Partial Diff Notes
Partial Diff Notes
Chapter 2
𝑓( 𝑥, 𝑦, 𝑧, 𝑝, 𝑞, 𝑟, 𝑠, 𝑡) = 0 ...(1)
The most general linear partial differential equation of order two in two independent
variables x and y with variable coefficients is of the form
𝑅𝑟 + 𝑆𝑠 + 𝑇𝑡 + 𝑃𝑝 + 𝑄𝑞 + 𝑍𝑧 = 𝐹 . . . (2)
where 𝑅, 𝑆, 𝑇, 𝑃, 𝑄, 𝑍, 𝐹 are functions of 𝑥 and 𝑦 only and not all 𝑅, 𝑆, 𝑇 are zero.
𝜕2𝑧
Sol. The given equation can be written as 𝜕𝑥 2 = 6𝑥 ...(1)
𝜕𝑧
Integrating (1) w. r. t. 𝑥 𝜕𝑥 = 3𝑥 2 + ∅1 (𝑦) ...(2)
𝑥 𝑧 = 𝑥 3 + 𝑥 ∅1 (𝑦) + ∅2 (𝑦)
Ex.2. 𝑎𝑟 = 𝑥𝑦
𝜕2𝑧 1
Sol: Given equation can be written as 𝜕𝑥 2 = 𝑎 𝑥𝑦 ...(1)
𝜕𝑧 𝑦 𝑥2
= + ∅1(y) ...(2)
𝜕𝑥 𝑎 2
𝑦
or z = 2𝑎 + x ∅1(y) + ∅2(y)
𝜕2𝑧
Sol. Given equation can be written as 𝜕𝑦 2 = sin(𝑥𝑦)...(1)
Exercises:𝑥𝑦𝑠 = 1
𝜕2𝑧
Sol: We know that 𝑠 = 𝜕𝑥𝜕𝑦
𝜕2𝑧
Therefore 𝑥𝑦 𝜕𝑥𝜕𝑦 = 1
𝜕2𝑧 1
or = 𝑥𝑦
𝜕𝑥𝜕𝑦
𝜕𝑧 1
= log 𝑦 + 𝑓 𝑥
𝜕𝑥 𝑥
Again integrating w.r.t., x we get
𝑧 = log 𝑥 log 𝑦 + 𝑓 𝑥 𝑑𝑥 + 𝐹 𝑦
0r 𝑧 = log 𝑥 log 𝑦 + 𝑔 𝑥 + 𝐹 𝑦
Exercises:2𝑥 + 2𝑦 = 𝑠
𝜕2𝑧
= 2𝑥 + 2𝑦
𝜕𝑥𝜕𝑦
𝜕𝑧
= 𝑦 2 + 2𝑥𝑦 + 𝑓 𝑥
𝜕𝑥
Integrating w.r.t., 𝑥, we have
𝑧 = 𝑦2𝑥 + 𝑥2 𝑦 + 𝑓 𝑥 𝑑𝑥 + 𝐹 𝑦
∴ 𝑧 = 𝑦2𝑥 + 𝑥2𝑦 + 𝑔 𝑥 + 𝐹 𝑦
Exercises:𝑥𝑟 + 𝑝 = 9𝑥 2 𝑦 3
𝜕2𝑧
𝑥 + 𝑝 = 9𝑥 2 𝑦 3
𝜕𝑥 2
𝜕𝑝
⇒𝑥 + 𝑝 = 9𝑥 2 𝑦 3
𝜕𝑥
𝜕𝑝 𝑝
⇒ 𝜕𝑥 + 𝑥 = 9𝑥𝑦 3 … (1)
∴ I. F. is 𝑒 log 𝑥 = 𝑥
𝜕𝑝 𝑝
𝑥 + = 9𝑥 2 𝑦 3
𝜕𝑥 𝑥
⇒ 𝑝𝑥 = 9 𝑥 2 𝑦 3 𝑑𝑥
𝑥3𝑦3
⇒ 𝑝𝑥 = 9 +𝑓 𝑦
3
⇒ 𝑝𝑥 = 3𝑥 3 𝑦 3 + 𝑓 𝑦
3𝑥 3 𝑦 3 + 𝑓 𝑦
⇒𝑝=
𝑥
𝜕𝑧 𝑓 𝑦
⇒ = 3𝑥 2 𝑦 3 +
𝜕𝑥 𝑥
Integrating with respect to 𝑥 we get
𝑧 = 𝑥 3 𝑦 3 + 𝑓 𝑦 log 𝑥 + 𝐹 𝑦
Exercises:𝑦𝑡 − 𝑞 = 𝑥𝑦
Exercises:𝑡 − 𝑥𝑞 = 𝑥 2
Exercises:𝑟 = 2𝑦 2
𝜕2𝑧
= 2𝑦 2
𝜕𝑥 2
𝜕𝑝
⇒ = 2𝑥 2
𝜕𝑥
Integrating with respect to 𝑥 we get
𝑝 = 2𝑦 2 𝑥 + 𝑓 𝑦
𝜕𝑧
⇒ = 2𝑦 2 𝑥 + 𝑓 𝑦
𝜕𝑥
Integrating we get
𝑧 = 𝑦2𝑥2 + 𝑓 𝑦 𝑑𝑥 + 𝐹 𝑦
⇒ 𝑧 = 𝑦 2 𝑥 2 + 𝑥𝑓 𝑦 + 𝐹 𝑦
Exercises:𝑡 = sin 𝑥𝑦
Exercises:log 𝑠 = 𝑥 + 𝑦
𝜕𝑞
log =𝑥+𝑦
𝜕𝑥
𝜕𝑞
⇒ = 𝑒 𝑥+𝑦
𝜕𝑥
𝜕𝑞
⇒ = 𝑒 𝑥 𝑒𝑦
𝜕𝑥
Integrating w.r.t. 𝑥 we get
𝑞 = 𝑒 𝑥 𝑒𝑦 + 𝑓 𝑦
𝜕𝑧
⇒ = 𝑒 𝑥 𝑒𝑦 + 𝑓 𝑦
𝜕𝑦
𝑧 = 𝑒 𝑥 𝑒𝑦 + 𝑓 𝑦 𝑑𝑦 + 𝐹 𝑥
or 𝑧 = 𝑒 𝑥 𝑒𝑦 + 𝑔 𝑦 + 𝐹 𝑥
𝑥
Exercises:𝑠 − 𝑡 = 𝑦 2
Exercises:𝑡 + 𝑠 + 𝑞 = 0
𝜕𝑞 𝜕𝑝 𝜕𝑧
+ + =0
𝜕𝑦 𝜕𝑦 𝜕𝑦
𝑞+𝑝+𝑧 =𝑓 𝑥
⇒𝑝+𝑞 = 𝑓 𝑥 −𝑧
It is of the form 𝑃𝑝 + 𝑄𝑞 = 𝑅
𝑑𝑥 = 𝑑𝑦
⇒ 𝑥−𝑦 =𝑎
𝑑𝑥 𝑑𝑧
=
1 𝑓 𝑥 −𝑧
⇒ 𝑓 𝑥 − 𝑧 𝑑𝑥 = 𝑑𝑧
𝑑𝑧
⇒ =𝑓 𝑥 −𝑧
𝑑𝑥
𝑑𝑧
⇒ −𝑧 =𝑓 𝑥
𝑑𝑥
It is first order linear differential equation in 𝑧
𝑑𝑥
Its integrating factor is 𝑒 = 𝑒𝑥
Therefore 𝑧𝑒 𝑥 = 𝑓 𝑥 𝑒 𝑥 𝑑𝑥
⇒ 𝑧𝑒 𝑥 = 𝑓 𝑥 𝑒 𝑥 𝑑𝑥 + 𝑓 𝑦
Exercise:𝑡 + 𝑠 + 𝑞 = 1
𝑦 2 = 4𝑎𝑥, 𝑧=0
𝑥𝑟 + 2𝑝 = 0
𝜕𝑝 2
⇒ 𝜕𝑥 + 𝑥 𝑝 = 0 … (1)
𝑥2 𝑝 = 0 𝑑𝑥
⇒ 𝑥2 𝑝 = 0 𝑑𝑥 + 𝑓(𝑦)
𝑓(𝑦)
⇒𝑝=
𝑥2
Integrating w. r. t. 𝑥 we have
1
𝑧 = −𝑥𝑓 𝑦 +𝐹 𝑦 … (2)
𝑦2
Using the given condition 𝑧 = 0, 𝑥 = 4𝑎 , in equation (2), we have
4𝑎
0=− 𝑓 𝑦 +𝐹 𝑦
𝑦2
4𝑎𝑓 𝑦
or 𝐹 𝑦 = … (3)
𝑦2
−𝑦 2
Also for 𝑧 = 1, and 𝑥 = , we have from (2) we have
4𝑎
4𝑎
1= 𝑓 𝑦 +𝐹 𝑦
𝑦2
8𝑎𝑓 𝑦
⇒1=
𝑦2
𝑦2
⇒𝑓 𝑦 =
8𝑎
Substituting 𝑓 𝑦 , in (3)
4𝑎 𝑦 2
𝐹 𝑦 = 2
𝑦 8𝑎
1
⇒𝐹 𝑦 =
2
−𝑦 2 1
𝑧= +
8𝑎𝑥 2
Which is the required surface passing through the parabolas.
𝑦 = 0 = 𝑧and 𝑦=1=𝑧
𝑡 = 6𝑥 3 𝑦
𝜕𝑞
⇒ = 6𝑥 3 𝑦
𝜕𝑦
6𝑥 3 𝑦 2
𝑞= +𝑓 𝑥
2
𝜕𝑧
⇒ = 3𝑥 3 𝑦 2 + 𝑓 𝑥
𝜕𝑦
Integrating w. r. t., 𝑦,
3𝑥 3 𝑦 3
𝑧= + 𝑦𝑓 𝑥 + 𝐹 𝑥
3
⇒ 𝑧 = 𝑥 3 𝑦 3 + 𝑦𝑓 𝑥 + 𝐹 𝑥 …(1)
0=0+0+𝐹 𝑥
⇒𝐹 𝑥 =0 … (2)
1 = 𝑥3 + 𝑓 𝑥 + 𝐹 𝑥
𝑓 𝑥 = 1 − 𝑥3 …(3)
𝑧 = 𝑥3 𝑦3 + 𝑦 1 − 𝑥3
Exercise: Find the surface satisfying 𝑟 + 𝑠 = 0, and touching the elliptic paraboloid
𝑧 = 4𝑥 2 + 𝑦 2 along the surface of plane 𝑦 = 2𝑥 + 1.
𝜕𝑝 𝜕𝑞
Sol: From the given equation we have 𝜕𝑥 + 𝜕𝑥 = 0.
𝑝+𝑞 =𝑓 𝑦
𝑑𝑥 𝑑𝑦
=
1 1
Integrating we get
𝑥 =𝑦+𝑎
⇒ 𝑥−𝑦 =𝑎 …(2)
𝑑𝑦 𝑑𝑧
=
1 𝑓 𝑦
⇒ 𝑑𝑧 = 𝑓 𝑦 𝑑𝑦
⇒𝑧 =𝜑 𝑦 +𝑏
or 𝑧 =𝜑 𝑦 +𝐹 𝑎
⇒𝑧 =𝜑 𝑦 +𝐹 𝑥−𝑦 … (3)
𝜕𝑧
q= 𝜕𝑦 = 𝜑 ′ 𝑦 − 𝐹 ′ 𝑥 − 𝑦 … (5)
Since 𝑧 = 4𝑥 2 + 𝑦 2
𝜕𝑧
∴ 𝑝 = 𝜕𝑥 = 8𝑥 …(6)
𝜕𝑧
& q= 𝜕𝑦 = 2𝑦 …(7)
𝐹 ′ 𝑥 − 𝑦 = 8𝑥 … (8)
𝜑 ′ 𝑦 − 𝐹 ′ 𝑥 − 𝑦 = 2𝑦 … (9)
𝜑 ′ 𝑦 = 8𝑥 + 2𝑦
8
= 𝑦 − 1 + 2𝑦
2
= 6𝑦 − 4
𝜑 𝑦 = 3𝑦 2 − 4𝑦 + 𝑏 … (10)
−𝐹 ′ 𝑥 − 𝑦 = 8𝑥 = −8 𝑦 − 𝑥 − 1 = 8 𝑥 − 𝑦 + 1
𝑧 = 3𝑦 2 − 4𝑦 + 𝑏 − 4 𝑥 − 𝑦 2
−8 𝑥−𝑦 +𝑐
= −4𝑥 2 − 𝑦 2 + 4𝑦 − 8𝑥 + 8𝑥𝑦 + 𝑑
4𝑥 2 + 2𝑥 + 1 2
= −4𝑥 2 − 2𝑥 + 1 2
+ 4 2𝑥 + 1 − 8𝑥 + 8𝑥 2𝑥 + 1 + 𝑑
⇒ 8𝑥 2 + 2 2𝑥 + 1 2
= 4 2𝑥 + 1 − 8𝑥 + 8𝑥 2𝑥 + 1 + 𝑑
⇒ 8𝑥 2 + 8𝑥 2 + 2 + 8𝑥 = 8𝑥 + 4 − 8𝑥 + 16𝑥 2 + 8𝑥 + 𝑑
⇒ 𝑑 = −2
or 𝜑 𝐷, 𝐷′ 𝑧 = 𝑓 𝑥, 𝑦
or 𝜑 𝐷, 𝐷′ 𝑧 = 0
.. .. .. ...
.. .. .. ...
.. .. .. ...
𝜕𝑛 𝑧
𝐷𝑛 𝑧 = 𝜕𝑥 𝑛 = 𝑚𝑛 𝐹 𝑛 𝑦 + 𝑚𝑥
And
𝜕𝑧
𝐷′ 𝑧 = 𝜕𝑦 = 𝐹 ′ 𝑦 + 𝑚𝑥
𝜕2𝑧
𝐷′2 𝑧 = 𝜕𝑦 2 = 𝐹 ′′ 𝑦 + 𝑚𝑥
𝑧 = 𝜑1 𝑦 + 𝑚1 𝑥 + 𝜑2 𝑦 + 𝑚2 𝑥 + . . . +𝜑𝑛 𝑦 + 𝑚𝑛 𝑥
∴ 𝜑 𝐷, 𝐷′ 𝑧 = 0
where the coefficients A0, A1, . . . An, B0, B1, . . . , Bn-1, M0, M1 and N0 are all
constants, then (1) is called a linear partial differential equation with constant
coefficients.
𝜕 𝜕
For convenience 𝜕𝑥 and 𝜕𝑦 will be denoted by D and 𝐷ʹ respectively.
2.1. Short method for finding the P.I. in certain cases of F(D,Dʹ) z = f(x, y)
1
= 𝑣𝑑𝑣𝑑𝑣, where v= x+y
12 +3.1.1+2.12
v2 1 𝑣3 1
= dv = 6 = 36 (x + y)3
2 6
−1
1 𝐷ʹ2
= 1 − 𝑎2 𝑥
𝐷2 𝐷2
1
= [1 + a2 (Dʹ2 /D2) +...]x
𝐷2
1
= 𝐷2x
x3
= ...(2)
6
Exercise: 1.2𝑟 + 5𝑠 + 2𝑡 = 0
2𝑚2 + 5𝑚 + 2 = 0
1
⇒ 𝑚 = − 2 , -2
1
Therefore the complementary function is 𝑧 = 𝜑1 𝑦 − 2 𝑥 + 𝜑2 𝑦 − 2𝑥
Exercise: 2.𝑟 = 𝑎2 𝑡
or 𝐷3 − 3𝐷2 𝐷′ + 2𝐷𝐷′2 𝑧 = 0
𝑚3 − 3𝑚2 + 2𝑚 = 0
⇒ 𝑚 𝑚2 − 3𝑚 + 2 = 0
⇒𝑚 𝑚−1 𝑚−2 =0
⇒ 𝑚 = 0, 1 , 2
Exercise: 6. 𝐷4 − 𝐷′4 𝑧 = 0
𝑚4 − 1 = 0
⇒ 𝑚2 − 1 𝑚2 + 1 = 0
⇒ 𝑚 = 1, −1, 𝑖, −𝑖
𝑧 = 𝜑1 𝑦 + 𝑥 + 𝜑2 𝑦 − 𝑥 + 𝜑3 𝑦 + 𝑖𝑥 + 𝜑4 𝑦 − 𝑖𝑥
𝑚2 − 2𝑚 + 1 = 0
⇒ 𝑚−1 𝑚−1 = 0
⇒ 𝑚 = 1, 1
𝑧 = 𝑓1 𝑦 + 𝑥 + 𝑥𝑓2 𝑦 + 𝑥
1
Also, P. I. = 12𝑥𝑦
𝐷−𝐷′ 2
−2
12 𝐷′
= 2 1− 𝑥𝑦
𝐷 𝐷
2 −1
12 𝐷 𝐷
= 2 1−2 + 𝑥𝑦
𝐷 𝐷′ 𝐷′
2
12 𝐷 𝐷
= 2 1 − −2 + + 𝑥𝑦
𝐷 𝐷′ 𝐷′
12 2
= 𝑥𝑦 + 𝑥
𝐷2 𝐷
12 𝑥 2 𝑦 𝑥 3
= +
𝐷 2 3
𝑥3 𝑦 𝑥4
= 12 +
6 12
= 2𝑥 3 𝑦+ 𝑥 4
i.e., 𝑧 = 𝑓1 𝑦 + 𝑥 + 𝑥𝑓2 𝑦 + 𝑥 + 2𝑥 3 𝑦 + 𝑥 4
𝑚3 − 1 = 0
⇒ 𝑚 − 1 𝑚2 + 𝑚 + 1 = 0
−1 + 𝑖 3 −1 − 𝑖 3
⇒ 𝑚 = 1, ,
2 2
−1+𝑖 3
Complementary function is 𝑧 = 𝜑1 𝑦 + 𝑥 + 𝜑2 𝑦 + 𝑥 + 𝜑3 𝑦 +
2
−1−𝑖 3
𝑥
2
1
P. I. = 𝑥3 𝑦3
𝐷 3 −𝐷′ 3
1
= 𝐷′ 3
𝑥3𝑦3
𝐷3 1−
𝐷
−1
1 𝐷′ 3
= 1− 𝑥3 𝑦3
𝐷3 𝐷
1 𝐷′ 3 𝐷′ 6
= 1+ + + . . . 𝑥3 𝑦3
𝐷3 𝐷 𝐷
1 𝐷′ 3 𝐷′ 6
= 𝑥3 𝑦3 + 𝑥3 𝑦3 + 𝑥3𝑦3 + . . .
𝐷3 𝐷 𝐷
1 𝐷′ 2
= 𝑥3𝑦3 + 3𝑥 3 𝑦 2 + 0 + . . .
𝐷3 𝐷3
1 𝐷′
= 𝑥 3 𝑦 3 + 𝐷 3 6𝑥 3 𝑦
𝐷3
1 1
= 𝑥 3 𝑦 3 + 𝐷 3 6𝑥 3
𝐷3
1 1 𝑥4
= 𝑥3 𝑦3 + 𝐷2 6
𝐷3 4
1 1 𝑥5
= 𝑥 3 𝑦 3 + 𝐷 6 20
𝐷3
1 𝑥6
= 𝑥 3 𝑦 3 + 20
𝐷3
1 𝑥4𝑦 3 𝑥7
= + 140
𝐷2 4
1 𝑥5𝑦 3 𝑥8
= + 1120
𝐷 20
𝑥6𝑦 3 𝑥9
= + 10080
120
−1 + 𝑖 3 −1 − 𝑖 3 𝑥6 𝑦3 𝑥9
𝑧 = 𝜑1 𝑦 + 𝑥 + 𝜑2 𝑦 + 𝑥 + 𝜑3 𝑦 + 𝑥 + +
2 2 120 10080
Exercise: Find the real function ′𝑣′ of 𝑥 and 𝑦 reducing to zero when 𝑦 = 0 and
𝜕2𝑣 𝜕2𝑣
satisfying + 𝜕𝑦 2 = −4𝜋 𝑥 2 + 𝑦 2
𝜕𝑥 2
1
= 𝐷′ 2
−4𝜋 𝑥 2 + 𝑦 2
𝐷 2 1+ 2
𝐷
−1
−4𝜋 𝐷′ 2
= 1+ 𝑥2 + 𝑦2
𝐷2 𝐷2
2
−4𝜋 𝐷′ 2 𝐷′ 2
= 1− + + . . . 𝑥2 + 𝑦2
𝐷2 𝐷2 𝐷2
−4𝜋 𝐷′ 2
= 𝑥2 + 𝑦2 − 𝑥2 + 𝑦2 + 0
𝐷2 𝐷2
−4𝜋 𝐷′
= 𝑥 2 + 𝑦 2 − 𝐷 2 2𝑦
𝐷2
−4𝜋 1
= 𝑥2 + 𝑦2 − 𝐷2 2
𝐷2
−4𝜋 1
= 𝑥 2 + 𝑦 2 − 𝐷 2𝑥
𝐷2
−4𝜋
= 𝑥2 + 𝑦2 − 𝑥2
𝐷2
−4𝜋
= 𝑥𝑦 2
𝐷
𝑥2𝑦2
= −4𝜋 2
= −2𝜋𝑥 2 𝑦 2
𝐹 𝐷, 𝐷′ 𝑢𝑟 = 0 , 𝑟 = 1, 2, . . . , 𝑛
∴ 𝐹 𝐷, 𝐷′ 𝑐𝑟 𝑢𝑟 = 𝑐𝑟 ′ 𝐹 𝐷, 𝐷′ 𝑢𝑟
and𝐹 𝐷, 𝐷′ 𝑢𝑟 = 𝐹 𝐷, 𝐷′ 𝑢𝑟
= 𝑐𝑟 𝐹 𝐷, 𝐷′ 𝑢𝑟
𝑟=1
= 0
𝑛
Therefore 𝑟=1 𝑐𝑟 𝑢𝑟 acts as a solution for the homogeneous system.
And
𝛾𝑟 𝑥
𝐷′𝑢𝑟 = −𝛼𝑟 𝑒𝑥𝑝. − 𝜑′ 𝛽𝑟 𝑥 − 𝛼𝑟 𝑦
𝛼𝑟
𝛾𝑟 𝑥 ′
∴ 𝛼𝑟 𝐷 + 𝛽𝑟 𝐷′ + 𝛾𝑟 𝑢𝑟 = −𝛾𝑟 𝑢𝑟 + 𝛼𝑟 𝛽𝑟 𝑒𝑥𝑝. − 𝜑 𝛽𝑟 𝑥 − 𝛼𝑟 𝑦
𝛼𝑟
𝛾𝑟 𝑥
−𝛼𝑟 𝛽𝑟 𝑒𝑥𝑝. − 𝜑 ′ 𝛽𝑟 𝑥 − 𝛼𝑟 𝑦 + 𝛾𝑟 𝑢𝑟 = 0 . . . (3)
𝛼𝑟
Since 𝛼𝑟 𝐷 + 𝛽𝑟 𝐷 ′ + 𝛾𝑟 is a factor of 𝐹 𝐷, 𝐷′
𝐹 𝐷, 𝐷′ 𝑢𝑟 = 0
Therefore 𝑢𝑟 is a solution of 𝐹 𝐷, 𝐷′ 𝑧 = 0
Let 𝐹 𝐷, 𝐷′ 𝑧 = 𝑓 𝑥, 𝑦 . . . (1)
𝐹 𝐷, 𝐷′ 𝑧 = 𝛼𝑟 𝐷 + 𝛽𝑟 𝐷′ + 𝛾𝑟 𝑧
𝑟=1
𝑑𝑥 𝑑𝑦 𝑑𝑧
= =
𝛼𝑟 𝛽𝑟 𝛾𝑟 𝑧
𝛽𝑟 𝑥 − 𝛼𝑟 𝑦 = 𝑐𝑟
𝑑𝑧 𝛾𝑟
= − 𝑑𝑥
𝑧 𝛼𝑟
𝛾
Integrating we get log 𝑧 = − 𝛼𝑟 𝑥 + 𝐴𝑟
𝑟
𝛾𝑟
⇒ 𝑧 = log 𝛽𝑟 exp − 𝑥 𝑒 𝐴𝑟 = 𝛽𝑟
𝛼𝑟
𝛾𝑟
⇒ 𝑧 = 𝜑 𝑟 exp − 𝑥
𝛼𝑟
𝛾𝑟
⇒ 𝑧 = 𝜑 𝛽𝑟 𝑥 − 𝛼𝑟 𝑦 exp − 𝑥
𝛼𝑟
Also
𝑑𝑧 𝛾𝑟
= − 𝑑𝑦
𝑧 𝛽𝑟
𝛾𝑟
⇒ 𝑧 = 𝜑 𝛽𝑟 𝑥 exp − 𝑦
𝛼𝑟
𝜕𝑧 𝜕𝑧 𝛾𝑟
⇒ 𝛼𝑟 + 𝛽𝑟 = 𝜑𝑟 𝛽𝑟 𝑥 − 𝛼𝑟 𝑦 exp − 𝑥 − 𝛾𝑟 𝑧
𝜕𝑥 𝜕𝑦 𝛼𝑟
Auxiliary system is
𝑑𝑥 𝑑𝑦 𝑑𝑧
= =
𝛼𝑟 𝛽𝑟 𝜑 𝛽 𝑥 − 𝛼 𝑦 exp − 𝛾𝑟 𝑥 − 𝛾 𝑧
𝑟 𝑟 𝑟 𝛼𝑟 𝑟
𝑑𝑥 𝑑𝑦
=
𝛼𝑟 𝛽𝑟
⇒ 𝛽𝑟 𝑑𝑥 = 𝛼𝑟 𝑑𝑦
⇒ 𝛽𝑟 𝑥 − 𝛼𝑟 𝑑𝑦 = 𝑐𝑟
𝑑𝑥 𝑑𝑧
=
𝛼𝑟 𝜑 𝛽 𝑥 − 𝛼 𝑦 exp − 𝛾𝑟 𝑥 − 𝛾 𝑧
𝑟 𝑟 𝑟 𝛼𝑟 𝑟
𝛾𝑟
𝑑𝑧 𝜑𝑟 𝛽𝑟 𝑥 − 𝛼𝑟 𝑦 exp − 𝛼𝑟 𝑥 − 𝛾𝑟 𝑧
⇒ =
𝑑𝑥 𝛼𝑟
𝛾𝑟
𝑑𝑧 𝛾𝑟 𝑧 𝜑𝑟 𝛽𝑟 𝑥 − 𝛼𝑟 𝑦 exp − 𝛼𝑟 𝑥
⇒ + =
𝑑𝑥 𝛼𝑟 𝛼𝑟
𝛾𝑟𝑥
Here I. F. is 𝑒 𝛼 𝑟 therefore the above equation can be written as
𝛾𝑟 𝑥 𝜑𝑟 𝛽𝑟 𝑥 − 𝛼𝑟 𝑦
𝑑 𝑧𝑒 𝛼 𝑟 =
𝛼𝑟
𝛾𝑟 𝑥 1
⇒ 𝑧𝑒 𝛼 𝑟 = 𝜑𝑟 𝛽𝑟 𝑥 − 𝛼𝑟 𝑦 𝑑𝑥 + 𝛽𝑟
𝛼𝑟
𝛾𝑟 𝑥 𝛾 𝑥
− 𝑟
⇒ 𝑧𝑒 𝛼 𝑟 = 𝑒 𝛼𝑟 𝜑𝑟 𝛽𝑟 𝑥 − 𝛼𝑟 𝑦 + 𝜓𝑟 𝛽𝑟 𝑥 − 𝛼𝑟 𝑦
Example: If 𝑧 = 𝑒 𝑎𝑥 +𝑏𝑦
Then 𝐹 𝐷, 𝐷′ 𝑧 = 𝐹 𝑎, 𝑏 𝑒 𝑎𝑥 +𝑏𝑦
⇒ 𝑚 − 1 𝑚2 − 𝑚 − 2 = 0
⇒ 𝑚 = 1, −1, 2
Therefore the C. F. is
𝑧 = 𝑓1 𝑦 + 𝑥 + 𝑓1 𝑦 − 𝑥 + 𝑓1 𝑦 + 2𝑥
1
P. I. = 𝐷 3 −2𝐷 2 𝐷 ′ −𝐷𝐷′ 2 +2𝐷′ 3
𝑒 𝑥+𝑦
1
= 2 𝑒 𝑥+𝑦
𝐷2 𝐷−2𝐷′ −𝐷 ′ 𝐷−𝐷′
1
= 2 𝑒 𝑥+𝑦
𝐷 2 −𝐷 ′ 𝐷−2𝐷′
1
= 𝑒 𝑥+𝑦
𝐷−𝐷 ′ 𝐷+𝐷′ 𝐷−2𝐷′
1
= 𝑒 𝑥+𝑦
𝐷−𝐷 ′ 1+1 1−2
1
= 𝑒 𝑥+𝑦
−2 𝐷−𝐷 ′
1
Now let 𝑤 = 𝐷−𝐷′ 𝑒 𝑥+𝑦
⇒ 𝐷 − 𝐷′ 𝑤 = 𝑒 𝑥+𝑦
𝑑𝑥 𝑑𝑦 𝑑𝑤
= = 𝑥+𝑦
1 −1 𝑒
From first two members we have
𝑑𝑥 𝑑𝑦
=
1 −1
⇒ 𝑑𝑥 + 𝑑𝑦 = 0
⇒𝑥+𝑦 =𝑐
𝑑𝑥 𝑑𝑤
= 𝑥+𝑦
1 𝑒
𝑑𝑤
⇒ 𝑑𝑥 =
𝑒𝑐
⇒ 𝑑𝑤 = 𝑒 𝑐 𝑑𝑥
⇒ 𝑤 = 𝑒𝑐 𝑥
⇒ 𝑤 = 𝑥𝑒 𝑥+𝑦
𝑤 1
Therefore the particular integral = − 2 = − 2 𝑥𝑒 𝑥+𝑦
1
𝑧 = 𝑓1 𝑦 + 𝑥 + 𝑓1 𝑦 − 𝑥 + 𝑓1 𝑦 + 2𝑥 + − 𝑥𝑒 𝑥+𝑦
2
Laplace Equation
2
𝜕2 𝜕2
∇ 𝑧= + 𝑧=0
𝜕𝑥 2 𝜕𝑦 2
𝜕2𝑧 𝜕2𝑧
⇒ 2 + 2 = e−x cos y
𝜕𝑥 𝜕𝑦
= 𝑎2 + 𝑏 2 𝑒 𝑎𝑥 +𝑏𝑦
Where 𝑎2 + 𝑏 2 = 0 = 𝐹 𝑎, 𝑏
∞
Therefore the complementary function is 𝐶. 𝐹. = 𝑟=0 𝐴𝑟 𝑒 𝑎𝑥 +𝑏𝑦
1
= cos 𝑦 𝐷 2 −1 e−x
1
= 𝑥 cos 𝑦 2𝐷 e−x
𝑥
= − 2 cos 𝑦 e−x
Using 𝑧 → 0 as 𝑥 → ∞, we write
⇒ 𝑏𝑟 = ±𝑖 𝑎𝑟2
= ±𝑖𝜆𝑟
∞
𝑥
𝑧 = 𝐴𝑟 𝑒 −𝜆 𝑟 𝑥 𝑒 ±𝑖𝜆 𝑟 𝑥 − cos 𝑦 e−x
2
𝑟=0
∞
𝑥
= 𝐵𝑟 𝑒 −𝜆 𝑟 𝑥 cos 𝜆𝑟 𝑦 + 𝜖𝑟 − cos 𝑦 e−x
2
𝑟=0
cos 𝑦 = 𝐵𝑟 cos 𝜆𝑟 𝑦 + 𝜖𝑟
𝑟=0
𝜕2𝑦 𝜕𝑦 𝜕2𝑦
Exercise: Show that the equation + 2𝑘 𝜕𝑡 = 𝑐 2 𝜕𝑥 2
𝜕𝑡 2
Similarly we get
⇒ 𝑎2 + 2𝑘𝑎 + 𝑐 2 𝑏 2 𝑒 𝑎𝑡 +𝑏𝑥 = 0
⇒ 𝑎2 + 2𝑘𝑎 + 𝑐 2 𝑏 2 = 0
−2𝑘 ± 4𝑘 2 + 4𝑏 2 𝑐 2
⇒ 𝑎=
2
⇒ 𝑎 = −𝑘 ± 𝑘 2 + 𝑏 2 𝑐 2
In general 𝑎𝑟 = −𝑘 ± 𝑘 2 + 𝑏𝑟2 𝑐 2
If 𝑏𝑟2 = −𝛼𝑟2
Then 𝑎𝑟 = −𝑘 ± 𝑘 2 − 𝛼𝑟2 𝑐 2
= −𝑘 ± 𝑖𝜔𝑟
Therefore 𝑦 = 𝑒 𝑏𝑟 𝑥 𝑒 𝑎 𝑟 𝑡
𝜕 2 𝑧 1 𝜕𝑧 2
𝜕2𝑧
− = 4𝑥
𝜕𝑥 2 𝑥 𝜕𝑥 𝜕𝑦 2
Therefore C. F. is 𝑓1 𝑦 + 𝑥 + 𝑓2 𝑦 − 𝑥
1
P. I. is 2 𝑥−𝑦
𝐷 2 −𝐷 ′
1
= 𝑥−𝑦
𝐷′ 2
𝐷2 1− 2
𝐷
2 −1
1 𝐷
= 2 1− 𝑥−𝑦
𝐷 𝐷′
2 4
1 𝐷 𝐷
= 2 1+ + + . . . 𝑥−𝑦
𝐷 𝐷′ 𝐷′
2 4
1 𝐷 𝐷
= 𝑥−𝑦 + 𝑥−𝑦 + 𝑥−𝑦 + . . .
𝐷2 𝐷′ 𝐷′
1
= 𝑥−𝑦 +0+0 . . .
𝐷2
1 𝑥2
= − 𝑦𝑥
𝐷 2
𝑥 3 𝑦𝑥 2
= −
3 2
𝑥3
Therefore the complete solution is 𝑧 = 𝐶. 𝐹. +𝑃. 𝐼. = 𝑓1 𝑦 + 𝑥 + 𝑓2 𝑦 − 𝑥 + −
3
𝑦𝑥2
2
𝜕 2 𝑧 1 𝜕𝑧
=
𝜕𝑥 2 𝑘 𝜕𝑡
Possesses solution of the form
∞
2𝑡
𝑐𝑛 cos 𝑛𝑥 + 𝜖𝑛 𝑒 −𝑘𝑛
𝑟=0
(a) 𝐷2 − 𝐷′ 𝑧 = 2𝑦 − 𝑥 2
(b) 𝐷2 − 𝐷′ 𝑧 = 𝑒 2𝑥−𝑦
(c) 𝑟 + 𝑠 − 2𝑡 = 𝑒 𝑥+𝑦
(d) 𝑟 − 𝑠 + 2𝑞 − 𝑧 = 𝑥 2 𝑦 2
1
Sol: (a) P. I. is 2𝑦 − 𝑥 2
𝐷 2 −𝐷 ′
1
= 2𝑦 − 𝑥 2
𝐷′
𝐷2 1− 2
𝐷
−1
1 𝐷′
= 2 1− 2 2𝑦 − 𝑥 2
𝐷 𝐷
1 𝐷′ 𝐷′2
= 2 1+ 2+ 4 + . . . 2𝑦 − 𝑥 2
𝐷 𝐷 𝐷
1 2
𝐷′ 𝐷′2
= 2 2𝑦 − 𝑥 + 2 2𝑦 − 𝑥 + 4 2𝑦 − 𝑥 2 + . . .
2
𝐷 𝐷 𝐷
1 1
= 2𝑦 − 𝑥 2 + 2
𝐷2 𝐷2
1 1
= 2𝑦 − 𝑥 2 + 2𝑥
𝐷2 𝐷
1
= 2𝑦 − 𝑥 2 + 𝑥 2
𝐷2
1
= 2𝑦
𝐷2
1
= 2𝑥𝑦
𝐷
= 𝑥2 𝑦
1
= 𝑥2𝑦2
𝐷′ 2𝐷′ 1
𝐷2 1− 𝐷 + 2 − 2
𝐷 𝐷
−1
1 𝐷′ 2𝐷′ 1
= 2 1− − 2 + 2 𝑥2𝑦2
𝐷 𝐷 𝐷 𝐷
2
1 𝐷′ 2𝐷′ 1 𝐷′ 2𝐷′ 1
= 2 1+ − 2 + 2 + − 2 + 2 + . . . 𝑥2𝑦2
𝐷 𝐷 𝐷 𝐷 𝐷 𝐷 𝐷
=
1 1 2 1 𝐷′
𝑥 2 𝑦 2 + 𝐷 2𝑥 2 𝑦 − 𝐷 2 2𝑥 2 𝑦 + 𝐷 2 𝑥 2 𝑦 2 + 𝐷 2 𝑥 2 𝑦 2 +
𝐷2
𝐷′ 2 1 𝐷′ 2 4 2𝐷′
4 𝐷4 𝑥2 𝑦2 + 𝐷4 𝑥2 𝑦2 − 4 𝐷3 𝑥2 𝑦 2 − 𝐷4 𝑥2 𝑦2 + 𝑥2𝑦2 + . . .
𝐷3
1 2 1 1 1 2
= 𝑥 2 𝑦 2 + 3 𝑥 3 𝑦 − 12 𝑥 4 𝑦 + 6 𝑥 4 + 40 𝑥 6 − 15 𝑥 5 −
𝐷2
1 1
𝑥 6 𝑦 + 15 𝑥 5 𝑦
45
Chapter-3
Classification of second order partial differential equation
Definition: A second order partial differential equation which is linear w. r. t., the
second order partial derivatives i.e. 𝑟, 𝑠 and 𝑡 is said to be a quasi linear PDE of
second order. For example the equation
𝑅𝑟 + 𝑆𝑠 + 𝑇𝑡 + 𝑓 𝑥, 𝑦, 𝑧, 𝑝, 𝑞 = 0 … . (1)
where 𝑓(𝑥, 𝑦, 𝑧, 𝑝, 𝑞) need not be linear, is a quasi linear partial differential equation.
Here the coefficients R, S, T may be functions of x and y, however for the sake of
simplicity we assume them to be constants.
There are mainly three types of boundary value problems for Laplace equation. If f
∈ C0, and is prescribed on the boundary C of some finite region R, the problem of
determining a function ∅ (𝑥, 𝑦, 𝑧) such that∇2∅ = 0 within R and satisfying ∅ = f on
C,is called the boundary value problem of first kind or Dirichlet problem. The second
type of boundary value problem (BVP) is to determine the function ∅ (𝑥, 𝑦, 𝑧) so that
𝜕∅ 𝜕∅
∇2∅ = 0 within R while 𝜕𝑛 is sepecified at every point of C, where 𝜕𝑛 is the normal
derivative of ∅. This problem is called the Neumann problem.
The third type of boundary value problem is concerned with the determination of the
function ∅ (𝑥, 𝑦, 𝑧) such that ∇ 2 ∅ = 0 within R,while a boundary condition of the
𝜕∅
form 𝜕𝑛 + h ∅ = f,where h≥ 0 is specified at every point of the boundary C.This is
called a mixed boundary value problem or Churchill’s problem.
𝜕 2 (𝑢) 𝜕 2 (𝑢)
𝛻2 ∅ = + = 0 … (1)
𝜕𝑥 2 𝜕𝑦 2
then solution is
𝑑2𝑋 𝑑2𝑌
Case II: let k = 0 then 𝑑𝑥 2 = 0 and = 0
𝑑𝑦
u(x,y) = (C9 cos px + C10 sin px) (C11 epy + C12 e-py ) ...(5)
Where x ≥ 0 and 0≤ y ≤ a.
C1 = 0 ∀ y
As u(x,0) = 0, we get
⇒ pa = n𝜋,n∈ I ⇒ p = n𝜋 / a, n = 0, ∓ 1,...
𝑛𝜋𝑦
u(x, y) = 𝐴n e-n𝜋x/a sin 𝑎
Ex. Show that the two dimensional Laplace equation ∇12 V = 0, in the plane polar
coordinates r and 𝜃 has the solution of the form (Arn + Brn) e (∓ in𝜃),
2V 1 V 1 2V 2V
V 2 0
2
r r r r 2 2 z 2
V R 2V 2R
Z and Z
r r r 2 r 2
2V 2 2V 2Z
RZ , R
2 2 z 2 z 2
2R 1 R 1 2 2Z
V Z
2
Z RZ R
r 2 r r r 2 2 z 2
1 d 2 R 1 dR 1 d 2 1 d 2 Z
Or 0
R dr 2 r dr r 2 d 2 Z dz 2
1 d 2Z d 2Z
m , m Z 0
2 2
Let 2
then 2
Z dz dz
mz
Z e
1 d 2 d 2 in
n n 0 e
2 2
Now let
d 2
d 2
d 2 R 1 dR 2 2
Now R
m n 0
dr 2
r dr 2
r
which is Bessel’s equation
R(r) = A J m B Y m
n n r n n r
J m B Y m }e e
im imz
Therefore V(r,𝜃,z) = {A n r n n r
n
Home Assignments
2 1 1 2
0
r 2 r r r 2 2
Subject to the condition v 0 at r=a
r
And v r U cos , v 1 U sin as r→ ∞
r
∇2 u = 00 ≤ 𝑟 ≤ 10
0≤𝜃≤𝜋
400
𝑢 10, 𝜃 = 𝜋𝜃 − 𝜃 2
𝜋
𝑢 𝑟, 0 = 0 = 𝑢 𝑟, 𝜋
And 𝑢 0, 𝜃 is finite
Exercise: Show that the solution of the Cauchy problem for the Laplace equation
𝜕2𝑢 𝜕2𝑢
+ =0
𝜕𝑥 2 𝜕𝑥 2
Subject to the condition 𝑢 𝑥, 0 = 0
1
𝑢𝑦 𝑥, 0 = sin 𝑛𝑥
𝑛
Where 𝑛 is a positive integer and
1
𝑢 𝑥, 𝑦 = sinh 𝑛𝑦 sin 𝑛𝑥
𝑛2
Interior Dirichlet problem for a circle
The Dirichlet problem for a circle is defined as follows:
To find the value of u at any point in the interior of the circle r = a in terms of its
values on the boundary such that u is the single valued and continuous function
within and on the circular region and satisfies the equation ∇2u = 0 ; 0≤ 𝑟 ≤ 𝑎 subject
to u(a,𝜃) = f(𝜃) ; 0≤ 𝜃 ≤ 2𝜋
We have
𝜕𝑢
∇2u = 𝜕2u / 𝜕r 2 + 1/r 𝜕𝑟 + 1/r2𝜕2u / 𝜕𝜃 2 = 0
We know that
u(r,𝜃) = ∞
𝑛=0(𝐶 nr
n
+ Dnr-n ) (Anʹ cos n𝜃 + Bnʹ sin n𝜃 )
Since the function u is defined for all the values within and on the boundary of a
circle
∞ n
Thus u(r,𝜃 ) = 𝑛 =0 𝑟 (Ancos n𝜃 + Bn sin n𝜃 )
∞ n
= a0 / 2 + 𝑛=1 𝑟 (Ancos n𝜃 + Bn sin n𝜃 ) where A0 = a0 / 2
∞ n
= a0 / 2 + 𝑛=1 𝑟 (ancos n𝜃 + bn sin n𝜃 )
∞ n
Therefore f(𝜃) = = a0 / 2 + 𝑛=1 𝑎 (ancos n𝜃 + bn sin n𝜃 )
2𝜋
an = 1/𝜋𝑎𝑛 0
𝑓 𝜑 cos 𝑛𝜑 𝑑𝜑
2𝜋
bn = 1/𝜋𝑎𝑛 0
𝑓 𝜑 sin 𝑛𝜑 𝑑𝜑
2𝜋 ∞ n 2𝜋
u(r,𝜃 )=1/2𝜋 0
𝑓 𝜑 𝑑𝜑 + 1/𝜋 𝑛=1(𝑟/𝑎) 0 cos 𝑛𝜃 cos 𝑛𝜑 + sin 𝑛𝜃 sin 𝑛𝜑d𝜑
2𝜋 ∞ n
u(r,𝜃 )=1/𝜋 { 0
𝑓 𝜑 [1/2 + 𝑛=1(𝑟/𝑎) cos 𝑛𝜃 cos 𝑛𝜑 + sin 𝑛𝜃 sin 𝑛𝜑]d𝜑 }
2𝜋 ∞ n
u(r,𝜃 )=1/𝜋 { 0
𝑓 𝜑 [1/2 + 𝑛=1(𝑟/𝑎) 𝑐𝑜𝑠 n(𝜑 − 𝜃)]d𝜑 } ...(1)
∞ n
let c = 𝑛=1(𝑟/𝑎) 𝑐𝑜𝑠 n(𝜑 − 𝜃)
∞ n
S= 𝑛=1(𝑟/𝑎) 𝑠𝑖𝑛 n(𝜑 − 𝜃)
∞
C + is = 𝑛=1 {(𝑟/𝑎) 𝑒 𝑖(𝜑−𝜃) }n
2 2
C = r/a [cos(𝜑 − 𝜃) ± 𝑟 𝑎2 ] / 1-(2 r/a) ( cos 𝜑 − 𝜃 + 𝑟 𝑎2 ]
r 𝑟2
1 2𝜋 cos φ−θ + 2
a 𝑎
u(r,𝜃 )=2𝜋 { 0
𝑓 𝜑 [1/2 + 𝑟 𝑟2
]}
1−2 cos 𝜑 −𝜃 + 2
𝑎 𝑎
1 2𝜋 𝑎 2 −𝑟 2
u(r,𝜃 )= 2𝜋 { 0 𝑎 2 −2𝑎𝑟𝑐𝑜𝑠 (𝜑−𝜃)
𝑓 𝜑 𝑑𝜑
This is the Poisson’s integral formula for a circle.
0 0 ≤ 𝜃 ≤ 2𝜋
2
A cos n B
r , C n r Dn r sin n
n n
n n
n 0
Cn 0n
r , r n An cos n Bn sin n
n 0
r a cos n bn sin n
= a0
n
n
...(1)
2 n 1
a a cos n bn sin n
f(𝜃)= a 0
n
...(2)
n
2 n 1
2
f d
1
a0 0
n 2
a
a f cosn d
n
0
n 2
a
b f sin n d
n
0
2 2 n
a 1
r , f d cos n cos n sin n sin n f d
1
2 n 1 0
0
r
2 n 2
a
r ,
1
f d
1
1
cos n f d
2 0
n1 r 0
2 n
1 a
r , f { cos }d
1
0
2 n1 r
...(3)
n
a
Let C = cos n
n 1 r
n
a
And S = sin n
n 1 r
n
a
Therefore 𝐶 + 𝑖𝑆 = e i ( )
n 1 r
a i ( )
e
= r
a i ( )
1 e
r
cos( ) a2
a
r r
C= 2
1 cos( ) a2
2a
r r
2
cos( ) a2
a
2 r
r , r
f { 2
1 1
}d ...(4)
2
1 cos( ) a2
0 2a
r r
2
(r a ) f ( )d
2
2
1
=
2 2
2ar cos( ) a
2
0 r
INTERIOR NEUMANN PROBLEM FOR A CIRCLE
To find the value of U at any point in the interior of the circle r=a such that
u 0;
2
0≤ 𝑟 < 𝑎; 0≤ 𝜃 ≤ 2𝜋
u u r ,
And g ( ) on r=a
n r
By the method of separation of variable the general solution of the given equation is
given by
u r , (C r D r ) A cos n B sin n
n n
n n n n
n 0
a cos n b sin n
Therefore u r , r
n
n n
n 0
2 r a b n 1
n n
u
a cos n b sin n
Therefore nr n1
r n 1 n n
u (r , )
Now g ( ) where
r
2
1
a n
na n1 g ( ) cosnd
0
2
1
bn na n1 g ( ) sin nd
0
Therefore
n 2
a r
u (r , ) g ( )cos n cos n sin n sin n d
0
2 na n 1
n 1
0
Home Assignment
Exercise: By separating the variables, show that the equation ∇2 v = 0 has a solution
of the form 𝐴 exp ±𝑛𝑥 ± 𝑖𝑛𝑦 where A and n are constants
Where 𝐴𝑟 ′𝑠 being constants are plane harmonic functions satisfying the conditions
𝑣 𝑥, 0 = 0, 𝑣 𝑥, 𝑎 = 0, 𝑣 𝑥, 𝑦 → 0, 𝑎𝑠 𝑥 → ∞
The general equation for heat transfer is governed by the following equations
𝜕𝑇
= 𝑘∇2 T
𝜕𝑡
𝜕𝑇 𝜕2 𝜕2 𝜕2
Where is the time derivative and ∇2 = 𝜕𝑥 2 + 𝜕𝑦 2 + 𝜕𝑧 2 represents the derivative
𝜕𝑡
w.r.t., space.
𝜕𝑇 𝜕2𝑇
Heat Equation: The heat conduction equation = 𝑘 𝜕𝑥 2
𝜕𝑡
May have numerous solutions unless a set of initial and boundary conditions are
satisfied. The boundary conditions are mainly of three types and are briefly given
below.
Boundary condition I: The temperature is prescribed all over the boundary surface.
This type of boundary condition depends on the problem under investigation. Some
times the temperature on the boundary surface is a function of position only or is a
function of time only or a constant. A special case includes T(r, t ) = 0 on the surface
of boundary, which is called a homogenous boundary condition.
𝜕𝑇
Boundary condition II: The flux of heat, i.e. the normal derivative of temperature 𝜕𝑛
is prescribed on the surface of boundary. This is called the Neumann condition. A
𝜕𝑇
special case includes 𝜕𝑛 =0 on the boundary.This homogenous boundary condition is
also called insulated boundary condition which states that the heat flow across the
surface is zero.
Boundary condition III: A linear combination of the temperature and the heat flux is
prescribed on the boundary
𝜕𝑇
i.e. K +ht = G(𝑥, t )
𝜕𝑛
this type of boundary condition is called Robins condition. It means that the boundary
surface dissipates heat by convection. By Newton’s law of cooling, we have
𝜕𝑇
K𝜕𝑛 = h(T – Ta )
𝜕𝑇
𝐾 + 𝑇 = 0
𝜕𝑛
The other boundary conditions such as the heat transfer due to radiation obeying the
fourth power temperature law and these associated with change of phase like melting,
ablation etc. give rise to non linear boundary conditions.
𝜕𝑇 𝜕2𝑇
= 𝑘 𝜕𝑥 2 ...(1)
𝜕𝑡
be the solution of the differential equation (1) substituting from (2) into (1) we obtain
𝑋ʹʹ 1 𝑌ʹ
= = 𝜆 (separation parameter) then we have
𝑋 𝐾𝑌
𝑑 2 𝑋ʹʹ
− 𝜆X = 0 ...(3)
𝑑𝑥 2
𝑑𝑌
− 𝐾𝜆𝑌 = 0 ...(4)
𝑑𝑡
Case II: let 𝜆 = - 𝛼 2 , 𝛼 is positive, then solution will have the form
2 𝑘𝑡
Which provide us X = C1 cos 𝛼x+ C2 sin 𝛼x, and Y = C3𝑒 −𝛼 ...(6)
Thus various possible solutions of the one dimensional heat conduction equation (1)
are
2
𝑇(𝑥, 𝑡) = (A𝑒 𝛼𝑥 +B𝑒 −𝛼𝑥 ) 𝑒 𝑘𝛼 t
2
𝑇(𝑥, 𝑡) = (𝐴 cos 𝛼𝑥 + 𝐵 sin 𝛼𝑥 ) 𝑒 − 𝛼 kt
...(8)
𝜕𝑇 𝜕2𝑇
= 2 . . . (1)
𝜕𝑡 𝜕𝑥
Satisfying the conditions
(1) T → 0, as t → ∞
(2) T = 0, for x = 0 and x = a for all t > 0
(3) T = x, when t = 0 and 0 < 𝑥 < 𝑎 is
∞ 𝑛−1 𝑛𝜋
𝑇(𝑥, 𝑡) = 2a /𝜋 𝑛=1(−1) /n ) sin ( 𝑎 x) exp[-(n𝜋 /𝑎)2 t]
Clearly solutions represented by (1) and (2) does not satisfy the given conditions.
Therefore the most feasible solution for the equation (1) can be treated (2)
2
𝑇(𝑥, 𝑡) = (𝐴 cos 𝛼𝑥 + 𝐵 sin 𝛼𝑥 ) 𝑒 −𝛼 𝑡 using the boundary condition (2) we have
2𝑡
0 = [A(1) + B(0) ] 𝑒 −𝛼
2𝑡
Or 0 = A𝑒 −𝛼
Or A = 0
2𝑡
Also T(0,t) = 0 = (B sin 𝛼𝑎)𝑒 −𝛼
2𝑡
Since B ≠ 0 and 𝑒 −𝛼 ≠0
⇒ sin 𝛼𝑎 = 0
𝑛𝜋 𝑛2𝜋2
= Bsin 𝑥 exp − 𝑡
𝑎 𝑎2
Since the heat conduction equation is linear therefore the most general solution is
obtained by applying the principle of superposition
∞ 𝑛𝜋 𝑛2𝜋2
i.e. T(x, t) = 𝑛=1 𝐵 n sin 𝑎
𝑥 exp − 𝑎2
𝑡
T = x for t = 0
∞ 𝑛𝜋
⇒x== 𝑛=1 𝐵 n sin 𝑥 ×1 (∵ t = 0)
𝑎
𝑎 𝑛𝜋
Which is a half range Fourier sine series therefore Bn = 2/a 0
𝑥𝑠𝑖𝑛( x) dx
𝑎
𝑛𝜋
Let 𝑥=z
𝑎
𝑛𝜋
𝑑𝑥 =d z
𝑎
For x = 0, z= 0
For x = a, z = n𝜋
2 𝑛𝜋 𝑎 2
Therefore 𝐵𝑛 = 𝑧𝑠𝑖𝑛𝑧𝑑𝑧
𝑎 0 𝑛𝜋 2
2𝑎 −1𝑛 +1
= 𝜋 𝑛
2𝑎 ∞ −1 𝑛 +1 𝑛𝜋 𝑛2𝜋2
Therefore 𝑇(𝑥, 𝑡) = 𝑛=1 sin 𝑥 exp − 𝑡
𝜋 𝑛 𝑎 𝑎2
EX: The ends A and B of a rod, 10 𝑐𝑚 in length are kept at temperature 00c and 1000
c respectively until the steady state conditions prevails. Suddenly the temperature at
the end A is increased to 200 c and the end B is decreased to 600 c. Find the
temperature distribution in rod at time at t.
𝑇 (0, 𝑡) = 10
𝑇(10, 𝑡) = 100
𝑑2𝑇
For steady state =0
𝑑𝑥 2
Ts(x) = 10x
Similarly when the temperature at the ends A and B are changed to 200c and 600C,
the final steady temperature in rod is
Ts(x) = 4x + 20
Which will be attained after long time. At any instant of time the temperature
n 2 n
𝑇(𝑥, 𝑡) = 4𝑥 + 20 + C n exp k t sin x
n 1 10 10
n
10𝑥 = 4𝑥 + 20 + C
n 1
n sin x
10
Where Cn =
2 10
6 x 20 sin n x dx
10 0
10
=
1
5
1 800
n
n
200
n
R 1 R 1 z 1
2
2
R r R r z k
R 1 R 1 z
(say)
2
2
2
R r R r z
The equation in Z, R and becomes
z 2 z 0 ...(3)
R 1 R
R r R
2 2 r 2
2
Therefore
2 0 ...(4)
2 2
And R R 2 R 0
1 2
...(5)
r r
Equations (2) and (4) have particular solutions of the form
e
2
k t
C cos D sin
z z
z Ae B e
( )r
2 2 2
)r C 2 Y
2
R(r) = CJ (
1
Where J
(r)and Y (r ) are Bessel functions of order of first and second kind
respectively. Equation (5) is singular for r = 0, the physically meaningful solution
must be twice continuously differentiable in 0≤ 𝑟 ≤ 𝑎.
Hence equation (5) has only one bounded solution
i.e. R(r) =
J 2 r
2
Finally the general solution of equation (1) is given as
T(r,𝜃, 𝑧, 𝑡) = exp K t A e B e
2
z z
C cos D sin j
( )r
2 2
...(6)
Assignment
EX: Find the solution of the diffusion equation
T
= K 2T
t
Ex: A uniform rod of length l with thermally insulated surface is initially at
temperature 0 At t=0, one end is suddenly cooled to 00 C
And subsequently maintained at this temperature, the other end remains thermally
insulated. Find the temperature distribution 𝜃(𝑥, 𝑡).
EX: Find the solution of the 1-D diffusion equation satisfying the following
conditions
(i) 𝑇 is bounded as 𝑡 → ∞
𝜕𝑇
(ii) = 0, ∀ 𝑡
𝜕𝑥 𝑥=0
(iii) 𝑇 𝑥, 0 = 𝑥 𝑎 − 𝑥 ; 0<𝑥<0
EX: Solve the boundary value problem
𝜕𝑢 2
𝜕2𝑢
=𝛼 , 0<𝑥<𝑡
𝜕𝑡 𝜕𝑥 2
Subject to the conditions
𝜕𝑢 0,𝑡
(i) =0
𝜕𝑥
𝜕𝑢 𝑙,0
(ii) =0
𝜕𝑥
(iii) 𝑢 𝑥, 0 = 𝑥
EX: Solve the following equation
𝜕𝑢 𝜕 2 𝑢
=
𝜕𝑡 𝜕𝑥 2
Subject to the conditions
(i) 𝑢 𝑥, 0 = 3 sin 𝑛𝜋𝑥
(ii) 𝑢 0, 𝑡 = 0 = 𝑢 𝑙, 𝑡 , 0 < 𝑥 < 𝑙, 𝑡 > 0.
𝜕𝑢 𝜕 2 𝑢
=
𝜕𝑡 𝜕𝑥 2
Subject to the conditions
(i) 𝑢 𝑥, 0 = 3 sin 𝑛𝜋𝑥
(ii) 𝑢 0, 𝑡 = 0 = 𝑢 𝑙, 𝑡 , 0 < 𝑥 < 𝑙, 𝑡 > 0.
EX: Find the solution of the equation
𝜕𝑣 𝜕2𝑣
=𝑘 2
𝜕𝑡 𝜕𝑥
Subject to the conditions
(i) 𝑣 = 𝑣0 sin 𝑛𝑡 where 𝑥 = 0 ∀𝑡
(ii) 𝑣=0 𝑥→∞
Exercise : show that in the absence of a charge, the electric field and the magnetic
field in the Maxwell’s equation satisfy the wave equation.
Solution: we have
1 H
Curl E = E
C t
1 H
Consider E
C t
1
= H
C t
1 2 E
This implies E
C 2 t 2
E 2 E 4 2 E
= - 2 E
1 2 E
2 E
C 2 t 2
2E
Or C 2 2E
t 2
Which is a wave equation satisfied by E
Similarly we can observe the magnetic field H satisfies the wave equation
2H
c2 2H .
t 2
2u 2 u
2
We have c ...(1)
t 2 x 2
2U 2U
XT and TX
t 2 x 2
T X
XT C 2TX C2 (say)
T X
T T 0 ...(2)
And C 2 X X 0 ...(3)
t t
T = Ae Be
CASE I: If 0 say k
2
Similarly
C 2 X K 2 X 0
K K
X ( x) D e C E e
X X
C
K K
Kt
U ( x, t ) ( A e B e ) ( D eC E e
Kt X X
C )
Case II: if 0
Then T 0 and C 2 X =0
T Ct D and X Ax B
Therefore 𝑈(𝑥, 𝑡) = ( Ax B )( Ct D )
T C 2 X
K 2 and K 2
T X
T K 2T 0 and C 2 X K 2 X 0
K
So T = A cos Kt B sin Kt and
K
X = D cos x E sin x
C C
K
Therefore U(x, t) = A cos Kt B sin Kt D cos x E sin x
K
C C
REMARK: From the above solutions of the wave equation for 0 ≤ 𝑥 ≤ 𝑙 and 𝑡 > 0
𝑈(0, 𝑡) = 0 ; 𝑡 > 0,
𝑈(𝑙, 𝑡) = 0
0 = U (0,t) = Ae Kt
BeK
t
D E
D+E = 0 ...(5)
Now U(l, t) =0
Kl
K
U(l, t) = Ae Kt
Be Kt De C Ee C l =0
Kl K
De Ee
C C
l
=0
K
De E0
2 l
C ...(6)
K
e 1 By comparing coefficients in (5) and (6)
2 l
C
K K
2 l 0 l 0
C C
K
Either l 0 or 0
C
B0
Implies A =0
Implies A=0=B
U(x, t) = A cos Kt B sin Kt D cos x E sin
K K
x
C C
Now U (0,t) =0
A cos Kt B sin Kt ( D) 0
⇒ D =0
K
Also 0 = 𝑈 (𝑙, 𝑡) = A cos Kt B sin Kt E sin l
C
K
⇒ E sin l 0
C
K
⇒ l n
C
K n
⇒
C l
n
U ( x, t ) A cos Kt B sin Kt E sin x
l
n cn cn
U ( x, t ) En sin x An cos t Bn sin t
l l l
n
Ex: By the separation of variables, show that one dimensional wave equation
2Z 1 2Z
x 2 c 2 t 2
2V 2V
LC
x 2 t 2
Appropriate to the case when the periodic e.m.f. V0 cos pt is applied at the end x=0
of the line.
Exercise: A tightly stretched string with fixed end points 𝑥 = 0 and 𝑥 = 𝑙 is
initially in a position given
x
y y 0 sin 3
l
dy
0
dt
Let y(x, t) = X(x) T(t) be the solution
Then
2 y 2 X
T TX
x 2 t 2
2 y 2T
and X XT
x 2 t 2
We have
1
TX XT
C2
Therefore
Now,
y(0, t ) 0
A0
and y( x, t ) B sin x C cos ct D sin ct
y( x,0)
0
t
And
y
0 [ A cos x B sin x cD cos ct cC sin ct ]t 0
t t 0
Thus
y( x, t ) E sin x cosct
Where 𝐸 = 𝐵𝑐
E sin l cosct 0
sin l 0
l n
n
l
n n
y ( x, t ) E n sin x cos ct
n l l
y ( x,0) y 0 sin 3 x
l
Therefore
n
y 0 sin 3 x E n sin x
l n l
2 3
E1 sin x E 2 sin x E3 sin x ...
l l l
We know that
3 sin x sin 3x
sin 3 x
4
Or
x 3x
3 sin l sin l
y0 E sin x E sin 2x E sin 3x ...
1 2 3
4 l l l
3 y0 y0
E1 , E2 0, E3 , E4 E5 ... 0
4 4
3 y0 x y 3x
y ( x, t ) sin cos ct 0 sin cos ct
4 l l 4 l l
1 u 1 2u
r 2 2
r r r C t ...(1)
Assume that
U F (r ) e
iwt
Acts as a solution
U
F (r ) e
iwt
r
U
rF (r ) e
iwt
r
r
u
r F (r ) e rF (r ) e
iwt iwt
r r
2u
w 2 F (r ) e
iwt
t 2
1
r
1
F (r ) e rF (r ) e 2 w F (r ) e
iwt iwt
C
2 iwt
F (r ) w 2
F (r ) 2 F (c ) 0
r c
wr wr
F A J O B yo
c c
wr wr wr wr
F C1 J O i y C 2 J o i y
c o
c c o
c
wr wr iwt
U A J O B y e
c o
c
2u 2 u
2
C f ( x, t )
t 2 x 2
u ( x,0)
u ( x,0) ( x), ( x) ...(1)
t
2u 2 u
2
and C f ( x, t ) ...(2)
t 2 x 2
u 2 ( x,0)
u 2 ( x,0) 0 and 0 ...(3)
t
Integrating equation (2) over the region we get
2u 2 u
2
R t 2
C dxdt f ( x, t )dxdt
x 2 R
2
u 2 dt c u 2 dx f ( x, t )dxdt . . . (4)
R
t x R
dx
And along PA, c
dt
Cu 2 P Cu 2 B Cu 2 A Cu 2 P f ( x, t )dxdt
R
Using conditions given in (3), we have
u 2 A u 2 B 0
Therefore we have
2Cu 2 P f ( x, t )dxdt
R
x c ct
1 t0 t0
u 2 P
2C 0 xc ct
f ( x, t )dxdt
t0
Where P(x0,t0) is any arbitrary point. From the initial conditions associated with the
homogenous system, we know that
x ct
1
1
d
c x ct
u1 x, t x ct x ct
2
Hence the complete solution of the inhomogeneous wave equation in one dimensional
system is given by
u ( x, t ) u1 u 2
2u 2u 1 2u
x 2 y 2 C 2 t 2 ...(1)
u ( x, y, t ) X x Y y T t
Let
Now
2u 2u 2u
YT X , XTY , XYT
x 2 y 2 t 2
1
YTX XTY XYT
C2
X Y 1 T
2
X Y C T
d2X
2
k2X 0
dx
And
d 2Y
L2 X 0
dy 2
So that
d 2T
dt 2
K 2 L2 C 2T 0
X C1 cosKx C2 sin Kx
Y C3 cosLy C4 sin Ly
And
C 6
D Alembert’s solution of one dimensional wave equation:
2u 2 u
2
; x , 𝑡 > 0
t 2 c x 2
...(1)
𝑈(𝑥, 0) = ( x) f ( x) g ( x) ...(4)
Also,
u ( x,0)
=v(x) = C f ( x) g ( x)
t
( x) 1 x
f ( x) v( s)ds
2 2c 0
( x)
1
x
g ( x) v( s)ds
2 2c 0
x ct 1 x ct x ct 1 x ct
𝑈(𝑥, 𝑡) =
2
2c 0
v ( s ) ds
2
2c 0
v( s)ds
x ct x ct 1
x ct
2c x ct
𝑈(𝑥, 𝑡) = v( s)ds
2
Note: If v=0 i.e. the string is released from rest, the solution takes the form
x ct x ct
𝑈(𝑥, 𝑡) =
2
2v
C 2 2 v 0 ...(*)
t 2
vx,0, 0
v
x,0, F ( x, A) ... (1)
t
Where F ( x, A) denotes a continuous function defined for x in R3
t
ifu( x, t ) vx, t , d
0
xR , t 0
3
u x,0
u ( x,0) 0
t
Also for
t
u ( x, t ) vx, t , d
0 ...(3)
To be the solution of (2) where v(x, t-λ, λ) is one parameter family solution of (*)
Also v(x,0,λ)=0 for t=λ
Differentiating eq. (3) w. r. t., 𝑡 under the integral sign and using Leibnitz rule we
have
u v
t
v( x,0, ) x, t , d
t t
0 ...(4)
2u 2v
t
0 t 2 x,t , d
t 2 vt
x ,0,
using (*)
u v
2 2
v x,0, C u
2
u 2 2
t 2 t
u C u v x,0,
2
2 2
t 2 t
u C u F ( x,0, )
2
2 2
t 2
The function 𝑣(𝑥, 𝑡, 𝜆) is called the pulse function or the force function.
𝜕2𝑢 2
𝜕2𝑢 𝜕2𝑢
= 𝑐 + 0≤𝑥≤𝑎
𝜕𝑡 2 𝜕𝑥 2 𝜕𝑦 2
0≤𝑦≤𝑏 ...(1)
𝑢 0, 𝑦, 𝑡 = 𝑢 𝑥, 0, 𝑡 = 𝑢 𝑥, 𝑏, 𝑡 = 𝑢 𝑎, 𝑦, 𝑡 = 0
𝑢 𝑥, 𝑦, 0 = 𝑓 𝑥, 𝑦
𝜕𝑢 𝑥, 𝑦, 0
= 𝑔 𝑥, 𝑦
𝜕𝑡
Let 𝑢 𝑥, 𝑦, 𝑡 = 𝑋 𝑥 𝑌 𝑦 𝑇 𝑡 be the solution of (1)
Then 𝑇 ′′ + 𝜆2 𝑐 2 𝑇 = 0
𝑋′′ 𝑌′′
⇒ + 𝜆2 = − = 𝜇 2 (𝑠𝑎𝑦)
𝑋 𝑌
𝑋′′ 𝑌′′
Then + 𝜆2 = 𝜇 2 and = −𝜇 2
𝑋 𝑌
⇒ 𝑋 ′′ + 𝜆2 − 𝜇 2 𝑋 = 0 and 𝑌 ′′ + 𝜇 2 𝑌 = 0
𝑋 𝑥 = 𝐴 cos 𝑝𝑥 + 𝐵 sin 𝑝𝑥
𝑌 𝑦 = 𝐶 cos 𝑞𝑦 + 𝐷 sin 𝑞𝑦
𝑇 𝑡 = 𝐸 cos 𝑟𝑡 + 𝐹 sin 𝑟𝑡
Also 𝑢 𝑥, 0, 𝑡 = 0, ⇒𝐶=0
And 𝑢 𝑎, 𝑦, 𝑡 = 0 ⇒ sin 𝑝𝑎 = 0
𝑚𝜋
⇒ 𝑝𝑎 = 𝑚𝜋 ⇒ 𝑝 =
𝑎
Also 𝑢 𝑥, 𝑏, 𝑡 = 0 ⇒ sin 𝑞𝑏 = 0
𝑛𝜋
⇒ 𝑞𝑏 = 𝑛𝜋 ⇒ 𝑞 =
𝑏
Now using the principle of superposition we get,
∞ ∞ 𝑚𝜋 𝑛𝜋
𝑢 𝑥, 𝑦, 𝑡 = 𝑚 =1 𝑛=1 𝐴𝑚𝑛 cos 𝑟𝑐𝑡 + 𝐵𝑚𝑛 sin 𝑟𝑐𝑡 sin 𝑥 sin 𝑦 . . .
𝑎 𝑏
(A)
𝑚2 𝑛2
where 𝑟 2 = 𝑝2 + 𝑞 2 = 𝜋 2 + 𝑏2
𝑎2
𝑢 𝑥, 𝑦, 0 = 𝑓 𝑥, 𝑦
which implies
∞ ∞ 𝑚𝜋 𝑛𝜋
𝑓 𝑥, 𝑦 = 𝑚 =1 𝑛=1 𝐵𝑚𝑛 sin 𝑥 sin 𝑦 ... (B)
𝑎 𝑏
𝜕𝑢 𝑥,𝑦 ,0
And also = 𝑔 𝑥, 𝑦
𝜕𝑡
∞ ∞ 𝑚𝜋 𝑛𝜋
𝑔 𝑥, 𝑦 = 𝑐𝑟 𝑚 =1 𝑛=1 𝐵𝑚𝑛 sin 𝑥 sin 𝑦 ...(C)
𝑎 𝑏
where
𝑎 𝑏
4 𝑚𝜋 𝑛𝜋
𝐵𝑚𝑛 = 𝑔 𝑥, 𝑦 sin 𝑥 sin 𝑦 𝑑𝑥𝑑𝑦
𝑎𝑏𝑐𝑟 𝑎 𝑏
0 0