of separation of variables:

Let 𝑢𝑢 𝑥𝑥, 𝑡𝑡 = 𝑋𝑋 𝑥𝑥 𝑇𝑇(𝑡𝑡), the wave equation becomes
𝑋𝑋𝑇𝑇 ′′ = 100𝑋𝑋 ′′ 𝑇𝑇 (5.52)
Separate the variables, we have
𝑋𝑋𝑋𝑋 𝑇𝑇′′
= = 𝑝𝑝 43
𝑋𝑋 100𝑇𝑇
Hence 𝑋𝑋 and 𝑇𝑇 must satisfy
𝑋𝑋 ′′ − 𝑝𝑝𝑝𝑝 = 0 and 𝑇𝑇 ′′ − 100𝑝𝑝𝑝𝑝 = 0 (5.53)

Since 𝑢𝑢 𝑥𝑥, 𝑡𝑡 = 𝑋𝑋 𝑥𝑥 𝑇𝑇(𝑡𝑡) and 𝑇𝑇 𝑡𝑡 ≠ 0, the boundary conditions become

𝑢𝑢 0, 𝑡𝑡 = 𝑋𝑋 0 𝑇𝑇 𝑡𝑡 = 0 ⇒ 𝑋𝑋 0 = 0
𝑢𝑢 4, 𝑡𝑡 = 𝑋𝑋 4 𝑇𝑇 𝑡𝑡 = 0 ⇒ 𝑋𝑋 4 = 0

Now, we need to consider three possible cases for 𝑝𝑝.

The first two cases are similar to the previous example where they
produce trivial solution. I leave it to you to work out the first two cases.
We will move to the third case directly.
44
Case (3): 𝑝𝑝 < 0.
Let 𝑝𝑝 = −𝜆𝜆2 and 𝜆𝜆 > 0. Substitute 𝑝𝑝 = −𝜆𝜆2 into Eqn (5.53) leads to ODE
𝑋𝑋 ′′ + 𝜆𝜆2 𝑋𝑋 = 0 and 𝑇𝑇 ′′ + 100𝜆𝜆2 𝑇𝑇 = 0
Characteristic equation for 𝑋𝑋 ′′ + 𝜆𝜆2 𝑋𝑋 = 0:
𝑚𝑚2 + 𝜆𝜆2 = 0 ⇒ 𝑚𝑚 = ±𝜆𝜆𝑖𝑖
∴ 𝑋𝑋 𝑥𝑥 = 𝐴𝐴 cos 𝜆𝜆𝑥𝑥 + 𝐵𝐵 sin 𝜆𝜆𝜆𝜆 (5.54)
𝑋𝑋 0 = 0, 𝑋𝑋 0 = 𝐴𝐴 cos 0 + 𝐵𝐵 sin 0, 𝑋𝑋 4 = 0, 𝑋𝑋 4 = 𝐴𝐴 cos 4𝜆𝜆 + 𝐵𝐵 sin 4𝜆𝜆
∴ 0 = 𝐴𝐴 0 = 𝐵𝐵 sin 4𝜆𝜆
𝐵𝐵 ≠ 0, ∴ sin 4𝜆𝜆 = 0
Given sin 𝑛𝑛𝑛𝑛 = 0, 𝑛𝑛 = 1,2,3, …
4𝜆𝜆 = 𝑛𝑛𝜋𝜋, 𝑛𝑛 = 1,2,3, …
So, sin 4𝜆𝜆 = 0 implies 4𝜆𝜆 = 𝑛𝑛𝜋𝜋
𝑛𝑛𝜋𝜋
𝜆𝜆𝑛𝑛 = , 𝑛𝑛 = 1,2,3, … (5.55)
4

∴ 𝑋𝑋𝑛𝑛 𝑥𝑥 = 𝐵𝐵𝑛𝑛 sin 𝜆𝜆𝑛𝑛 𝑥𝑥

𝑛𝑛𝜋𝜋𝑥𝑥
= 𝐵𝐵𝑛𝑛 sin , 𝑛𝑛 = 1,2,3, … (5.56)
4 45
Characteristic equation for 𝑇𝑇 ′′ + 100𝜆𝜆2 𝑇𝑇 = 0:
𝑚𝑚2 + 100𝜆𝜆2 = 0
𝑚𝑚 = ±10𝜆𝜆𝑖𝑖
∴ 𝑇𝑇 𝑡𝑡 = 𝐶𝐶 cos 10𝜆𝜆𝑡𝑡 + 𝐷𝐷 sin 10𝜆𝜆𝑡𝑡
𝑛𝑛𝜋𝜋
Since 𝜆𝜆𝑛𝑛 = from Eqn. (5.55),
4
5𝑛𝑛𝜋𝜋𝑡𝑡 5𝑛𝑛𝜋𝜋𝜋𝜋
∴ 𝑇𝑇𝑛𝑛 𝑡𝑡 = 𝐶𝐶𝑛𝑛 cos + 𝐷𝐷𝑛𝑛 sin , 𝑛𝑛 = 1,2,3, … (5.57)
2 2
Hence,
𝑢𝑢𝑛𝑛 𝑥𝑥, 𝑡𝑡 = 𝑋𝑋𝑛𝑛 𝑥𝑥 𝑇𝑇𝑛𝑛 𝑡𝑡
𝑛𝑛𝜋𝜋𝑥𝑥 5𝑛𝑛𝜋𝜋𝜋𝜋 5𝑛𝑛𝜋𝜋𝜋𝜋
= 𝐵𝐵𝑛𝑛 sin 𝐶𝐶𝑛𝑛 cos + 𝐷𝐷𝑛𝑛 sin
4 2 2 Let 𝐸𝐸𝑛𝑛 = 𝐵𝐵𝑛𝑛 𝐶𝐶𝑛𝑛
5𝑛𝑛𝜋𝜋𝜋𝜋 5𝑛𝑛𝜋𝜋𝜋𝜋 𝑛𝑛𝜋𝜋𝑥𝑥 and 𝐹𝐹𝑛𝑛 = 𝐵𝐵𝑛𝑛 𝐷𝐷𝑛𝑛
= 𝐸𝐸𝑛𝑛 cos + 𝐹𝐹𝑛𝑛 sin sin
2 2 4 (5.58)
46
By superposition principle:
∞
5𝑛𝑛𝜋𝜋𝜋𝜋 5𝑛𝑛𝜋𝜋𝜋𝜋 𝑛𝑛𝜋𝜋𝑥𝑥
𝑢𝑢 𝑥𝑥, 𝑡𝑡 = � 𝐸𝐸𝑛𝑛 cos + 𝐹𝐹𝑛𝑛 sin sin (5.59)
2 2 4
𝑛𝑛=1

0.005𝑥𝑥, 0 ≤ 𝑥𝑥 ≤ 2
Apply initial condition 𝑢𝑢 𝑥𝑥, 0 = 𝑓𝑓 𝑥𝑥 = � to the Eqn.
0.02 − 0.005𝑥𝑥, 2 ≤ 𝑥𝑥 ≤ 4
(5.59),
∞
𝑛𝑛𝜋𝜋𝑥𝑥
𝑓𝑓(𝑥𝑥) = � 𝐸𝐸𝑛𝑛 sin (5.60)
4
𝑛𝑛=1

47
 By applying Fourier Sine series to Eqn. (5.60):
2 𝐿𝐿 𝑛𝑛𝜋𝜋𝑥𝑥
𝐸𝐸𝑛𝑛 = � 𝑓𝑓 𝑥𝑥 sin 𝑑𝑑𝑑𝑑
𝐿𝐿 0 𝐿𝐿
2 2 𝑛𝑛𝜋𝜋𝑥𝑥 4
𝑛𝑛𝜋𝜋𝑥𝑥
= � 0.005𝑥𝑥 sin 𝑑𝑑𝑑𝑑 + � (0.02 − 0.005𝑥𝑥) sin 𝑑𝑑𝑑𝑑
4 0 4 2 4

Tabular Method: Tabular Method:

Diff. Integrate Diff. Integrate
+ 0.005𝑥𝑥 𝑛𝑛𝜋𝜋𝑥𝑥 + 0.02 − 0.005𝑥𝑥 𝑛𝑛𝜋𝜋𝑥𝑥
sin sin
4 4
− 0.005 4 𝑛𝑛𝜋𝜋𝑥𝑥 − −0.005 4 𝑛𝑛𝜋𝜋𝑥𝑥
− cos − cos
𝑛𝑛𝜋𝜋 4 𝑛𝑛𝜋𝜋 4
+ 0 16 𝑛𝑛𝜋𝜋𝑥𝑥 + 0 16 𝑛𝑛𝜋𝜋𝑥𝑥
− 2 2 sin − 2 2 sin
𝑛𝑛 𝜋𝜋 4 𝑛𝑛 𝜋𝜋 4
2 4
1 0.02𝑥𝑥 𝑛𝑛𝜋𝜋𝑥𝑥 0.08 𝑛𝑛𝜋𝜋𝑥𝑥 4 0.005𝑥𝑥 − 0.02 𝑛𝑛𝜋𝜋𝑥𝑥 0.08 𝑛𝑛𝜋𝜋𝑥𝑥
𝐸𝐸𝑛𝑛 = − cos + 2 2 sin � + cos − 2 2 sin �
2 𝑛𝑛𝜋𝜋 4 𝑛𝑛 𝜋𝜋 4 0 𝑛𝑛𝜋𝜋 4 𝑛𝑛 𝜋𝜋 4 2
48
 2 4
1 0.02𝑥𝑥 𝑛𝑛𝜋𝜋𝑥𝑥 0.08 𝑛𝑛𝜋𝜋𝑥𝑥 4 0.005𝑥𝑥 − 0.02 𝑛𝑛𝜋𝜋𝑥𝑥 0.08 𝑛𝑛𝜋𝜋𝑥𝑥
𝐸𝐸𝑛𝑛 = − cos + 2 2 sin � + cos − 2 2 sin �
2 𝑛𝑛𝜋𝜋 4 𝑛𝑛 𝜋𝜋 4 0 𝑛𝑛𝜋𝜋 4 𝑛𝑛 𝜋𝜋 4 2

1 0.04 𝑛𝑛𝜋𝜋 0.08 𝑛𝑛𝜋𝜋 0.04 𝑛𝑛𝜋𝜋 0.08 𝑛𝑛𝜋𝜋

= − cos + 2 2 sin + 0− − cos − 2 2 sin
2 𝑛𝑛𝜋𝜋 2 𝑛𝑛 𝜋𝜋 2 𝑛𝑛𝜋𝜋 2 𝑛𝑛 𝜋𝜋 2

1 0.16 𝑛𝑛𝜋𝜋
= sin
2 𝑛𝑛2 𝜋𝜋 2 2
0.08 𝑛𝑛𝜋𝜋
= 2 2 sin
𝑛𝑛 𝜋𝜋 2

49
Differentiate Eqn. (5.59) with respect to 𝑡𝑡 ,
∞
5𝑛𝑛𝜋𝜋 5𝑛𝑛𝜋𝜋𝜋𝜋 5𝑛𝑛𝜋𝜋 5𝑛𝑛𝜋𝜋𝜋𝜋 𝑛𝑛𝜋𝜋𝑥𝑥
𝑢𝑢𝑡𝑡 𝑥𝑥, 𝑡𝑡 = � − 𝐸𝐸𝑛𝑛 sin + 𝐹𝐹𝑛𝑛 cos sin (5.61)
2 2 2 2 4
𝑛𝑛=1

Apply initial velocity, 𝑢𝑢𝑡𝑡 𝑥𝑥, 0 = 0 into Eqn. (5.61):

∞
5𝑛𝑛𝜋𝜋 𝑛𝑛𝜋𝜋𝑥𝑥
0=� 𝐹𝐹𝑛𝑛 sin
2 4
𝑛𝑛=1

By applying Fourier sine series:

5𝑛𝑛𝜋𝜋 2 𝐿𝐿 𝑛𝑛𝜋𝜋𝑥𝑥
𝐹𝐹𝑛𝑛 = � 𝑓𝑓 𝑥𝑥 sin 𝑑𝑑𝑑𝑑
2 𝐿𝐿 0 𝐿𝐿
2 4 𝑛𝑛𝜋𝜋𝑥𝑥
= � (0) sin 𝑑𝑑𝑑𝑑
4 0 4
=0
50
∴ 𝐹𝐹𝑛𝑛 = 0
Finally, substitute 𝐸𝐸𝑛𝑛 and 𝐹𝐹𝑛𝑛 into Eqn. (5.59),
∞
5𝑛𝑛𝜋𝜋𝜋𝜋 5𝑛𝑛𝜋𝜋𝜋𝜋 𝑛𝑛𝜋𝜋𝑥𝑥
𝑢𝑢 𝑥𝑥, 𝑡𝑡 = � 𝐸𝐸𝑛𝑛 cos + 𝐹𝐹𝑛𝑛 sin sin
2 2 4
𝑛𝑛=1
∞
0.08 𝑛𝑛𝜋𝜋 5𝑛𝑛𝜋𝜋𝜋𝜋 𝑛𝑛𝜋𝜋𝑥𝑥
=� sin cos sin
𝑛𝑛2 𝜋𝜋 2 2 2 4
𝑛𝑛=1

51
Exercise 5.9:
1) Solve the following wave equation by using method of separation of
variables:
𝑢𝑢𝑡𝑡𝑡𝑡 = 𝑢𝑢𝑥𝑥𝑥𝑥 , 0 < 𝑥𝑥 < 𝜋𝜋, 𝑡𝑡 > 0.
𝑢𝑢 0, 𝑡𝑡 = 𝑢𝑢 𝜋𝜋, 𝑡𝑡 = 0; 𝑢𝑢 𝑥𝑥, 0 = 4 sin 𝑥𝑥 − 3 sin 2𝑥𝑥 ; 𝑢𝑢𝑡𝑡 𝑥𝑥, 0 = 0.5.
2 1
[Ans: 𝑢𝑢 𝑥𝑥, 𝑡𝑡 = 4 cos 𝑡𝑡 sin 𝑥𝑥 − 3 cos 2𝑡𝑡 sin 2𝑥𝑥 + ∑∞ sin 𝑛𝑛𝑛𝑛 sin 𝑛𝑛𝑛𝑛]
𝜋𝜋 𝑛𝑛 𝑜𝑜𝑜𝑜𝑜𝑜 𝑛𝑛2

2) A stretched string of length 20cm is set oscillating by displacing its

midpoint a distance 1cm from its rest position and releasing it with
zero initial velocity. Solve the wave equation 𝑢𝑢𝑡𝑡𝑡𝑡 = 𝑢𝑢𝑥𝑥𝑥𝑥 to determine
the resulting motion, 𝑢𝑢 𝑥𝑥, 𝑡𝑡 .
8 1 𝑛𝑛𝜋𝜋 𝑛𝑛𝜋𝜋𝑡𝑡 𝑛𝑛𝜋𝜋𝑥𝑥
[Ans: 𝑢𝑢 𝑥𝑥, 𝑡𝑡 = 𝜋𝜋2 ∑∞
𝑛𝑛=1 (sin ) cos sin ]
𝑛𝑛2 2 20 20 52
References

1. Abd Wahid M. R. and Mohd Nor M. (2018). Differential equations

For Engineering Students. Penerbit UTM.

2. Muzalna M. J., Irmawani J., Rahifa R., Nurilyana A. A. (2018).

Module 2: Differential Equations, Penerbit UTeM.

3. Edwards C. H., Penny D.E.& Calvis D.(2016). Differential

Equations and Boundary Value Problems, Fifth Edition. Pearson
Education Inc.

53
 Thank You

Questions & Answer?

54