CECB423 Numerical Methods for Civil Engineers

Topic 9-Introduction to Finite Element and Finite Difference Methods

9.1 Boundary-Value Problems
9.2 Finite Element Method

Boundary Value Problem

(BVP)

Boundary condition Boundary condition with Boundary condition

without derivative derivative (Initial) with derivative (End)

a) Boundary condition without derivative

Consider the general linear two point BVP,

𝑦𝑦 ′′ (𝑥𝑥) + 𝑝𝑝(𝑥𝑥)𝑦𝑦 ′ (𝑥𝑥) + 𝑞𝑞(𝑥𝑥)𝑦𝑦(𝑥𝑥) = 𝑟𝑟(𝑥𝑥); 𝑎𝑎 ≤ 𝑥𝑥 ≤ 𝑏𝑏

With the boundary condition (without derivative)

𝑦𝑦(𝑎𝑎) = 𝛼𝛼 and 𝑦𝑦(𝑏𝑏) = 𝛽𝛽

To derive the finite-difference equation, the interval [a,b] is divided by step size h,

h h h h h

𝑥𝑥0 = 𝑎𝑎 𝑥𝑥1 𝑥𝑥2 𝑥𝑥3 = 𝑥𝑥4 = 𝑥𝑥𝑛𝑛 = 𝑏𝑏

= 𝑎𝑎 + ℎ = 𝑎𝑎
𝑦𝑦0 = 𝛼𝛼 + 2ℎ 𝑦𝑦3 = 𝑦𝑦4 = 𝑦𝑦𝑛𝑛 = 𝛽𝛽
𝑦𝑦1 =
𝑦𝑦 =

First boundary condition at Second boundary condition at

𝑥𝑥 = 𝑎𝑎 𝑥𝑥 = 𝑏𝑏
𝑦𝑦(𝑎𝑎) = 𝛼𝛼 𝑦𝑦(𝑏𝑏) = 𝛽𝛽
𝑦𝑦(𝑥𝑥0 ) = 𝛼𝛼 𝑦𝑦(𝑥𝑥𝑛𝑛 ) = 𝛽𝛽
𝑦𝑦0 = 𝛼𝛼 𝑦𝑦𝑛𝑛 = 𝛽𝛽
 CECB423 Numerical Methods for Civil Engineers

At grid points 𝑥𝑥 = 𝑥𝑥𝑖𝑖 , the differential equation can be written as

𝑦𝑦 ′′ (𝑥𝑥𝑖𝑖 ) + 𝑝𝑝(𝑥𝑥𝑖𝑖 )𝑦𝑦 ′ (𝑥𝑥𝑖𝑖 ) + 𝑞𝑞(𝑥𝑥𝑖𝑖 )𝑦𝑦(𝑥𝑥𝑖𝑖 ) = 𝑟𝑟(𝑥𝑥𝑖𝑖 )

By using central differentiation

𝑦𝑦𝑖𝑖+1 − 𝑦𝑦𝑖𝑖−1
𝑦𝑦 ′ (𝑥𝑥𝑖𝑖 ) ≈
2ℎ
′′ (𝑥𝑥 )
𝑦𝑦𝑖𝑖+1 − 2𝑦𝑦𝑖𝑖 + 𝑦𝑦𝑖𝑖−1
𝑦𝑦 𝑖𝑖 ≈
ℎ2

Substituting the approximation into above equation,

𝑦𝑦𝑖𝑖+1 − 2𝑦𝑦𝑖𝑖 + 𝑦𝑦𝑖𝑖−1 𝑦𝑦𝑖𝑖+1 − 𝑦𝑦𝑖𝑖−1

� 2 � + 𝑝𝑝𝑖𝑖 � � + 𝑞𝑞𝑖𝑖 𝑦𝑦𝑖𝑖 = 𝑟𝑟𝑖𝑖
ℎ 2ℎ

And multiplying throughout by h2

ℎ
𝑦𝑦𝑖𝑖+1 − 2𝑦𝑦𝑖𝑖 + 𝑦𝑦𝑖𝑖−1 + 𝑝𝑝𝑖𝑖 (𝑦𝑦𝑖𝑖+1 − 𝑦𝑦𝑖𝑖−1 ) + ℎ2 𝑞𝑞𝑖𝑖 𝑦𝑦𝑖𝑖 = ℎ2 𝑟𝑟𝑖𝑖
2

Rearrange the equation,

ℎ ℎ
(1 − 𝑝𝑝𝑖𝑖 )𝑦𝑦𝑖𝑖−1 + (−2 + ℎ2 𝑞𝑞𝑖𝑖 )𝑦𝑦𝑖𝑖 + (1 + 𝑝𝑝𝑖𝑖 )𝑦𝑦𝑖𝑖+1 = ℎ2 𝑟𝑟𝑖𝑖
2 2

Expanding the finite difference equation for 𝑖𝑖 = 1,2,3, … . , 𝑛𝑛

i.e:

at i=1 ℎ ℎ
(1 − 𝑝𝑝1 )𝑦𝑦0 + (−2 + ℎ2 𝑞𝑞1 )𝑦𝑦1 + (1 + 𝑝𝑝1 )𝑦𝑦2 = ℎ2 𝑟𝑟1
2 2

at i=2 ℎ ℎ
(1 − 𝑝𝑝2 )𝑦𝑦1 + (−2 + ℎ2 𝑞𝑞2 )𝑦𝑦2 + (1 + 𝑝𝑝2 )𝑦𝑦3 = ℎ2 𝑟𝑟2
2 2

at i=3 ℎ ℎ
(1 − 𝑝𝑝3 )𝑦𝑦2 + (−2 + ℎ2 𝑞𝑞3 )𝑦𝑦3 + (1 + 𝑝𝑝3 )𝑦𝑦4 = ℎ2 𝑟𝑟3
2 2

Combine all equation in linear system of from 𝐴𝐴𝐴𝐴 = 𝑏𝑏

Solve the value of 𝑦𝑦1 , 𝑦𝑦2 , 𝑦𝑦3 , … … … . , 𝑦𝑦𝑛𝑛

 CECB423 Numerical Methods for Civil Engineers

Example:

Solve the BVP

1 1
𝑦𝑦 ′′ + 𝑦𝑦 ′ + 2 𝑦𝑦(𝑥𝑥𝑖𝑖 ) = 3; 𝑦𝑦(1) = 2, 𝑦𝑦(2) = 3
𝑥𝑥 𝑥𝑥

For x at interval [1,2] with h=0.2. Use four decimal places in all calculation.

0.2

𝑥𝑥0 = 1 𝑥𝑥1 = 1.2 𝑥𝑥2 = 1.4 𝑥𝑥3 = 1.6 𝑥𝑥4 = 1.8 𝑥𝑥5 = 2

𝑦𝑦0 = 2 𝑦𝑦1 = 𝑦𝑦2 = 𝑦𝑦3 = 𝑦𝑦4 = 𝑦𝑦5 = 3

1 1
𝑦𝑦 ′′ + 𝑦𝑦 ′ + 2 𝑦𝑦(𝑥𝑥𝑖𝑖 ) = 3
𝑥𝑥 𝑥𝑥
𝑦𝑦 ′′ (𝑥𝑥) + 𝑝𝑝(𝑥𝑥)𝑦𝑦 ′ (𝑥𝑥) + 𝑞𝑞(𝑥𝑥)𝑦𝑦(𝑥𝑥) = 𝑟𝑟(𝑥𝑥)

First boundary condition at Second boundary condition at

𝑦𝑦(1) = 2 𝑦𝑦(2) = 3
𝑦𝑦(𝑥𝑥0 ) = 2 𝑦𝑦(𝑥𝑥5 ) = 3
𝑦𝑦0 = 2 𝑦𝑦5 = 3

ℎ ℎ
(1 − 𝑝𝑝𝑖𝑖 )𝑦𝑦𝑖𝑖−1 + (−2 + ℎ2 𝑞𝑞𝑖𝑖 )𝑦𝑦𝑖𝑖 + (1 + 𝑝𝑝𝑖𝑖 )𝑦𝑦𝑖𝑖+1 = ℎ2 𝑟𝑟𝑖𝑖
2 2

Substitute p(x), q(x) and r(x)

ℎ 1 1 ℎ 1
�1 − � ∗ �� 𝑦𝑦𝑖𝑖−1 + �−2 + �ℎ2 ∗ 2 �� 𝑦𝑦𝑖𝑖 + �1 + � ∗ �� 𝑦𝑦𝑖𝑖+1 = ℎ2 ∗ 3
2 𝑥𝑥𝑖𝑖 𝑥𝑥𝑖𝑖 2 𝑥𝑥𝑖𝑖

Substitute the h=0.2

0.2 1 1 0.2 1
�1 − � ∗ �� 𝑦𝑦𝑖𝑖−1 + �−2 + �(0.2)2 ∗ 2 �� 𝑦𝑦𝑖𝑖 + �1 + � ∗ �� 𝑦𝑦𝑖𝑖+1 = (0.2)2 ∗
2 𝑥𝑥𝑖𝑖 𝑥𝑥𝑖𝑖 2 𝑥𝑥𝑖𝑖

at i=1 0.2 1 1 0.2 1

�1 − � ∗ �� 𝑦𝑦0 + �−2 + �0.22 ∗ �� 𝑦𝑦1 + �1 + � ∗ �� 𝑦𝑦2 = (0.2)2 ∗ 3
(x1=1.2) 2 𝑥𝑥1 (𝑥𝑥1 )2 2 𝑥𝑥1
(y0=2)
0.2 1 1 0.2 1
�1 − � ∗ �� (2) + �−2 + �0.22 ∗ 2
�� 𝑦𝑦1 + �1 + � ∗ �� 𝑦𝑦2 = (0.2)2 ∗ 3
2 1.2 (1.2) 2 1.2

1.8333 − 2.0278𝑦𝑦1 + 1.0833𝑦𝑦2 = 0.12

 CECB423 Numerical Methods for Civil Engineers

−𝟐𝟐. 𝟎𝟎𝟎𝟎𝟎𝟎𝟎𝟎𝒚𝒚𝟏𝟏 + 𝟏𝟏. 𝟎𝟎𝟎𝟎𝟎𝟎𝟎𝟎𝒚𝒚𝟐𝟐 = −𝟏𝟏. 𝟕𝟕𝟕𝟕𝟕𝟕𝟕𝟕

at i=2 0.2 1 1 0.2 1

�1 − � ∗ �� 𝑦𝑦1 + �−2 + �0.22 ∗ �� 𝑦𝑦2 + �1 + � ∗ �� 𝑦𝑦3 = (0.2)2 ∗ 3
(x2=1.4) 2 𝑥𝑥2 (𝑥𝑥2 )2 2 𝑥𝑥2

0.2 1 1 0.2 1
�1 − � ∗ �� 𝑦𝑦1 + �−2 + �0.22 ∗ 2
�� 𝑦𝑦2 + �1 + � ∗ �� 𝑦𝑦3 = (0.2)2 ∗ 3
2 1.4 (1.4) 2 1.4

𝟎𝟎. 𝟗𝟗𝟗𝟗𝟗𝟗𝟗𝟗𝒚𝒚𝟏𝟏 − 𝟐𝟐. 𝟎𝟎𝟎𝟎𝟎𝟎𝟎𝟎𝒚𝒚𝟐𝟐 + 𝟏𝟏. 𝟎𝟎𝟎𝟎𝟎𝟎𝟎𝟎𝒚𝒚𝟑𝟑 = 𝟎𝟎. 𝟏𝟏𝟏𝟏

at i=3 0.2 1 1 0.2 1

�1 − � ∗ �� 𝑦𝑦2 + �−2 + �0.22 ∗ �� 𝑦𝑦3 + �1 + � ∗ �� 𝑦𝑦4 = (0.2)2 ∗ 3
(x3=1.6) 2 𝑥𝑥3 (𝑥𝑥3 )2 2 𝑥𝑥3

0.2 1 1 0.2 1
�1 − � ∗ �� 𝑦𝑦2 + �−2 + �0.22 ∗ 2
�� 𝑦𝑦3 + �1 + � ∗ �� 𝑦𝑦4 = (0.2)2 ∗ 3
2 1.6 (1.6) 2 1.6

𝟎𝟎. 𝟗𝟗𝟗𝟗𝟗𝟗𝟗𝟗𝒚𝒚𝟐𝟐 − 𝟐𝟐. 𝟎𝟎𝟎𝟎𝟎𝟎𝟎𝟎𝒚𝒚𝟑𝟑 + 𝟏𝟏. 𝟎𝟎𝟎𝟎𝟎𝟎𝟎𝟎𝒚𝒚𝟒𝟒 = 𝟎𝟎. 𝟏𝟏𝟏𝟏

at i=4 0.2 1 1 0.2 1

�1 − � ∗ �� 𝑦𝑦3 + �−2 + �0.22 ∗ �� 𝑦𝑦4 + �1 + � ∗ �� 𝑦𝑦5 = (0.2)2 ∗ 3
(x4=1.8) 2 𝑥𝑥4 (𝑥𝑥4 )2 2 𝑥𝑥4
(y5=3)
0.2 1 1 0.2 1
�1 − � ∗ �� 𝑦𝑦3 + �−2 + �0.22 ∗ 2
�� 𝑦𝑦4 + �1 + � ∗ �� (3) = (0.2)2 ∗ 3
2 1.8 (1.8) 2 1.8

0.9444𝑦𝑦3 − 2.0123𝑦𝑦4 + 3.1668 = 0.12

𝟎𝟎. 𝟗𝟗𝟗𝟗𝟗𝟗𝟗𝟗𝒚𝒚𝟑𝟑 − 𝟐𝟐. 𝟎𝟎𝟎𝟎𝟎𝟎𝟎𝟎𝒚𝒚𝟒𝟒 = −𝟑𝟑. 𝟎𝟎𝟎𝟎𝟎𝟎𝟎𝟎

Solve the system using trigonal system

−2.0278 1.0833 0 0 𝑦𝑦1 −1.7133

0.9286 −2.0204 1.0714 0 𝑦𝑦2 0.1200
� � �𝑦𝑦 � = � �
0 0.9375 −2.0156 1.0625 3 0.1200
0 0 0.9444 −2.0123 𝑦𝑦4 −3.0468

𝑦𝑦1 1.9082
𝑦𝑦2 1.9904
�𝑦𝑦 � = � �
3 2.2115
𝑦𝑦4 2.5520
 CECB423 Numerical Methods for Civil Engineers

b) Boundary condition with derivative (Initial)

Consider the general linear two point BVP,

𝑦𝑦 ′′ (𝑥𝑥) + 𝑝𝑝(𝑥𝑥)𝑦𝑦 ′ (𝑥𝑥) + 𝑞𝑞(𝑥𝑥)𝑦𝑦(𝑥𝑥) = 𝑟𝑟(𝑥𝑥); 𝑎𝑎 ≤ 𝑥𝑥 ≤ 𝑏𝑏

With the boundary condition (with derivative 𝑥𝑥 = 𝑎𝑎)

𝑘𝑘1 𝑦𝑦(𝑎𝑎) + 𝑘𝑘2 𝑦𝑦′(𝑎𝑎) = 𝛼𝛼 and 𝑦𝑦(𝑏𝑏) = 𝛽𝛽

To derive the finite-difference equation, the interval [a,b] is divided by step size h,

h h h h h

𝑥𝑥−1 = 𝑥𝑥0 = 𝛼𝛼 𝑥𝑥1 = 𝑥𝑥2 = 𝑥𝑥3 = 𝑥𝑥𝑛𝑛 = 𝑏𝑏

𝑦𝑦−1 = 𝑦𝑦0 = 𝑦𝑦1 = 𝑦𝑦2 = 𝑦𝑦3 = 𝑦𝑦𝑛𝑛 = 𝛽𝛽

𝑦𝑦𝑖𝑖+1 − 𝑦𝑦𝑖𝑖−1
𝑦𝑦 ′ (𝑥𝑥𝑖𝑖 ) ≈
2ℎ
′′ (𝑥𝑥 )
𝑦𝑦𝑖𝑖+1 − 2𝑦𝑦𝑖𝑖 + 𝑦𝑦𝑖𝑖−1
𝑦𝑦 𝑖𝑖 ≈
ℎ2
First boundary condition at Second boundary condition at
𝑘𝑘1 𝑦𝑦(𝑎𝑎) + 𝑘𝑘2 𝑦𝑦′(𝑎𝑎) = 𝛼𝛼 𝑥𝑥 = 𝑏𝑏
𝑘𝑘1 𝑦𝑦(𝑥𝑥0 ) + 𝑘𝑘2 𝑦𝑦′(𝑥𝑥0 ) = 𝛼𝛼 𝑦𝑦(𝑏𝑏) = 𝛽𝛽
𝑘𝑘1 𝑦𝑦0 + 𝑘𝑘2 𝑦𝑦′0 = 𝛼𝛼 𝑦𝑦(𝑥𝑥𝑛𝑛 ) = 𝛽𝛽
𝑦𝑦1 − 𝑦𝑦−1 𝑦𝑦𝑛𝑛 = 𝛽𝛽
𝑘𝑘1 𝑦𝑦0 + 𝑘𝑘2 � � = 𝛼𝛼
2ℎ
2ℎ𝑘𝑘1
𝑦𝑦−1 =
𝑘𝑘2

At grid points 𝑥𝑥 = 𝑥𝑥𝑖𝑖 , the differential equation can be written as

𝑦𝑦 ′′ (𝑥𝑥𝑖𝑖 ) + 𝑝𝑝(𝑥𝑥𝑖𝑖 )𝑦𝑦 ′ (𝑥𝑥𝑖𝑖 ) + 𝑞𝑞(𝑥𝑥𝑖𝑖 )𝑦𝑦(𝑥𝑥𝑖𝑖 ) = 𝑟𝑟(𝑥𝑥𝑖𝑖 )

By using central differentiation

𝑦𝑦𝑖𝑖+1 − 𝑦𝑦𝑖𝑖−1
𝑦𝑦 ′ (𝑥𝑥𝑖𝑖 ) ≈
2ℎ
𝑦𝑦𝑖𝑖+1 − 2𝑦𝑦𝑖𝑖 + 𝑦𝑦𝑖𝑖−1
𝑦𝑦 ′′ (𝑥𝑥𝑖𝑖 ) ≈
ℎ2
Substituting the approximation into above equation,
 CECB423 Numerical Methods for Civil Engineers

𝑦𝑦𝑖𝑖+1 − 2𝑦𝑦𝑖𝑖 + 𝑦𝑦𝑖𝑖−1 𝑦𝑦𝑖𝑖+1 − 𝑦𝑦𝑖𝑖−1

� � + 𝑝𝑝𝑖𝑖 � � + 𝑞𝑞𝑖𝑖 𝑦𝑦𝑖𝑖 = 𝑟𝑟𝑖𝑖
ℎ2 2ℎ

And multiplying throughout by h2

ℎ
𝑦𝑦𝑖𝑖+1 − 2𝑦𝑦𝑖𝑖 + 𝑦𝑦𝑖𝑖−1 + 𝑝𝑝𝑖𝑖 (𝑦𝑦𝑖𝑖+1 − 𝑦𝑦𝑖𝑖−1 ) + ℎ2 𝑞𝑞𝑖𝑖 𝑦𝑦𝑖𝑖 = ℎ2 𝑟𝑟𝑖𝑖
2

Rearrange the equation,

ℎ ℎ
(1 − 𝑝𝑝𝑖𝑖 )𝑦𝑦𝑖𝑖−1 + (−2 + ℎ2 𝑞𝑞𝑖𝑖 )𝑦𝑦𝑖𝑖 + (1 + 𝑝𝑝𝑖𝑖 )𝑦𝑦𝑖𝑖+1 = ℎ2 𝑟𝑟𝑖𝑖
2 2

Expanding the finite difference equation for 𝑖𝑖 = 1,2,3, … . , 𝑛𝑛

i.e:

at i=0 ℎ ℎ
(1 − 𝑝𝑝0 )𝑦𝑦−1 + (−2 + ℎ2 𝑞𝑞0 )𝑦𝑦0 + (1 + 𝑝𝑝0 )𝑦𝑦1 = ℎ2 𝑟𝑟0
2 2
ℎ 2ℎ𝑘𝑘1 ℎ
(1 − 𝑝𝑝0 )( ) + (−2 + ℎ2 𝑞𝑞0 )𝑦𝑦0 + (1 + 𝑝𝑝0 )𝑦𝑦1 = ℎ2 𝑟𝑟0
2 𝑘𝑘2 2

at i=1 ℎ ℎ
(1 − 𝑝𝑝1 )𝑦𝑦0 + (−2 + ℎ2 𝑞𝑞1 )𝑦𝑦1 + (1 + 𝑝𝑝1 )𝑦𝑦2 = ℎ2 𝑟𝑟1
2 2

at i=2 ℎ ℎ
(1 − 𝑝𝑝2 )𝑦𝑦1 + (−2 + ℎ2 𝑞𝑞2 )𝑦𝑦2 + (1 + 𝑝𝑝2 )𝑦𝑦3 = ℎ2 𝑟𝑟2
2 2

Combine all equation in linear system of from 𝐴𝐴𝐴𝐴 = 𝑏𝑏

Solve the value of 𝑦𝑦0 , 𝑦𝑦1 , 𝑦𝑦2 , … … … . , 𝑦𝑦𝑛𝑛

 CECB423 Numerical Methods for Civil Engineers

Example:

Solve the BVP

1 1
𝑦𝑦 ′′ + 𝑦𝑦 ′ + 2 𝑦𝑦(𝑥𝑥𝑖𝑖 ) = 3; 𝑦𝑦 ′ (1) + 𝑦𝑦(1) = 1, 𝑦𝑦(2) = 3
𝑥𝑥 𝑥𝑥

For x at interval [1,2] with h=0.2. Use four decimal places in all calculation.

0.2

𝑥𝑥−1 = 𝑥𝑥0 = 1 𝑥𝑥1 = 1.2 𝑥𝑥2 = 1.4 𝑥𝑥3 = 1.6 𝑥𝑥4 = 1.8 𝑥𝑥5 = 2
𝑦𝑦−1 = 𝑦𝑦0 = 𝑦𝑦1 = 𝑦𝑦2 = 𝑦𝑦3 = 𝑦𝑦4 = 𝛽𝛽 𝑦𝑦5 = 3

𝑦𝑦𝑖𝑖+1 − 𝑦𝑦𝑖𝑖−1
𝑦𝑦 ′ (𝑥𝑥𝑖𝑖 ) ≈
2ℎ
𝑦𝑦𝑖𝑖+1 − 2𝑦𝑦𝑖𝑖 + 𝑦𝑦𝑖𝑖−1
𝑦𝑦 ′′ (𝑥𝑥𝑖𝑖 ) ≈
ℎ2
Second boundary condition at
First boundary condition at 𝑦𝑦(2) = 3
𝑦𝑦 ′ (1) + 𝑦𝑦(1) = 1, 𝑦𝑦(𝑥𝑥5 ) = 3
𝑦𝑦′0 + 𝑦𝑦0 = 1 𝑦𝑦5 = 3
𝑦𝑦1 − 𝑦𝑦−1
+ 𝑦𝑦0 = 1
2ℎ
𝑦𝑦1 − 𝑦𝑦−1 + 2ℎ𝑦𝑦0 = 2ℎ
−2ℎ + 𝑦𝑦1 + 2ℎ𝑦𝑦0 = 𝑦𝑦−1
𝑦𝑦−1 = −0.4 + 0.4𝑦𝑦0 + 𝑦𝑦1

1 1
𝑦𝑦 ′′ + 𝑦𝑦 ′ + 2 𝑦𝑦(𝑥𝑥𝑖𝑖 ) = 3
𝑥𝑥 𝑥𝑥
𝑦𝑦 ′′ (𝑥𝑥) + 𝑝𝑝(𝑥𝑥)𝑦𝑦 ′ (𝑥𝑥) + 𝑞𝑞(𝑥𝑥)𝑦𝑦(𝑥𝑥) = 𝑟𝑟(𝑥𝑥)

ℎ ℎ
(1 − 𝑝𝑝𝑖𝑖 )𝑦𝑦𝑖𝑖−1 + (−2 + ℎ2 𝑞𝑞𝑖𝑖 )𝑦𝑦𝑖𝑖 + (1 + 𝑝𝑝𝑖𝑖 )𝑦𝑦𝑖𝑖+1 = ℎ2 𝑟𝑟𝑖𝑖
2 2

at i=0 ℎ ℎ
(1 − 𝑝𝑝0 )𝑦𝑦−1 + (−2 + ℎ2 𝑞𝑞0 )𝑦𝑦0 + (1 + 𝑝𝑝0 )𝑦𝑦1 = ℎ2 𝑟𝑟0
2 2
x0=1.0
ℎ 1 1 ℎ 1
�1 − ∗ � (−0.4 + 0.4𝑦𝑦0 + 𝑦𝑦1 ) + �−2 + ℎ2 ∗ 2
� 𝑦𝑦0 + �1 + ∗ � 𝑦𝑦1 = ℎ2 ∗ 3
2 𝑥𝑥0 (𝑥𝑥0 ) 2 𝑥𝑥0
0.2 1 1 0.2 1
�1 − ∗ � (−0.4 + 0.4𝑦𝑦0 + 𝑦𝑦1 ) + �−2 + 0.22 ∗ � 𝑦𝑦0 + �1 + ∗ � 𝑦𝑦 = 0.22 ∗ 3
2 1.0 1.0 2 2 1.0 1
 CECB423 Numerical Methods for Civil Engineers

0.90(−0.4 + 0.4𝑦𝑦0 + 𝑦𝑦1 ) − 2.0400𝑦𝑦0 + 1.1000𝑦𝑦1 = 0.12

−𝟏𝟏. 𝟔𝟔𝟔𝟔𝟔𝟔𝟔𝟔𝒚𝒚𝟎𝟎 + 𝟐𝟐. 𝟎𝟎𝟎𝟎𝟎𝟎𝟎𝟎𝒚𝒚𝟏𝟏 = 𝟎𝟎. 𝟒𝟒𝟒𝟒𝟒𝟒𝟒𝟒

at i=1 ℎ ℎ
(1 − 𝑝𝑝1 )𝑦𝑦0 + (−2 + ℎ2 𝑞𝑞1 )𝑦𝑦1 + (1 + 𝑝𝑝1 )𝑦𝑦2 = ℎ2 𝑟𝑟1
2 2
x1=1.2
ℎ 1 1 ℎ 1
�1 − ∗ � 𝑦𝑦0 + �−2 + ℎ2 ∗ 2
� 𝑦𝑦1 + �1 + ∗ � 𝑦𝑦2 = ℎ2 ∗ 3
2 𝑥𝑥1 (𝑥𝑥1 ) 2 𝑥𝑥1
0.2 1 1 0.2 1
�1 − ∗ � 𝑦𝑦0 + �−2 + 0.22 ∗ � 𝑦𝑦1 + �1 + ∗ � 𝑦𝑦 = 0.22 ∗ 3
2 1.2 1.2 2 2 1.2 2
𝟎𝟎. 𝟗𝟗𝟗𝟗𝟗𝟗𝟗𝟗𝒚𝒚𝟎𝟎 − 𝟐𝟐. 𝟎𝟎𝟎𝟎𝟎𝟎𝟎𝟎𝒚𝒚𝟏𝟏 + 𝟏𝟏. 𝟎𝟎𝟎𝟎𝟎𝟎𝟎𝟎𝒚𝒚𝟐𝟐 = 𝟎𝟎. 𝟏𝟏𝟏𝟏

at i=2 ℎ ℎ
(1 − 𝑝𝑝2 )𝑦𝑦1 + (−2 + ℎ2 𝑞𝑞2 )𝑦𝑦2 + (1 + 𝑝𝑝2 )𝑦𝑦3 = ℎ2 𝑟𝑟2
2 2
x2=1.4
ℎ 1 1 ℎ 1
�1 − ∗ � 𝑦𝑦1 + �−2 + ℎ2 ∗ 2
� 𝑦𝑦2 + �1 + ∗ � 𝑦𝑦3 = ℎ2 ∗ 3
2 𝑥𝑥2 (𝑥𝑥2 ) 2 𝑥𝑥2
0.2 1 1 0.2 1
�1 − ∗ � 𝑦𝑦1 + �−2 + 0.22 ∗ � 𝑦𝑦2 + �1 + ∗ � 𝑦𝑦 = 0.22 ∗ 3
2 1.4 1.4 2 2 1.4 3
𝟎𝟎. 𝟗𝟗𝟗𝟗𝟗𝟗𝟗𝟗𝒚𝒚𝟏𝟏 − 𝟐𝟐. 𝟎𝟎𝟎𝟎𝟎𝟎𝟎𝟎𝒚𝒚𝟐𝟐 + 𝟏𝟏. 𝟎𝟎𝟎𝟎𝟎𝟎𝟎𝟎𝒚𝒚𝟑𝟑 = 𝟎𝟎. 𝟏𝟏𝟏𝟏

at i=3 ℎ ℎ
(1 − 𝑝𝑝3 )𝑦𝑦2 + (−2 + ℎ2 𝑞𝑞3 )𝑦𝑦3 + (1 + 𝑝𝑝3 )𝑦𝑦4 = ℎ2 𝑟𝑟3
2 2
x3=1.6
ℎ 1 1 ℎ 1
�1 − ∗ � 𝑦𝑦2 + �−2 + ℎ2 ∗ 2
� 𝑦𝑦3 + �1 + ∗ � 𝑦𝑦4 = ℎ2 ∗ 3
2 𝑥𝑥3 (𝑥𝑥3 ) 2 𝑥𝑥3
0.2 1 1 0.2 1
�1 − ∗ � 𝑦𝑦2 + �−2 + 0.22 ∗ � 𝑦𝑦3 + �1 + ∗ � 𝑦𝑦 = 0.22 ∗ 3
2 1.6 1.6 2 2 1.6 4
𝟎𝟎. 𝟗𝟗𝟗𝟗𝟗𝟗𝟗𝟗𝒚𝒚𝟐𝟐 − 𝟐𝟐. 𝟎𝟎𝟎𝟎𝟎𝟎𝟎𝟎𝒚𝒚𝟑𝟑 + 𝟏𝟏. 𝟎𝟎𝟎𝟎𝟎𝟎𝟎𝟎𝒚𝒚𝟒𝟒 = 𝟎𝟎. 𝟏𝟏𝟏𝟏

at i=4 ℎ ℎ
(1 − 𝑝𝑝4 )𝑦𝑦3 + (−2 + ℎ2 𝑞𝑞4 )𝑦𝑦4 + (1 + 𝑝𝑝4 )𝑦𝑦5 = ℎ2 𝑟𝑟4
2 2
x4=1.8
ℎ 1 1 ℎ 1
�1 − ∗ � 𝑦𝑦3 + �−2 + ℎ2 ∗ 2
� 𝑦𝑦4 + �1 + ∗ � 𝑦𝑦5 = ℎ2 ∗ 3
2 𝑥𝑥4 (𝑥𝑥4 ) 2 𝑥𝑥4
0.2 1 1 0.2 1
�1 − ∗ � 𝑦𝑦3 + �−2 + 0.22 ∗ 2
� 𝑦𝑦4 + �1 + ∗ � (3) = 0.22 ∗ 3
2 1.8 1.8 2 1.8
0.9444𝑦𝑦3 − 2.0123𝑦𝑦4 + 3.1668 = 0.12
𝟎𝟎. 𝟗𝟗𝟗𝟗𝟗𝟗𝟗𝟗𝒚𝒚𝟑𝟑 − 𝟐𝟐. 𝟎𝟎𝟎𝟎𝟎𝟎𝟎𝟎𝒚𝒚𝟒𝟒 = −𝟑𝟑. 𝟎𝟎𝟎𝟎𝟎𝟎𝟎𝟎
 CECB423 Numerical Methods for Civil Engineers

Solve the system using trigonal system

−1.6800 2.0000 0 0 0 𝑦𝑦0 0.4800

𝑦𝑦1
⎛ 0.9167 −0.0278 1.0833 0 0 ⎞ ⎛ ⎞ ⎛ 0.1200 ⎞
⎜ 0 0.9286 −2.0204 1.0714 0 ⎟ ⎜𝑦𝑦2 ⎟ = ⎜ 0.1200 ⎟
0 0 0.9375 −2.0156 1.0625 𝑦𝑦3 0.1200
⎝ 0 0 0 0.9444 𝑦𝑦
−2.0123⎠ ⎝ 4 ⎠ ⎝−3.0468⎠

𝑦𝑦0 1.9081
𝑦𝑦1
⎛ ⎞ ⎛1.8428⎞
⎜𝑦𝑦2 ⎟ = ⎜1.9455⎟
𝑦𝑦3 2.1837
𝑦𝑦
⎝ 4 ⎠ ⎝2.5388⎠

c) Boundary condition with derivative (End)

Consider the general linear two point BVP,

𝑦𝑦 ′′ (𝑥𝑥) + 𝑝𝑝(𝑥𝑥)𝑦𝑦 ′ (𝑥𝑥) + 𝑞𝑞(𝑥𝑥)𝑦𝑦(𝑥𝑥) = 𝑟𝑟(𝑥𝑥); 𝑎𝑎 ≤ 𝑥𝑥 ≤ 𝑏𝑏

With the boundary condition (with derivative 𝑥𝑥 = 𝑏𝑏)

𝑦𝑦(𝑎𝑎) = 𝛼𝛼 and 𝑘𝑘1 𝑦𝑦(𝑏𝑏) + 𝑘𝑘2 𝑦𝑦′(𝑏𝑏) = 𝛽𝛽

To derive the finite-difference equation, the interval [a,b] is divided by step size h,

h h h h h

𝑥𝑥0 = 𝑎𝑎 𝑥𝑥1 = 𝑥𝑥2 = 𝑥𝑥3 = 𝑥𝑥𝑛𝑛 = 𝑏𝑏 𝑥𝑥𝑛𝑛+1 =

𝑦𝑦0 = 𝛼𝛼 𝑦𝑦1 = 𝑦𝑦2 = 𝑦𝑦3 = 𝑦𝑦𝑛𝑛 = 𝑦𝑦𝑛𝑛+1 =

First boundary condition at Second boundary condition at

𝑥𝑥 = 𝑎𝑎 𝑘𝑘1 𝑦𝑦(𝑏𝑏) + 𝑘𝑘2 𝑦𝑦′(𝑏𝑏) = 𝛽𝛽
𝑦𝑦(𝑎𝑎) = 𝛼𝛼 𝑘𝑘1 𝑦𝑦(𝑥𝑥𝑛𝑛 ) + 𝑘𝑘2 𝑦𝑦′(𝑥𝑥𝑛𝑛 ) = 𝛽𝛽
 CECB423 Numerical Methods for Civil Engineers

𝑦𝑦(𝑥𝑥0 ) = 𝛼𝛼 𝑘𝑘1 𝑦𝑦𝑛𝑛 + 𝑘𝑘2 𝑦𝑦′𝑛𝑛 = 𝛽𝛽

𝑦𝑦0 = 𝛼𝛼 𝑦𝑦𝑛𝑛+1 − 𝑦𝑦𝑛𝑛−1
𝑘𝑘1 𝑦𝑦𝑛𝑛 + 𝑘𝑘2 � � = 𝛽𝛽
2ℎ
2ℎ𝑘𝑘1 2ℎ
𝑦𝑦𝑛𝑛+1 = 𝑦𝑦𝑛𝑛−1 − 𝑦𝑦𝑛𝑛 + 𝛽𝛽
𝑘𝑘2 𝑘𝑘2
At grid points 𝑥𝑥 = 𝑥𝑥𝑖𝑖 , the differential equation can be written as

𝑦𝑦 ′′ (𝑥𝑥𝑖𝑖 ) + 𝑝𝑝(𝑥𝑥𝑖𝑖 )𝑦𝑦 ′ (𝑥𝑥𝑖𝑖 ) + 𝑞𝑞(𝑥𝑥𝑖𝑖 )𝑦𝑦(𝑥𝑥𝑖𝑖 ) = 𝑟𝑟(𝑥𝑥𝑖𝑖 )

By using central differentiation

𝑦𝑦𝑖𝑖+1 − 𝑦𝑦𝑖𝑖−1
𝑦𝑦 ′ (𝑥𝑥𝑖𝑖 ) ≈
2ℎ
𝑦𝑦𝑖𝑖+1 − 2𝑦𝑦𝑖𝑖 + 𝑦𝑦𝑖𝑖−1
𝑦𝑦 ′′ (𝑥𝑥𝑖𝑖 ) ≈
ℎ2
Substituting the approximation into above equation,

𝑦𝑦𝑖𝑖+1 − 2𝑦𝑦𝑖𝑖 + 𝑦𝑦𝑖𝑖−1 𝑦𝑦𝑖𝑖+1 − 𝑦𝑦𝑖𝑖−1

� 2 � + 𝑝𝑝𝑖𝑖 � � + 𝑞𝑞𝑖𝑖 𝑦𝑦𝑖𝑖 = 𝑟𝑟𝑖𝑖
ℎ 2ℎ

And multiplying throughout by h2

ℎ
𝑦𝑦𝑖𝑖+1 − 2𝑦𝑦𝑖𝑖 + 𝑦𝑦𝑖𝑖−1 + 𝑝𝑝𝑖𝑖 (𝑦𝑦𝑖𝑖+1 − 𝑦𝑦𝑖𝑖−1 ) + ℎ2 𝑞𝑞𝑖𝑖 𝑦𝑦𝑖𝑖 = ℎ2 𝑟𝑟𝑖𝑖
2

Rearrange the equation,

ℎ ℎ
(1 − 𝑝𝑝𝑖𝑖 )𝑦𝑦𝑖𝑖−1 + (−2 + ℎ2 𝑞𝑞𝑖𝑖 )𝑦𝑦𝑖𝑖 + (1 + 𝑝𝑝𝑖𝑖 )𝑦𝑦𝑖𝑖+1 = ℎ2 𝑟𝑟𝑖𝑖
2 2

Expanding the finite difference equation for 𝑖𝑖 = 1,2,3, … . , 𝑛𝑛

i.e:

at i=1 ℎ ℎ
(1 − 𝑝𝑝1 )𝑦𝑦0 + (−2 + ℎ2 𝑞𝑞1 )𝑦𝑦1 + (1 + 𝑝𝑝1 )𝑦𝑦2 = ℎ2 𝑟𝑟1
2 2

at i=2 ℎ ℎ
(1 − 𝑝𝑝2 )𝑦𝑦1 + (−2 + ℎ2 𝑞𝑞2 )𝑦𝑦2 + (1 + 𝑝𝑝2 )𝑦𝑦3 = ℎ2 𝑟𝑟2
2 2

at i=n ℎ ℎ
(1 − 𝑝𝑝𝑛𝑛 )𝑦𝑦𝑛𝑛−1 + (−2 + ℎ2 𝑞𝑞𝑛𝑛 )𝑦𝑦𝑛𝑛 + (1 + 𝑝𝑝𝑛𝑛 )𝑦𝑦𝑛𝑛+1 = ℎ2 𝑟𝑟𝑛𝑛
2 2
ℎ ℎ 2ℎ𝑘𝑘1 2ℎ
(1 − 𝑝𝑝𝑛𝑛 )𝑦𝑦𝑛𝑛−1 + (−2 + ℎ2 𝑞𝑞𝑛𝑛 )𝑦𝑦𝑛𝑛 + (1 + 𝑝𝑝𝑛𝑛 )(𝑦𝑦𝑛𝑛−1 − 𝑦𝑦𝑛𝑛 + 𝛽𝛽) = ℎ2 𝑟𝑟𝑛𝑛
2 2 𝑘𝑘2 𝑘𝑘2

Combine all equation in linear system of from 𝐴𝐴𝐴𝐴 = 𝑏𝑏

 CECB423 Numerical Methods for Civil Engineers

Solve the value of 𝑦𝑦1 , 𝑦𝑦2 , 𝑦𝑦3 , … … … . , 𝑦𝑦𝑛𝑛

Example:

Solve the BVP

𝑦𝑦 ′′ + 2𝑥𝑥𝑦𝑦 ′ − 3𝑦𝑦 = −6𝑒𝑒 −𝑥𝑥 (1 + 𝑥𝑥 2 ); 𝑦𝑦(0) = 0, 𝑦𝑦 ′ (1) = 0,

For x at interval [0,1] with h=0.2. Use four decimal places in all calculation.

0.2

𝑥𝑥0 = 0 𝑥𝑥1 = 0.2 𝑥𝑥2 = 0.4 𝑥𝑥3 = 0.6 𝑥𝑥4 = 0.8 𝑥𝑥5 = 1.0 𝑥𝑥6 =
𝑦𝑦0 = 0 𝑦𝑦1 = 𝑦𝑦2 = 𝑦𝑦3 = 𝑦𝑦4 = 𝑦𝑦5 = 𝑦𝑦6 =

First boundary condition at Second boundary condition at

𝑦𝑦(0) = 0 𝑦𝑦 ′ (1) = 0
𝑦𝑦(𝑥𝑥0 ) = 0 𝑦𝑦 ′ (𝑥𝑥5 ) = 0
𝑦𝑦0 = 0 𝑦𝑦6 − 𝑦𝑦4
=0
2ℎ
𝑦𝑦6 − 𝑦𝑦4 = 0
𝑦𝑦6 = 𝑦𝑦4

𝑦𝑦 ′′ + 2𝑥𝑥𝑦𝑦 ′ − 3𝑦𝑦 = −6𝑒𝑒 −𝑥𝑥 (1 + 𝑥𝑥 2 )

𝑦𝑦 ′′ (𝑥𝑥) + 𝑝𝑝(𝑥𝑥)𝑦𝑦 ′ (𝑥𝑥) + 𝑞𝑞(𝑥𝑥)𝑦𝑦(𝑥𝑥) = 𝑟𝑟(𝑥𝑥)

ℎ ℎ
(1 − 𝑝𝑝𝑖𝑖 )𝑦𝑦𝑖𝑖−1 + (−2 + ℎ2 𝑞𝑞𝑖𝑖 )𝑦𝑦𝑖𝑖 + (1 + 𝑝𝑝𝑖𝑖 )𝑦𝑦𝑖𝑖+1 = ℎ2 𝑟𝑟𝑖𝑖
2 2

at i=1 ℎ ℎ
(1 − 𝑝𝑝1 )𝑦𝑦0 + (−2 + ℎ2 𝑞𝑞1 )𝑦𝑦1 + (1 + 𝑝𝑝1 )𝑦𝑦2 = ℎ2 𝑟𝑟1
2 2
x1=0.2 ℎ ℎ
�1 − ∗ 2𝑥𝑥1 � 𝑦𝑦0 + (−2 + ℎ2 ∗ (−3))𝑦𝑦1 + �1 + ∗ 2𝑥𝑥1 � 𝑦𝑦2 = ℎ2 ∗ −6𝑒𝑒 −𝑥𝑥1 (1 + (𝑥𝑥1 )2 )
2 2
0.2 0.2
�1 − ∗ 2(0.2)� (0) + (−2 + 0.22 ∗ (−3))𝑦𝑦1 + �1 + ∗ 2(0.2)� 𝑦𝑦2 = 0.22 ∗ −6𝑒𝑒 −0.2 (1 + (0.2)2 )
2 2
 CECB423 Numerical Methods for Civil Engineers

−𝟐𝟐. 𝟏𝟏𝟏𝟏𝒚𝒚𝟏𝟏 + 𝟏𝟏. 𝟎𝟎𝟎𝟎𝒚𝒚𝟐𝟐 = −𝟎𝟎. 𝟐𝟐𝟐𝟐𝟐𝟐𝟐𝟐

at i=2 ℎ ℎ
(1 − 𝑝𝑝2 )𝑦𝑦1 + (−2 + ℎ2 𝑞𝑞2 )𝑦𝑦2 + (1 + 𝑝𝑝2 )𝑦𝑦3 = ℎ2 𝑟𝑟2
2 2
x2=0.4 ℎ ℎ
�1 − ∗ 2𝑥𝑥2 � 𝑦𝑦1 + (−2 + ℎ2 ∗ (−3))𝑦𝑦2 + �1 + ∗ 2𝑥𝑥2 � 𝑦𝑦3 = ℎ2 ∗ −6𝑒𝑒 −𝑥𝑥2 (1 + (𝑥𝑥2 )2 )
2 2
0.2 0.2
�1 − ∗ 2(0.4)� 𝑦𝑦1 + (−2 + 0.22 ∗ (−3))𝑦𝑦2 + �1 + ∗ 2(0.4)� 𝑦𝑦3 = 0.22 ∗ −6𝑒𝑒 −0.4 (1 + (0.4)2 )
2 2
𝟎𝟎. 𝟗𝟗𝟗𝟗𝒚𝒚𝟏𝟏 − 𝟐𝟐. 𝟏𝟏𝟏𝟏𝒚𝒚𝟐𝟐 + 𝟏𝟏. 𝟎𝟎𝟎𝟎𝒚𝒚𝟑𝟑 = −𝟎𝟎. 𝟏𝟏𝟏𝟏𝟏𝟏𝟏𝟏

at i=3 ℎ ℎ
(1 − 𝑝𝑝3 )𝑦𝑦2 + (−2 + ℎ2 𝑞𝑞3 )𝑦𝑦3 + (1 + 𝑝𝑝3 )𝑦𝑦4 = ℎ2 𝑟𝑟3
2 2
x3=0.6 ℎ ℎ
�1 − ∗ 2𝑥𝑥3 � 𝑦𝑦2 + (−2 + ℎ2 ∗ (−3))𝑦𝑦3 + �1 + ∗ 2𝑥𝑥3 � 𝑦𝑦4 = ℎ2 ∗ −6𝑒𝑒 −𝑥𝑥3 (1 + (𝑥𝑥3 )2 )
2 2
0.2 0.2
�1 − ∗ 2(0.6)� 𝑦𝑦2 + (−2 + 0.22 ∗ (−3))𝑦𝑦3 + �1 + ∗ 2(0.6)� 𝑦𝑦4 = 0.22 ∗ −6𝑒𝑒 −0.6 (1 + (0.6)2 )
2 2
𝟎𝟎. 𝟖𝟖𝟖𝟖𝒚𝒚𝟐𝟐 − 𝟐𝟐. 𝟏𝟏𝟏𝟏𝒚𝒚𝟑𝟑 + 𝟏𝟏. 𝟏𝟏𝟏𝟏𝒚𝒚𝟒𝟒 = −𝟎𝟎. 𝟏𝟏𝟏𝟏𝟏𝟏𝟏𝟏

at i=4 ℎ ℎ
(1 − 𝑝𝑝4 )𝑦𝑦3 + (−2 + ℎ2 𝑞𝑞4 )𝑦𝑦4 + (1 + 𝑝𝑝4 )𝑦𝑦5 = ℎ2 𝑟𝑟4
2 2
x4=0.8 ℎ ℎ
�1 − ∗ 2𝑥𝑥4 � 𝑦𝑦3 + (−2 + ℎ2 ∗ (−3))𝑦𝑦4 + �1 + ∗ 2𝑥𝑥3 � 𝑦𝑦5 = ℎ2 ∗ −6𝑒𝑒 −𝑥𝑥3 (1 + (𝑥𝑥4 )2 )
2 2
0.2 0.2
�1 − ∗ 2(0.8)� 𝑦𝑦3 + (−2 + 0.22 ∗ (−3))𝑦𝑦4 + �1 + ∗ 2(0.8)� 𝑦𝑦5 = 0.22 ∗ −6𝑒𝑒 −0.8 (1 + (0.8)2 )
2 2
𝟎𝟎. 𝟖𝟖𝟖𝟖𝒚𝒚𝟐𝟐 − 𝟐𝟐. 𝟏𝟏𝟏𝟏𝒚𝒚𝟑𝟑 + 𝟏𝟏. 𝟏𝟏𝟏𝟏𝒚𝒚𝟒𝟒 = −𝟎𝟎. 𝟏𝟏𝟏𝟏𝟏𝟏𝟏𝟏

at i=5 ℎ ℎ
(1 − 𝑝𝑝5 )𝑦𝑦4 + (−2 + ℎ2 𝑞𝑞5 )𝑦𝑦5 + (1 + 𝑝𝑝5 )𝑦𝑦6 = ℎ2 𝑟𝑟5
2 2
x5=1.0 ℎ ℎ
�1 − ∗ 2𝑥𝑥5 � 𝑦𝑦4 + (−2 + ℎ2 ∗ (−3))𝑦𝑦5 + �1 + ∗ 2𝑥𝑥5 � 𝑦𝑦6 = ℎ2 ∗ −6𝑒𝑒 −𝑥𝑥5 (1 + (𝑥𝑥5 )2 )
2 2
0.2 0.2
�1 − ∗ 2(1.0)� 𝑦𝑦4 + (−2 + 0.22 ∗ (−3))𝑦𝑦5 + �1 + ∗ 2(1.0)� 𝑦𝑦6 = 0.22 ∗ −6𝑒𝑒 −1 (1 + (1.0)2 )
2 2
𝟎𝟎. 𝟖𝟖𝟖𝟖𝒚𝒚𝟒𝟒 − 𝟐𝟐. 𝟏𝟏𝟏𝟏𝒚𝒚𝟓𝟓 + 𝟏𝟏. 𝟐𝟐𝒚𝒚𝟔𝟔 = −𝟎𝟎. 𝟏𝟏𝟏𝟏𝟏𝟏𝟏𝟏
Based on second derivative
0.80𝑦𝑦4 − 2.12𝑦𝑦5 + 1.2𝑦𝑦4 = −0.1766
𝟐𝟐. 𝟎𝟎𝟎𝟎𝒚𝒚𝟒𝟒 − 𝟐𝟐. 𝟏𝟏𝟏𝟏𝒚𝒚𝟓𝟓 = −𝟎𝟎. 𝟏𝟏𝟏𝟏𝟏𝟏𝟏𝟏
 CECB423 Numerical Methods for Civil Engineers

Solve the system using trigonal system

−2.12 1.04 0 0 0 𝑦𝑦1 −0.2044

0.92 −2.12 1.04 0 0 𝑦𝑦2
⎛ ⎞ ⎛ ⎞ ⎛−0.1866⎞
⎜ 0 0.88 −2.12 1.12 0 ⎟ ⎜𝑦𝑦3 ⎟ = ⎜−0.1791⎟
0 0 0.84 −2.12 1.16 𝑦𝑦4 −0.1769
⎝ 0 0 0 𝑦𝑦
2.00 −2.12⎠ ⎝ 5 ⎠ ⎝−0.1766⎠

𝑦𝑦1 0.4897
𝑦𝑦2
⎛ ⎞ ⎛0.8017⎞
⎜𝑦𝑦3 ⎟ = ⎜0.9838⎟
𝑦𝑦4 1.0724
𝑦𝑦
⎝ 5 ⎠ ⎝1.0950⎠

Example BVP both derivative boundary condition

Solve the BVP

𝑦𝑦 ′′ + 𝑥𝑥𝑦𝑦 ′ − 𝑥𝑥 2 𝑦𝑦(𝑥𝑥) = 2𝑥𝑥 3 ; 𝑦𝑦′(0) = 1, 𝑦𝑦′(1) = −1

For x at interval [0,1] with h=0.2. Use four decimal places in all calculation.

0.2

𝑥𝑥−1 = 𝑥𝑥0 = 0 𝑥𝑥1 = 0.2 𝑥𝑥2 = 0.4 𝑥𝑥3 = 0.6 𝑥𝑥4 = 0.8 𝑥𝑥5 = 1 𝑥𝑥6 =
𝑦𝑦−1 = 𝑦𝑦0 = 𝑦𝑦1 = 𝑦𝑦2 = 𝑦𝑦3 = 𝑦𝑦4 = 𝑦𝑦5 = 𝑦𝑦6 =

First boundary condition at Second boundary condition at

𝑦𝑦 ′ (0) = 1, 𝑦𝑦′(1) = −1
𝑦𝑦′0 = 1 𝑦𝑦5′ = −1
𝑦𝑦1 − 𝑦𝑦−1 𝑦𝑦6 − 𝑦𝑦4
=1 = −1
2ℎ 2ℎ
𝑦𝑦1 − 𝑦𝑦−1 = 2(0.2) 𝑦𝑦6 − 𝑦𝑦4 = −2(0.2)
𝑦𝑦−1 = 𝑦𝑦1 − 0.4 𝑦𝑦6 = −0.4 + 𝑦𝑦4

𝑦𝑦 ′′ + 𝑥𝑥𝑦𝑦 ′ − 𝑥𝑥 2 𝑦𝑦(𝑥𝑥) = 2𝑥𝑥 3

 CECB423 Numerical Methods for Civil Engineers

𝑦𝑦 ′′ (𝑥𝑥) + 𝑝𝑝(𝑥𝑥)𝑦𝑦 ′ (𝑥𝑥) + 𝑞𝑞(𝑥𝑥)𝑦𝑦(𝑥𝑥) = 𝑟𝑟(𝑥𝑥)

𝑦𝑦𝑖𝑖+1 − 2𝑦𝑦𝑖𝑖 + 𝑦𝑦𝑖𝑖−1 𝑦𝑦𝑖𝑖+1 − 𝑦𝑦𝑖𝑖−1

� 2 � + 𝑥𝑥𝑖𝑖 � � − 𝑥𝑥𝑖𝑖 2 𝑦𝑦𝑖𝑖 = 2𝑥𝑥𝑖𝑖 3
ℎ 2ℎ

ℎ
[𝑦𝑦𝑖𝑖+1 − 2𝑦𝑦𝑖𝑖 + 𝑦𝑦𝑖𝑖−1 ] + 𝑥𝑥𝑖𝑖 [𝑦𝑦𝑖𝑖+1 − 𝑦𝑦𝑖𝑖−1 ] − ℎ2 𝑥𝑥𝑖𝑖 2 𝑦𝑦𝑖𝑖 = ℎ2 2𝑥𝑥𝑖𝑖 3
2

ℎ ℎ
𝑦𝑦𝑖𝑖+1 − 2𝑦𝑦𝑖𝑖 + 𝑦𝑦𝑖𝑖−1 + 𝑥𝑥𝑖𝑖 𝑦𝑦𝑖𝑖+1 − 𝑥𝑥𝑖𝑖 𝑦𝑦𝑖𝑖−1 − ℎ2 𝑥𝑥𝑖𝑖 2 𝑦𝑦𝑖𝑖 = ℎ2 2𝑥𝑥𝑖𝑖 3
2 2

ℎ ℎ
(1 − 𝑥𝑥𝑖𝑖 )𝑦𝑦𝑖𝑖−1 + (−2 − ℎ2 𝑥𝑥𝑖𝑖 2 )𝑦𝑦𝑖𝑖 + �1 + 𝑥𝑥𝑖𝑖 � 𝑦𝑦𝑖𝑖+1 = ℎ2 . 2𝑥𝑥𝑖𝑖3
2 2

0.2 0.2
(1 − 𝑥𝑥𝑖𝑖 )𝑦𝑦𝑖𝑖−1 + (−2 − (0.2)2 𝑥𝑥𝑖𝑖 2 )𝑦𝑦𝑖𝑖 + �1 + 𝑥𝑥 � 𝑦𝑦 = (0.2)2 . 2𝑥𝑥𝑖𝑖3
2 2 𝑖𝑖 𝑖𝑖+1

(1 − 0.1𝑥𝑥𝑖𝑖 )𝑦𝑦𝑖𝑖−1 + (−2 − 0.04𝑥𝑥𝑖𝑖 2 )𝑦𝑦𝑖𝑖 + (1 + 0.1𝑥𝑥𝑖𝑖 )𝑦𝑦𝑖𝑖+1 = 0.08𝑥𝑥𝑖𝑖3

at i=0 (1 − 0.1𝑥𝑥0 )𝑦𝑦−1 + (−2 − 0.04𝑥𝑥0 2 )𝑦𝑦0 + (1 + 0.1𝑥𝑥0 )𝑦𝑦1 = 0.08𝑥𝑥03

x0=0 �1 − 0.1(0)�𝑦𝑦−1 + (−2 − 0.04(0)2 )𝑦𝑦0 + �1 + 0.1(0)�𝑦𝑦1 = 0.08(0)3
𝑦𝑦−1 − 2𝑦𝑦0 + 𝑦𝑦1 = 0
𝑦𝑦1 − 0.4 − 2𝑦𝑦0 + 𝑦𝑦1 = 0
−𝟐𝟐𝒚𝒚𝟎𝟎 + 𝟐𝟐𝒚𝒚𝟏𝟏 = 𝟎𝟎. 𝟒𝟒
at i=1 (1 − 0.1𝑥𝑥1 )𝑦𝑦0 + (−2 − 0.04𝑥𝑥1 2 )𝑦𝑦1 + (1 + 0.1𝑥𝑥1 )𝑦𝑦2 = 0.08𝑥𝑥13
x1=0.2 �1 − 0.1(0.2)�𝑦𝑦0 + (−2 − 0.04(0.2)2 )𝑦𝑦1 + �1 + 0.1(0.2)�𝑦𝑦2 = 0.08(0.2)3
𝟎𝟎. 𝟗𝟗𝟗𝟗𝒚𝒚𝟎𝟎 − 𝟐𝟐. 𝟎𝟎𝟎𝟎𝟎𝟎𝟎𝟎𝒚𝒚𝟏𝟏 + 𝟏𝟏. 𝟎𝟎𝟎𝟎𝒚𝒚𝟐𝟐 = 𝟎𝟎. 𝟎𝟎𝟎𝟎𝟎𝟎𝟎𝟎

at i=2 (1 − 0.1𝑥𝑥2 )𝑦𝑦1 + (−2 − 0.04𝑥𝑥2 2 )𝑦𝑦2 + (1 + 0.1𝑥𝑥2 )𝑦𝑦3 = 0.08𝑥𝑥23

x2=0.4 �1 − 0.1(0.4)�𝑦𝑦1 + (−2 − 0.04(0.4)2 )𝑦𝑦2 + �1 + 0.1(0.4)�𝑦𝑦3 = 0.08(0.4)3
𝟎𝟎. 𝟗𝟗𝟗𝟗𝒚𝒚𝟏𝟏 − 𝟐𝟐. 𝟎𝟎𝟎𝟎𝟎𝟎𝟎𝟎𝒚𝒚𝟐𝟐 + 𝟏𝟏. 𝟎𝟎𝟎𝟎𝒚𝒚𝟑𝟑 = 𝟎𝟎. 𝟎𝟎𝟎𝟎𝟎𝟎𝟎𝟎

at i=3 (1 − 0.1𝑥𝑥3 )𝑦𝑦2 + (−2 − 0.04𝑥𝑥3 2 )𝑦𝑦3 + (1 + 0.1𝑥𝑥3 )𝑦𝑦4 = 0.08𝑥𝑥33

x3=0.6 �1 − 0.1(0.6)�𝑦𝑦2 + (−2 − 0.04(0.6)2 )𝑦𝑦3 + �1 + 0.1(0.6)�𝑦𝑦4 = 0.08(0.6)3
𝟎𝟎. 𝟗𝟗𝟗𝟗𝒚𝒚𝟐𝟐 − 𝟐𝟐. 𝟎𝟎𝟎𝟎𝟎𝟎𝟎𝟎𝒚𝒚𝟑𝟑 + 𝟏𝟏. 𝟎𝟎𝟎𝟎𝒚𝒚𝟒𝟒 = 𝟎𝟎. 𝟎𝟎𝟎𝟎𝟎𝟎𝟎𝟎
 CECB423 Numerical Methods for Civil Engineers

at i=4 (1 − 0.1𝑥𝑥4 )𝑦𝑦3 + (−2 − 0.04𝑥𝑥4 2 )𝑦𝑦4 + (1 + 0.1𝑥𝑥4 )𝑦𝑦5 = 0.08𝑥𝑥43

x4=0.8 �1 − 0.1(0.8)�𝑦𝑦3 + (−2 − 0.04(0.8)2 )𝑦𝑦4 + �1 + 0.1(0.8)�𝑦𝑦5 = 0.08(0.8)3
𝟎𝟎. 𝟗𝟗𝟗𝟗𝒚𝒚𝟑𝟑 − 𝟐𝟐. 𝟎𝟎𝟎𝟎𝟎𝟎𝟎𝟎𝒚𝒚𝟒𝟒 + 𝟏𝟏. 𝟎𝟎𝟎𝟎𝒚𝒚𝟓𝟓 = 𝟎𝟎. 𝟎𝟎𝟎𝟎𝟎𝟎𝟎𝟎

at i=5 (1 − 0.1𝑥𝑥5 )𝑦𝑦4 + (−2 − 0.04𝑥𝑥5 2 )𝑦𝑦5 + (1 + 0.1𝑥𝑥5 )𝑦𝑦6 = 0.08𝑥𝑥53

x5=1.0 �1 − 0.1(1.0)�𝑦𝑦4 + (−2 − 0.04(1.0)2 )𝑦𝑦5 + �1 + 0.1(1.0)�𝑦𝑦6 = 0.08(1.0)3
0.9𝑦𝑦4 − 2.04𝑦𝑦5 + 1.1𝑦𝑦6 = 0.08
0.9𝑦𝑦4 − 2.04𝑦𝑦5 + 1.1(−0.4 + 𝑦𝑦4 ) = 0.08
0.9𝑦𝑦4 − 2.04𝑦𝑦5 − 0.44 + 1.1𝑦𝑦4 = 0.08
𝟐𝟐. 𝟎𝟎𝒚𝒚𝟒𝟒 − 𝟐𝟐. 𝟎𝟎𝟎𝟎𝒚𝒚𝟓𝟓 = 𝟎𝟎. 𝟓𝟓𝟓𝟓

−𝟐𝟐𝒚𝒚𝟎𝟎 + 𝟐𝟐𝒚𝒚𝟏𝟏 = 𝟎𝟎. 𝟒𝟒

𝟎𝟎. 𝟗𝟗𝟗𝟗𝒚𝒚𝟎𝟎 − 𝟐𝟐. 𝟎𝟎𝟎𝟎𝟎𝟎𝟎𝟎𝒚𝒚𝟏𝟏 + 𝟏𝟏. 𝟎𝟎𝟎𝟎𝒚𝒚𝟐𝟐 = 𝟎𝟎. 𝟎𝟎𝟎𝟎𝟎𝟎𝟎𝟎
𝟎𝟎. 𝟗𝟗𝟗𝟗𝒚𝒚𝟏𝟏 − 𝟐𝟐. 𝟎𝟎𝟎𝟎𝟎𝟎𝟎𝟎𝒚𝒚𝟐𝟐 + 𝟏𝟏. 𝟎𝟎𝟎𝟎𝒚𝒚𝟑𝟑 = 𝟎𝟎. 𝟎𝟎𝟎𝟎𝟎𝟎𝟎𝟎
𝟎𝟎. 𝟗𝟗𝟗𝟗𝒚𝒚𝟐𝟐 − 𝟐𝟐. 𝟎𝟎𝟎𝟎𝟎𝟎𝟎𝟎𝒚𝒚𝟑𝟑 + 𝟏𝟏. 𝟎𝟎𝟎𝟎𝒚𝒚𝟒𝟒 = 𝟎𝟎. 𝟎𝟎𝟎𝟎𝟎𝟎𝟎𝟎
𝟎𝟎. 𝟗𝟗𝟗𝟗𝒚𝒚𝟑𝟑 − 𝟐𝟐. 𝟎𝟎𝟎𝟎𝟎𝟎𝟎𝟎𝒚𝒚𝟒𝟒 + 𝟏𝟏. 𝟎𝟎𝟎𝟎𝒚𝒚𝟓𝟓 = 𝟎𝟎. 𝟎𝟎𝟎𝟎𝟎𝟎𝟎𝟎
𝟐𝟐. 𝟎𝟎𝒚𝒚𝟒𝟒 − 𝟐𝟐. 𝟎𝟎𝟎𝟎𝒚𝒚𝟓𝟓 = 𝟎𝟎. 𝟓𝟓𝟓𝟓
Solve the system using trigonal system

𝑦𝑦0 −7.8494
𝑦𝑦1 −7.6494
⎛𝑦𝑦2 ⎞ ⎛ ⎞
= −7.4687
⎜
⎜𝑦𝑦3 ⎟ ⎜−7.3429⎟
⎟ ⎜ ⎟
𝑦𝑦4 −7.3148
⎝𝑦𝑦5 ⎠ ⎝−7.4262⎠
 CECB423 Numerical Methods for Civil Engineers

Example on simply supported beam

Figure 1 show the simply supported beam with a uniform load (q). Given the equation of
deflection (y) is
𝑦𝑦 ′′ (𝑥𝑥) = 𝑟𝑟(𝑥𝑥)

𝑑𝑑 2 𝑦𝑦 𝑞𝑞𝑥𝑥(𝐿𝐿 − 𝑥𝑥)2
=
𝑑𝑑𝑥𝑥 2 2𝐸𝐸𝐸𝐸
2
𝑦𝑦𝑖𝑖+1 − 2𝑦𝑦𝑖𝑖 + 𝑦𝑦𝑖𝑖−1 (10)𝑥𝑥(10 − 𝑥𝑥)
=
ℎ2 2(30)(120)

where
x = location along the beam
E = Young’s modulus of elasticity of the beam = 30 kN/𝑚𝑚2
I = second moment of area = 120 m4
q = uniform loading intensity = 10𝑘𝑘𝑘𝑘
L = length of beam = 10 m

Determine the maximum deflection of the beam using central difference approximation with step
size of ∆x = 2.5.

2
𝑦𝑦𝑖𝑖+1 − 2𝑦𝑦𝑖𝑖 + 𝑦𝑦𝑖𝑖−1 (10)𝑥𝑥(10 − 𝑥𝑥)
=
ℎ2 2(30)(120)

𝑥𝑥0 = 0 𝑥𝑥1 = 2.5 𝑥𝑥2 = 5.0 𝑥𝑥3 = 7.5 𝑥𝑥4 = 10

𝑦𝑦0 = 0 𝑦𝑦1 = 𝑦𝑦2 = 𝑦𝑦3 = 𝑦𝑦4 = 0

 CECB423 Numerical Methods for Civil Engineers

𝑑𝑑2 𝑦𝑦 𝑞𝑞𝑥𝑥(𝐿𝐿 − 𝑥𝑥)2

=
𝑑𝑑𝑥𝑥 2 2𝐸𝐸𝐸𝐸

𝑦𝑦𝑖𝑖+1 − 2𝑦𝑦𝑖𝑖 + 𝑦𝑦𝑖𝑖−1 10𝑥𝑥(10 − 𝑥𝑥)2

=
ℎ2 2(30)(120)

𝑦𝑦𝑖𝑖+1 − 2𝑦𝑦𝑖𝑖 + 𝑦𝑦𝑖𝑖−1 10𝑥𝑥(10 − 𝑥𝑥)2

=
(2.5)2 7200
At i=1,X1=2.5
𝑦𝑦2 − 2𝑦𝑦1 + 𝑦𝑦0 (10)(2.5)(10 − 2.5)2
=
(2.5)2 7200
𝑦𝑦2 − 2𝑦𝑦1 − 0 = 1.2207
𝒚𝒚𝟐𝟐 − 𝟐𝟐𝒚𝒚𝟏𝟏 = 𝟏𝟏. 𝟐𝟐𝟐𝟐𝟐𝟐𝟐𝟐
At i=2, X2=5.0
𝑦𝑦3 − 2𝑦𝑦2 + 𝑦𝑦1 (10)(5)(10 − 5.0)2
=
(2.5)2 7200
𝒚𝒚𝟑𝟑 − 𝟐𝟐𝒚𝒚𝟐𝟐 + 𝒚𝒚𝟏𝟏 = 𝟏𝟏. 𝟎𝟎𝟎𝟎𝟎𝟎𝟎𝟎
At i=3, X3=7.5
𝑦𝑦4 − 2𝑦𝑦3 + 𝑦𝑦2 (10)(7.5)(10 − 7.5)2
=
(2.5)2 7200
0 − 2𝑦𝑦3 + 𝑦𝑦2 = 0.4069
−𝟐𝟐𝒚𝒚𝟑𝟑 + 𝒚𝒚𝟐𝟐 = 𝟎𝟎. 𝟒𝟒𝟒𝟒𝟒𝟒𝟒𝟒

−2 1 0 𝑦𝑦1 1.2207
� 1 −2 1 � �𝑦𝑦2 � = �1.0851�
0 1 −2 𝑦𝑦3 0.4069
𝑦𝑦1 −1.5598
�𝑦𝑦2 � = �−1.8989�
𝑦𝑦3 −1.1529

The maximum deflection is 1.8989m

 CECB423 Numerical Methods for Civil Engineers

Partial Differential Equations (PDE).

Partial differential equation (PDE) are differential equations involving more than one
independent variable. We consider a general second-order PDE in two independent
variable x and y which is written as

𝛿𝛿 2 𝑢𝑢 𝛿𝛿 2 𝑢𝑢 𝛿𝛿 2 𝑢𝑢 𝛿𝛿𝛿𝛿 𝛿𝛿𝛿𝛿
𝐴𝐴 2
+ 𝐵𝐵 + 𝐶𝐶 2
+ 𝐷𝐷 + 𝐸𝐸 + 𝐹𝐹𝐹𝐹 = 𝐺𝐺
𝛿𝛿𝑥𝑥 𝛿𝛿𝛿𝛿𝛿𝛿𝛿𝛿 𝛿𝛿𝑦𝑦 𝛿𝛿𝛿𝛿 𝛿𝛿𝛿𝛿
Or
𝐴𝐴𝑢𝑢𝑥𝑥𝑥𝑥 + 𝐵𝐵𝑢𝑢𝑥𝑥𝑥𝑥 + 𝐶𝐶𝑢𝑢𝑦𝑦𝑦𝑦 + 𝐷𝐷𝑢𝑢𝑥𝑥 + 𝐸𝐸𝑢𝑢𝑦𝑦 + 𝐹𝐹𝐹𝐹 = 𝐺𝐺

Where A, B, C, D, E, F and G are constants or functions of x and y. These PDEs are

classified into three group.

Elliptic <0
2
PDE is called Parabolic if 𝐵𝐵 − 4𝐴𝐴𝐴𝐴 =0
Hyperbolic >0

Determine either the equation is elliptic, parabolic or hyperbolic.

𝛿𝛿 2 𝑢𝑢 𝛿𝛿 2 𝑢𝑢 A:1, C:1, G;0(0)2-(4*1*1)=-4 <0 [elliptic]

1. Laplace equation, 𝛿𝛿𝑥𝑥 2 + 𝛿𝛿𝑦𝑦 2 = 0

𝛿𝛿 2 𝑢𝑢 𝛿𝛿 2 𝑢𝑢
2. Poisson equation 𝛿𝛿𝑥𝑥 2 + 𝛿𝛿𝑦𝑦 2 = 𝑓𝑓(𝑥𝑥, 𝑦𝑦) A:1, C:1,G:f(x,y)(0)2-(4*1*1)=-4 <0 [elliptic]

𝛿𝛿 2 𝑢𝑢 𝛿𝛿𝛿𝛿 A:C; E:-1(0)2-(4*C*0)=0 [Parabolic]

3. Heat equation 𝑐𝑐 𝛿𝛿𝑥𝑥 2 = 𝛿𝛿𝛿𝛿

𝛿𝛿 2 𝑢𝑢 𝛿𝛿 2 𝑢𝑢
4. Wave equation 𝑐𝑐 𝛿𝛿𝑥𝑥 2 = 𝛿𝛿𝑡𝑡 2
A:C; C:-1(0)2-(4*C*-1)=4C >0 [Hyperbolic]

Formula partial differential equation using central approximation:

𝛿𝛿 2 𝑢𝑢 𝑢𝑢𝑖𝑖+1,𝑗𝑗 − 2𝑢𝑢𝑖𝑖,𝑗𝑗 + 𝑢𝑢𝑖𝑖−1,𝑗𝑗 𝛿𝛿𝛿𝛿 𝑢𝑢𝑖𝑖+1,𝑗𝑗 − 𝑢𝑢𝑖𝑖−1,𝑗𝑗

= =
𝛿𝛿𝑥𝑥 2 ℎ2 𝛿𝛿𝛿𝛿 2ℎ

𝛿𝛿 2 𝑢𝑢 𝑢𝑢𝑖𝑖,𝑗𝑗+1 − 2𝑢𝑢𝑖𝑖,𝑗𝑗 + 𝑢𝑢𝑖𝑖,𝑗𝑗−1 𝛿𝛿𝛿𝛿 𝑢𝑢𝑖𝑖,𝑗𝑗+1 − 𝑢𝑢𝑖𝑖,𝑗𝑗−1

= =
𝛿𝛿𝑦𝑦 2 𝑘𝑘 2 𝛿𝛿𝛿𝛿 2𝑘𝑘
 CECB423 Numerical Methods for Civil Engineers

Elliptic Equations
Finite difference method for the numerical solution of the PDEs of elliptic type.
Consider the Poisson equation,
𝛿𝛿 2 𝑢𝑢 𝛿𝛿 2 𝑢𝑢
+ = 𝑓𝑓(𝑥𝑥, 𝑦𝑦)
𝛿𝛿𝑥𝑥 2 𝛿𝛿𝑦𝑦 2
With boundary condition
𝑢𝑢(𝑥𝑥, 0) = 𝑓𝑓1 (𝑥𝑥), 𝑢𝑢(𝑥𝑥, 𝑏𝑏) = 𝑓𝑓2 (𝑥𝑥), 0 < 𝑥𝑥 < 𝑎𝑎
𝑢𝑢(0, 𝑦𝑦) = 𝑔𝑔1 (𝑥𝑥), 𝑢𝑢(𝑎𝑎, 𝑦𝑦) = 𝑔𝑔2 (𝑥𝑥), 0 < 𝑦𝑦 < 𝑏𝑏

Create Domain

In order to approximate the solution to PDE elliptic using finite difference method the
𝑎𝑎 𝑏𝑏
rectangular domain is partitioned into a grids ∆𝑥𝑥 = ℎ = 𝑛𝑛 and ∆𝑦𝑦 = 𝑘𝑘 = 𝑚𝑚 as shown in
figure below

𝛿𝛿 2 𝑢𝑢 𝛿𝛿 2 𝑢𝑢
+ = 𝑓𝑓(𝑥𝑥, 𝑦𝑦)
𝛿𝛿𝑥𝑥 2 𝛿𝛿𝑦𝑦 2
At point (𝑥𝑥𝑖𝑖 , 𝑦𝑦𝑗𝑗 )

𝛿𝛿 2 𝑢𝑢(𝑥𝑥𝑖𝑖 , 𝑦𝑦𝑗𝑗 ) 𝛿𝛿 2 𝑢𝑢(𝑥𝑥𝑖𝑖 , 𝑦𝑦𝑗𝑗 )

+ = 𝑓𝑓((𝑥𝑥𝑖𝑖 , 𝑦𝑦𝑗𝑗 )
𝛿𝛿𝑥𝑥 2 𝛿𝛿𝑦𝑦 2
 CECB423 Numerical Methods for Civil Engineers

using finite difference approximation

𝑢𝑢𝑖𝑖+1,𝑗𝑗 − 2𝑢𝑢𝑖𝑖,𝑗𝑗 + 𝑢𝑢𝑖𝑖−1,𝑗𝑗 𝑢𝑢𝑖𝑖,𝑗𝑗+1 − 2𝑢𝑢𝑖𝑖,𝑗𝑗 + 𝑢𝑢𝑖𝑖,𝑗𝑗−1

+ = 𝑓𝑓𝑖𝑖,𝑗𝑗
ℎ2 𝑘𝑘 2

multiply both side ℎ2

ℎ2
𝑢𝑢𝑖𝑖+1,𝑗𝑗 − 2𝑢𝑢𝑖𝑖,𝑗𝑗 + 𝑢𝑢𝑖𝑖−1,𝑗𝑗 + 2 �𝑢𝑢𝑖𝑖,𝑗𝑗+1 − 2𝑢𝑢𝑖𝑖,𝑗𝑗 + 𝑢𝑢𝑖𝑖,𝑗𝑗−1 � = ℎ2 𝑓𝑓𝑖𝑖,𝑗𝑗
𝑘𝑘

Create equation and solve using Ax=b system

 CECB423 Numerical Methods for Civil Engineers

Example 1:
𝛿𝛿 2 𝑢𝑢 𝛿𝛿 2 𝑢𝑢
+ = 𝑥𝑥𝑒𝑒 𝑦𝑦 0 < 𝑥𝑥 < 2 ; 0 < 𝑦𝑦 < 1
𝛿𝛿𝑥𝑥 2 𝛿𝛿𝑦𝑦 2
With boundary conditions

𝑢𝑢(0, 𝑦𝑦) = 0, 𝑢𝑢(2, 𝑦𝑦) = 2𝑒𝑒 𝑦𝑦 0 < 𝑦𝑦 < 1

𝑢𝑢(𝑥𝑥, 0) = 𝑥𝑥, 𝑢𝑢(𝑥𝑥, 1) = 𝑥𝑥𝑒𝑒 𝑦𝑦 0 < 𝑥𝑥 < 2

Using finite difference method with ∆𝑥𝑥 = ℎ = 0.5 and ∆𝑦𝑦 = 𝑘𝑘 = 0.5
Solve the linear system 𝐴𝐴𝐴𝐴 = 𝑏𝑏 using Gauss Jordan method. Use three decimal places in
all calculation.

j=2 1.0

𝑢𝑢(𝑥𝑥, 1) = 𝑥𝑥𝑒𝑒 𝑦𝑦

j=1 0.5 𝑢𝑢(0, 𝑦𝑦) = 0 𝑢𝑢(2, 𝑦𝑦) = 2𝑒𝑒 𝑦𝑦

𝑢𝑢(𝑥𝑥, 0) = 𝑥𝑥

j=0 0.0 x

0.0 0.5 1.0 1.5 2.0

i=0 i=1 i=2 i=3 i=4

𝑢𝑢(0, 𝑦𝑦) = 0 𝑢𝑢(2, 𝑦𝑦) = 2𝑒𝑒 𝑦𝑦 𝑢𝑢(𝑥𝑥, 0) = 𝑥𝑥 𝑢𝑢(𝑥𝑥, 1) = 𝑥𝑥𝑒𝑒 𝑦𝑦

𝑢𝑢(0,0) = 0 𝑢𝑢(2,0) = 2𝑒𝑒 0 = 2 𝑢𝑢(0,0) = 0 𝑢𝑢(0,1) = (0)𝑒𝑒 1 = 0

𝑢𝑢(0,0.5) = 0 𝑢𝑢(2,0.5) = 2𝑒𝑒 0.5 = 3.2974 𝑢𝑢(0.5,0) = 0.5 𝑢𝑢(0.5,1) = (0.5)𝑒𝑒 1 = 1.3591
𝑢𝑢(0,1.0) = 0 𝑢𝑢(2,1.0) = 2𝑒𝑒 1 = 5.4366 𝑢𝑢(1.0,0) = 1.0 𝑢𝑢(1.0,1) = (1)𝑒𝑒 1 = 2.7183
𝑢𝑢(1.5,0) = 1.5 𝑢𝑢(1.5,1) = (1.5)𝑒𝑒 1 = 4.0774
𝑢𝑢(2.0,0) = 2.0 𝑢𝑢(2.0,1) = (2)𝑒𝑒 1 = 5.4366
 CECB423 Numerical Methods for Civil Engineers

U0,2=0 U1,2=1.3591 U2,2=2.7183 U3,2=4.0774 U4,2=5.4366

U0,1=0 U1,1 U2,1 U3,1 U4,1=3.2974

x
U0,0=0 U1,0=0.5 U2,0=1.0 U3,0=1.5 U4,0=2.0

Approximate 𝑢𝑢1,1 , 𝑢𝑢2,1 𝑎𝑎𝑎𝑎𝑎𝑎 𝑢𝑢3,1

At point (𝑥𝑥𝑖𝑖 , 𝑦𝑦𝑗𝑗 )

𝛿𝛿 2 𝑢𝑢(𝑥𝑥𝑖𝑖 , 𝑦𝑦𝑗𝑗 ) 𝛿𝛿 2 𝑢𝑢(𝑥𝑥𝑖𝑖 , 𝑦𝑦𝑗𝑗 )

+ = 𝑥𝑥𝑖𝑖 𝑒𝑒 𝑦𝑦𝑖𝑖
𝛿𝛿𝑥𝑥 2 𝛿𝛿𝑦𝑦 2

using finite difference approximation

𝑢𝑢𝑖𝑖+1,𝑗𝑗 − 2𝑢𝑢𝑖𝑖,𝑗𝑗 + 𝑢𝑢𝑖𝑖−1,𝑗𝑗 𝑢𝑢𝑖𝑖,𝑗𝑗+1 − 2𝑢𝑢𝑖𝑖,𝑗𝑗 + 𝑢𝑢𝑖𝑖,𝑗𝑗−1

+ = 𝑥𝑥𝑖𝑖 𝑒𝑒 𝑦𝑦𝑖𝑖
ℎ2 𝑘𝑘 2

multiply both side ℎ2

ℎ2
𝑢𝑢𝑖𝑖+1,𝑗𝑗 − 2𝑢𝑢𝑖𝑖,𝑗𝑗 + 𝑢𝑢𝑖𝑖−1,𝑗𝑗 + �𝑢𝑢 − 2𝑢𝑢𝑖𝑖,𝑗𝑗 + 𝑢𝑢𝑖𝑖,𝑗𝑗−1 � = ℎ2 𝑥𝑥𝑖𝑖 𝑒𝑒 𝑦𝑦𝑖𝑖
𝑘𝑘 2 𝑖𝑖,𝑗𝑗+1

Given ℎ = 𝑘𝑘 = 0.5

0.52
𝑢𝑢𝑖𝑖+1,𝑗𝑗 − 2𝑢𝑢𝑖𝑖,𝑗𝑗 + 𝑢𝑢𝑖𝑖−1,𝑗𝑗 + �𝑢𝑢 − 2𝑢𝑢𝑖𝑖,𝑗𝑗 + 𝑢𝑢𝑖𝑖,𝑗𝑗−1 � = 0.52 𝑥𝑥𝑖𝑖 𝑒𝑒 𝑦𝑦𝑖𝑖
0.52 𝑖𝑖,𝑗𝑗+1

𝑢𝑢𝑖𝑖+1,𝑗𝑗 − 2𝑢𝑢𝑖𝑖,𝑗𝑗 + 𝑢𝑢𝑖𝑖−1,𝑗𝑗 + 𝑢𝑢𝑖𝑖,𝑗𝑗+1 − 2𝑢𝑢𝑖𝑖,𝑗𝑗 + 𝑢𝑢𝑖𝑖,𝑗𝑗−1 = 0.52 𝑥𝑥𝑖𝑖 𝑒𝑒 𝑦𝑦𝑖𝑖

𝑢𝑢𝑖𝑖+1,𝑗𝑗 + 𝑢𝑢𝑖𝑖−1,𝑗𝑗 − 4𝑢𝑢𝑖𝑖,𝑗𝑗 + 𝑢𝑢𝑖𝑖,𝑗𝑗+1 + 𝑢𝑢𝑖𝑖,𝑗𝑗−1 = 0.52 𝑥𝑥𝑖𝑖 𝑒𝑒 𝑦𝑦𝑖𝑖

 CECB423 Numerical Methods for Civil Engineers

U1,1 𝑢𝑢2,1 + 𝑢𝑢0,1 − 4𝑢𝑢1,1 + 𝑢𝑢1,2 + 𝑢𝑢1,0 = 0.52 𝑥𝑥1 𝑒𝑒 𝑦𝑦1

att i=1, j=1 𝑢𝑢2,1 + 0 − 4𝑢𝑢1,1 + 1.359 + 0.5 = 0.52 (0.5)𝑒𝑒 0.5
𝑢𝑢2,1 − 4𝑢𝑢1,1 = −1.653

U 2,1 𝑢𝑢3,1 + 𝑢𝑢1,1 − 4𝑢𝑢2,1 + 𝑢𝑢2,2 + 𝑢𝑢2,0 = 0.52 𝑥𝑥2 𝑒𝑒 𝑦𝑦1

at i=2, j=1 𝑢𝑢3,1 + 𝑢𝑢1,1 − 4𝑢𝑢2,1 + 2.718 + 1 = 0.52 (1)𝑒𝑒 0.5
𝑢𝑢3,1 + 𝑢𝑢1,1 − 4𝑢𝑢2,1 = −3.306

U 3,1 𝑢𝑢4,1 + 𝑢𝑢2,1 − 4𝑢𝑢3,1 + 𝑢𝑢3,2 + 𝑢𝑢3,0 = 0.52 𝑥𝑥3 𝑒𝑒 𝑦𝑦1

at i=3, j=1 3.297 + 𝑢𝑢2,1 − 4𝑢𝑢3,1 + 4.077 + 1.5 = 0.52 𝑥𝑥3 𝑒𝑒 𝑦𝑦1
𝑢𝑢2,1 − 4𝑢𝑢3,1 = −8.256

Equation (1) till (3) can be express in matrix form

−4 1 0 𝑢𝑢1,1 −1.653
� 1 −4 1 � �𝑢𝑢2,1 � = �−3.306�
0 1 −4 𝑢𝑢3,1 −8.256

𝑢𝑢1,1 0.8263
𝑢𝑢
� 2,1 � = �1.6524�
𝑢𝑢3,1 2.4771

Consider the Laplace equation,

𝛿𝛿 2 𝑢𝑢 𝛿𝛿 2 𝑢𝑢
+ =0
𝛿𝛿𝑥𝑥 2 𝛿𝛿𝑦𝑦 2
The method to solve Laplace equation similar to solve Poisson’s equation with
𝑓𝑓(𝑥𝑥, 𝑦𝑦) = 0
 CECB423 Numerical Methods for Civil Engineers

Example 2:
𝛿𝛿 2 𝑢𝑢 𝛿𝛿 2 𝑢𝑢
+ =0 0 < 𝑥𝑥 < 3 ; 0 < 𝑦𝑦 < 1
𝛿𝛿𝑥𝑥 2 𝛿𝛿𝑦𝑦 2
With boundary conditions

𝑢𝑢(0, 𝑦𝑦) = 0, 𝑢𝑢(3, 𝑦𝑦) = 0 0 < 𝑦𝑦 < 1

𝑢𝑢(𝑥𝑥, 0) = 𝑥𝑥(3 − 𝑥𝑥), 𝑢𝑢(𝑥𝑥, 1) = 0 0 < 𝑥𝑥 < 3

Using finite difference method with ∆𝑥𝑥 = ℎ = 1 and ∆𝑦𝑦 = 𝑘𝑘 = 1/3

Solve the linear system 𝐴𝐴𝐴𝐴 = 𝑏𝑏 using Gauss Jordan method. Use three decimal places in
all calculation.

3/3 y
j=3

j=2 2/3 𝑢𝑢(𝑥𝑥, 1) = 0

𝑢𝑢(3, 𝑦𝑦) = 0
𝑢𝑢(0, 𝑦𝑦) = 0
j=1 1/3

𝑢𝑢(𝑥𝑥, 0) = 𝑥𝑥(3 − 𝑥𝑥)

j=0 0.0 x

0.0 1.0 2.0 3.0

i=0 i=1 i=2 i=3

𝑢𝑢(0, 𝑦𝑦) = 0 𝑢𝑢(3, 𝑦𝑦) = 0 𝑢𝑢(𝑥𝑥, 1) = 0 𝑢𝑢(𝑥𝑥, 0) = 𝑥𝑥(3 − 𝑥𝑥)

𝑢𝑢(0,0) = 0 𝑢𝑢(3,0) = 0 𝑢𝑢(0,1) = 0 𝑢𝑢(0,0) = 0(3 − 0) = 0

𝑢𝑢(0,1/3) = 0 𝑢𝑢(3,1/3) = 0 𝑢𝑢(1,1) = 0 𝑢𝑢(1,0) = 1(3 − 1) = 2
𝑢𝑢(0,2/3) = 0 𝑢𝑢(3,2/3) = 0 𝑢𝑢(2,1) = 0 𝑢𝑢(2,0) = 2(3 − 2) = 2
𝑢𝑢(0,3/3) = 0 𝑢𝑢(3,3/3) = 0 𝑢𝑢(3,1) = 0 𝑢𝑢(3,0) = 3(3 − 3) = 0
 CECB423 Numerical Methods for Civil Engineers

U0,3=0 U1,3=0 U2,3=0 U3,3=0

U0,2=0 U1,2 U2,2 U3,2=0

U0,1=0 U1,1 U2,1 U3,1=0

U0,0=0 U1,0=2 U2,0=2 U3,0=0

Approximate 𝑢𝑢1,1 , 𝑢𝑢2,1 , 𝑢𝑢1,2 𝑎𝑎𝑎𝑎𝑎𝑎 𝑢𝑢2,2

At point (𝑥𝑥𝑖𝑖 , 𝑦𝑦𝑗𝑗 )

𝛿𝛿 2 𝑢𝑢(𝑥𝑥𝑖𝑖 , 𝑦𝑦𝑗𝑗 ) 𝛿𝛿 2 𝑢𝑢(𝑥𝑥𝑖𝑖 , 𝑦𝑦𝑗𝑗 )

+ =0
𝛿𝛿𝑥𝑥 2 𝛿𝛿𝑦𝑦 2

using finite difference approximation

𝑢𝑢𝑖𝑖+1,𝑗𝑗 − 2𝑢𝑢𝑖𝑖,𝑗𝑗 + 𝑢𝑢𝑖𝑖−1,𝑗𝑗 𝑢𝑢𝑖𝑖,𝑗𝑗+1 − 2𝑢𝑢𝑖𝑖,𝑗𝑗 + 𝑢𝑢𝑖𝑖,𝑗𝑗−1

+ =0
ℎ2 𝑘𝑘 2
multiply both side ℎ2
ℎ2
𝑢𝑢𝑖𝑖+1,𝑗𝑗 − 2𝑢𝑢𝑖𝑖,𝑗𝑗 + 𝑢𝑢𝑖𝑖−1,𝑗𝑗 + �𝑢𝑢 − 2𝑢𝑢𝑖𝑖,𝑗𝑗 + 𝑢𝑢𝑖𝑖,𝑗𝑗−1 � = 0
𝑘𝑘 2 𝑖𝑖,𝑗𝑗+1
Given ℎ = 1, 𝑘𝑘 = 1/3
12
𝑢𝑢𝑖𝑖+1,𝑗𝑗 − 2𝑢𝑢𝑖𝑖,𝑗𝑗 + 𝑢𝑢𝑖𝑖−1,𝑗𝑗 + �𝑢𝑢 − 2𝑢𝑢𝑖𝑖,𝑗𝑗 + 𝑢𝑢𝑖𝑖,𝑗𝑗−1 � = 0
1/32 𝑖𝑖,𝑗𝑗+1
 CECB423 Numerical Methods for Civil Engineers

𝑢𝑢𝑖𝑖+1,𝑗𝑗 − 2𝑢𝑢𝑖𝑖,𝑗𝑗 + 𝑢𝑢𝑖𝑖−1,𝑗𝑗 + 9�𝑢𝑢𝑖𝑖,𝑗𝑗+1 − 2𝑢𝑢𝑖𝑖,𝑗𝑗 + 𝑢𝑢𝑖𝑖,𝑗𝑗−1 � = 0

𝑢𝑢𝑖𝑖+1,𝑗𝑗 − 2𝑢𝑢𝑖𝑖,𝑗𝑗 + 𝑢𝑢𝑖𝑖−1,𝑗𝑗 + 9𝑢𝑢𝑖𝑖,𝑗𝑗+1 − 18𝑢𝑢𝑖𝑖,𝑗𝑗 + 9𝑢𝑢𝑖𝑖,𝑗𝑗−1 = 0
𝑢𝑢𝑖𝑖+1,𝑗𝑗 + 𝑢𝑢𝑖𝑖−1,𝑗𝑗 − 20𝑢𝑢𝑖𝑖,𝑗𝑗 + 9𝑢𝑢𝑖𝑖,𝑗𝑗+1 + 9𝑢𝑢𝑖𝑖,𝑗𝑗−1 = 0

U1,1 𝑢𝑢21 + 𝑢𝑢01 − 20𝑢𝑢11 + 9𝑢𝑢12 + 9𝑢𝑢10 = 0

at i=1, j=1 𝑢𝑢21 + 0 − 20𝑢𝑢11 + 9𝑢𝑢12 + 9(2) = 0
𝒖𝒖𝟐𝟐𝟐𝟐 − 𝟐𝟐𝟐𝟐𝒖𝒖𝟏𝟏𝟏𝟏 + 𝟗𝟗𝒖𝒖𝟏𝟏𝟏𝟏 = −𝟏𝟏𝟏𝟏
U1,2 𝑢𝑢22 + 𝑢𝑢02 − 20𝑢𝑢12 + 9𝑢𝑢13 + 9𝑢𝑢11 = 0
at i=1, j=2 𝑢𝑢22 + 0 − 20𝑢𝑢12 + 9(0) + 9𝑢𝑢11 = 0
𝒖𝒖𝟐𝟐𝟐𝟐 − 𝟐𝟐𝟐𝟐𝒖𝒖𝟏𝟏𝟏𝟏 + 𝟗𝟗𝒖𝒖𝟏𝟏𝟏𝟏 = 𝟎𝟎
U2,1 𝑢𝑢31 + 𝑢𝑢11 − 20𝑢𝑢21 + 9𝑢𝑢22 + 9𝑢𝑢20 = 0
at i=2, j=1 0 + 𝑢𝑢11 − 20𝑢𝑢21 + 9𝑢𝑢22 + 9(2) = 0
𝒖𝒖𝟏𝟏𝟏𝟏 − 𝟐𝟐𝟐𝟐𝒖𝒖𝟐𝟐𝟐𝟐 + 𝟗𝟗𝒖𝒖𝟐𝟐𝟐𝟐 = −𝟏𝟏𝟏𝟏
U2,2 𝑢𝑢32 + 𝑢𝑢12 − 20𝑢𝑢22 + 9𝑢𝑢23 + 9𝑢𝑢21 = 0
at i=2, j=2 0 + 𝑢𝑢12 − 20𝑢𝑢22 + 9(0) + 9𝑢𝑢21 = 0
𝒖𝒖𝟏𝟏𝟏𝟏 − 𝟐𝟐𝟐𝟐𝒖𝒖𝟐𝟐𝟐𝟐 + 𝟗𝟗𝒖𝒖𝟐𝟐𝟐𝟐 = 𝟎𝟎

If U1,1=a, U1,2=b, U2,1=c, U2,2=d

𝑐𝑐 − 20𝑎𝑎 + 9𝑏𝑏 = −18
𝑑𝑑 − 20𝑏𝑏 + 9𝑎𝑎 = 0
𝑎𝑎 − 20𝑐𝑐 + 9𝑑𝑑 = −18
𝑏𝑏 − 20𝑑𝑑 + 9𝑐𝑐 = 0

−20 9 1 0 𝑎𝑎 −18
9 −20 0 1 𝑏𝑏 0
� �� � = � �
1 0 −20 9 𝑐𝑐 −18
0 1 9 −20 𝑑𝑑 0

𝑼𝑼𝟏𝟏𝟏𝟏 = 𝟏𝟏. 𝟐𝟐𝟐𝟐𝟐𝟐𝟐𝟐, 𝑼𝑼𝟏𝟏𝟏𝟏 = 𝟎𝟎. 𝟓𝟓𝟓𝟓𝟓𝟓𝟓𝟓, 𝑼𝑼𝟐𝟐𝟐𝟐 = 𝟏𝟏. 𝟐𝟐𝟐𝟐𝟐𝟐𝟐𝟐, 𝑼𝑼𝟐𝟐𝟐𝟐 = 𝟎𝟎. 𝟓𝟓𝟓𝟓𝟓𝟓𝟓𝟓
 CECB423 Numerical Methods for Civil Engineers

Example 3
Solve the Laplace equation using finite difference method with ∆𝑥𝑥 = ℎ = 1 and ∆𝑦𝑦 = 𝑘𝑘 = 1
𝛿𝛿 2 𝑢𝑢 𝛿𝛿 2 𝑢𝑢
+ =0 0 < 𝑥𝑥 < 2 ; 0 < 𝑦𝑦 < 2
𝛿𝛿𝑥𝑥 2 𝛿𝛿𝑦𝑦 2

Solve the linear system 𝐴𝐴𝐴𝐴 = 𝑏𝑏 using Gauss Jordan method. Use four decimal places in all
calculation.
𝜕𝜕𝜕𝜕
=0
𝜕𝜕𝜕𝜕

𝑈𝑈0,2 𝑈𝑈1,2 𝑈𝑈2,2

𝜕𝜕𝜕𝜕
=1 𝑈𝑈𝑖𝑖,𝑗𝑗 = 0
𝜕𝜕𝜕𝜕
𝑈𝑈0,1 𝑈𝑈1,1 𝑈𝑈2,1

𝑈𝑈0,0 𝑈𝑈1,0 𝑈𝑈2,0

𝑈𝑈𝑖𝑖,𝑗𝑗 = 0

𝑈𝑈0,2 𝑈𝑈1,2 𝑈𝑈2,2 = 0

𝑈𝑈0,1 𝑈𝑈1,1 𝑈𝑈2,1 = 0

𝑈𝑈0,0 = 0 𝑈𝑈1,0 = 0 𝑈𝑈2,0 = 0

 CECB423 Numerical Methods for Civil Engineers

𝛿𝛿 2 𝑢𝑢 𝛿𝛿 2 𝑢𝑢
+ =0
𝛿𝛿𝑥𝑥 2 𝛿𝛿𝑦𝑦 2
using finite difference approximation

𝑢𝑢𝑖𝑖+1,𝑗𝑗 − 2𝑢𝑢𝑖𝑖,𝑗𝑗 + 𝑢𝑢𝑖𝑖−1,𝑗𝑗 𝑢𝑢𝑖𝑖,𝑗𝑗+1 − 2𝑢𝑢𝑖𝑖,𝑗𝑗 + 𝑢𝑢𝑖𝑖,𝑗𝑗−1

+ =0
ℎ2 𝑘𝑘 2
multiply both side ℎ2
ℎ2
𝑢𝑢𝑖𝑖+1,𝑗𝑗 − 2𝑢𝑢𝑖𝑖,𝑗𝑗 + 𝑢𝑢𝑖𝑖−1,𝑗𝑗 + 2 �𝑢𝑢𝑖𝑖,𝑗𝑗+1 − 2𝑢𝑢𝑖𝑖,𝑗𝑗 + 𝑢𝑢𝑖𝑖,𝑗𝑗−1 � = 0
𝑘𝑘
Given ℎ = 1, 𝑘𝑘 = 1

𝑢𝑢𝑖𝑖+1,𝑗𝑗 − 2𝑢𝑢𝑖𝑖,𝑗𝑗 + 𝑢𝑢𝑖𝑖−1,𝑗𝑗 + 𝑢𝑢𝑖𝑖,𝑗𝑗+1 − 2𝑢𝑢𝑖𝑖,𝑗𝑗 + 𝑢𝑢𝑖𝑖,𝑗𝑗−1 = 0

𝑢𝑢𝑖𝑖+1,𝑗𝑗 + 𝑢𝑢𝑖𝑖−1,𝑗𝑗 − 4𝑢𝑢𝑖𝑖,𝑗𝑗 + 𝑢𝑢𝑖𝑖,𝑗𝑗+1 + 𝑢𝑢𝑖𝑖,𝑗𝑗−1 = 0

At 𝑈𝑈0,1
𝑢𝑢1,1 + 𝑢𝑢−1,1 − 4𝑢𝑢0,1 + 𝑢𝑢0,2 + 𝑢𝑢0,0 = 0
𝑢𝑢1,1 + 𝑢𝑢1,1 − 4𝑢𝑢0,1 + 𝑢𝑢0,2 + 0 = 0
𝒖𝒖𝟏𝟏,𝟏𝟏 − 𝟒𝟒𝒖𝒖𝟎𝟎,𝟏𝟏 + 𝒖𝒖𝟎𝟎,𝟐𝟐 = 𝟎𝟎−→ (𝟏𝟏)

External boundary
𝜕𝜕𝜕𝜕
=0
𝜕𝜕𝜕𝜕
𝑢𝑢𝑖𝑖+1,𝑗𝑗 − 𝑢𝑢𝑖𝑖−1,𝑗𝑗
=0
2ℎ
𝑢𝑢1,1 − 𝑢𝑢−1,1
=0
2ℎ

𝑢𝑢1,1 − 𝑢𝑢−1,1 = 0

𝑢𝑢1,1 = 𝑢𝑢−1,1

At 𝑈𝑈1,1
𝑢𝑢2,1 + 𝑢𝑢0,1 − 4𝑢𝑢1,1 + 𝑢𝑢1,2 + 𝑢𝑢1,0 = 0
0 + 𝑢𝑢0,1 − 4𝑢𝑢1,1 + 𝑢𝑢1,2 + 0 = 0
𝒖𝒖𝟎𝟎,𝟏𝟏 − 𝟒𝟒𝒖𝒖𝟏𝟏,𝟏𝟏 + 𝒖𝒖𝟏𝟏,𝟐𝟐 = 𝟎𝟎−→ (𝟐𝟐)
 CECB423 Numerical Methods for Civil Engineers

At 𝑈𝑈0,2
𝑢𝑢1,2 + 𝑢𝑢−1,2 − 4𝑢𝑢0,2 + 𝑢𝑢0,3 + 𝑢𝑢0,1 = 0
𝑢𝑢1,2 + 𝑢𝑢1,2 − 4𝑢𝑢0,2 + (2 + 𝑢𝑢0,1 ) + 𝑢𝑢0,1 = 0
𝟐𝟐𝒖𝒖𝟏𝟏,𝟐𝟐 − 𝟒𝟒𝒖𝒖𝟎𝟎,𝟐𝟐 +𝟐𝟐𝟐𝟐𝟎𝟎,𝟏𝟏 = −𝟏𝟏−→ (𝟑𝟑)

External boundary

𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕
=0 =1
𝜕𝜕𝜕𝜕 𝜕𝜕𝜕𝜕
𝑢𝑢𝑖𝑖,𝑗𝑗+1 − 𝑢𝑢𝑖𝑖,𝑗𝑗−1
𝑢𝑢1,2 − 𝑢𝑢−1,2 =1
=0 2𝑘𝑘
2ℎ
𝑢𝑢0,3 − 𝑢𝑢0,1
=1
𝑢𝑢1,2 − 𝑢𝑢−1,2 = 0 2(1)

𝑢𝑢1,2 = 𝑢𝑢−1,2 𝑢𝑢0,3 − 𝑢𝑢0,1 = 2

𝑢𝑢0,3 = 2 + 𝑢𝑢0,1

At 𝑈𝑈1,2

𝑢𝑢2,2 + 𝑢𝑢0,2 − 4𝑢𝑢1,2 + 𝑢𝑢1,3 + 𝑢𝑢1,1 = 0

0 + 𝑢𝑢0,2 − 4𝑢𝑢1,2 + (2 + 𝑢𝑢1,1 ) + 𝑢𝑢1,1 = 0
𝒖𝒖𝟎𝟎,𝟐𝟐 − 𝟒𝟒𝒖𝒖𝟏𝟏,𝟐𝟐 + 𝟐𝟐𝒖𝒖𝟏𝟏,𝟏𝟏 = −𝟐𝟐−→ (𝟒𝟒)

External boundary
𝜕𝜕𝜕𝜕
=1
𝜕𝜕𝜕𝜕

𝑢𝑢𝑖𝑖,𝑗𝑗+1 − 𝑢𝑢𝑖𝑖,𝑗𝑗−1
=1
2𝑘𝑘
𝑢𝑢1,3 − 𝑢𝑢1,1
=1
2(1)

𝑢𝑢1,3 − 𝑢𝑢1,1 = 2

𝑢𝑢1,3 = 2+𝑢𝑢1,1
 CECB423 Numerical Methods for Civil Engineers

𝑢𝑢1,1 − 4𝑢𝑢0,1 + 𝑢𝑢0,2 = 0−→ (1)

𝑢𝑢0,1 − 4𝑢𝑢1,1 + 𝑢𝑢1,2 = 0−→ (2)
2𝑢𝑢1,2 − 4𝑢𝑢0,2 +2𝑢𝑢0,1 = −1−→ (3)
𝑢𝑢0,2 − 4𝑢𝑢1,2 + 2𝑢𝑢1,1 = −2−→ (4)

𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴 𝑢𝑢0,1 = 𝑎𝑎, 𝑢𝑢0,2 = 𝑏𝑏, 𝑢𝑢1,1 = 𝑐𝑐 𝑎𝑎𝑎𝑎𝑎𝑎 𝑢𝑢1,2 = 𝑑𝑑

𝑐𝑐 − 4𝑎𝑎 + 𝑏𝑏 = 0−→ (1)

𝑎𝑎 − 4𝑐𝑐 + 𝑑𝑑 = 0−→ (2)
2𝑑𝑑 − 4𝑏𝑏 + 2𝑎𝑎 = −1−→ (3)
𝑏𝑏 − 4𝑑𝑑 + 2𝑐𝑐 = −2−→ (4)

Rearrange the eqn

−4𝑎𝑎 + 𝑏𝑏 + 𝑐𝑐 = 0−→ (1)
𝑎𝑎 − 4𝑐𝑐 + 𝑑𝑑 = −→ (2)
2𝑎𝑎 − 4𝑏𝑏 + 2𝑑𝑑 = −1−→ (3)
𝑏𝑏 + 2𝑐𝑐 − 4𝑑𝑑 = −2−→ (4)

𝒖𝒖𝟎𝟎,𝟏𝟏 = 𝒂𝒂 = 𝟎𝟎. 𝟐𝟐𝟐𝟐𝟐𝟐𝟐𝟐

𝒖𝒖𝟎𝟎,𝟐𝟐 = 𝒃𝒃 = 𝟎𝟎. 𝟖𝟖𝟖𝟖𝟖𝟖𝟖𝟖
𝒖𝒖𝟏𝟏,𝟏𝟏 = 𝒄𝒄 = 𝟎𝟎. 𝟐𝟐𝟐𝟐𝟐𝟐𝟐𝟐
𝒖𝒖𝟏𝟏,𝟐𝟐 = 𝒅𝒅 = 𝟎𝟎. 𝟖𝟖𝟖𝟖𝟖𝟖𝟑𝟑