0% found this document useful (0 votes)
9 views36 pages

Power Flows-Part 1 (Iterative Solutions) .v2

Uploaded by

Cesar Manalo Jr.
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
9 views36 pages

Power Flows-Part 1 (Iterative Solutions) .v2

Uploaded by

Cesar Manalo Jr.
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 36

Malayan Colleges Laguna

Department of
Electrical Engineering

EE 134 POWER FLOWS


Power System
Analysis 2 Part 1:
Iterative Solutions to Algebraic Equations:
Jacobi, Gauss-Seidel, and the Newton-
Raphson
Prepared by: Cesar G. Manalo, Jr.
LEARNING OBJECTIVES
• Derive the iterative equations for solving linear algebraic equations using Jacobi
and Gauss-Seidel methods.
• Derive the iterative equations for solving nonlinear algebraic equations using the
Newton-Raphson method.
• Show examples using the above iterative methods.
ITERATIVE SOLUTION TO LINEAR ALGERAIC
EQUATIONS:
THE JACOBI METHOD
• Given the linear algebraic equations,

[ ][ ] [ ]
𝐴 11 𝐴 12 ⋯ 𝐴1 𝑁
𝑥1 𝑦1
𝐴 21 𝐴 22 ⋯ 𝐴2 𝑁
𝑥2 𝑦2
⋮ ⋮ ⋮ ⋮
⋮ ⋮
𝐴 𝑘1
𝐴 ( 𝑘+1 ) 1
𝐴𝑘 2 ⋯
𝐴( 𝑘+1 ) 2 ⋯
𝐴 𝑘𝑁
𝐴( 𝑘+1 ) 𝑁
𝑥𝑘 = 𝑦𝑘 𝐀𝐱 = 𝐲
𝑥𝑘 +1 𝑦 𝑘 +1
⋮ ⋮ ⋮ ⋮
⋮ ⋮
𝐴𝑁 1 𝐴𝑁 2 ⋯ 𝐴 𝑁𝑁
𝑥𝑁 𝑦𝑁

where A is an N x N matrix, x is a vector whose values are to be determined, and


y is a given vector. Note that matrices and vectors are written in boldface.
• Extracting the kth equation,

𝑦 𝑘= 𝐴 𝑘 1 𝑥 1+ 𝐴𝑘 2 𝑥 2 +⋯ + 𝐴𝑘𝑘 𝑥 𝑘 +⋯ + 𝐴 𝑘𝑁 𝑥 𝑁
ITERATIVE SOLUTION TO LINEAR ALGERAIC
EQUATIONS:
THE JACOBI METHOD
𝑦 𝑘= 𝐴 𝑘 1 𝑥 1+ 𝐴𝑘 2 𝑥 2 +⋯ + 𝐴𝑘𝑘 𝑥 𝑘 +⋯ + 𝐴 𝑘𝑁 𝑥 𝑁
Solving in the kth equation,
1
𝑥 𝑘= [ 𝑦 − ( 𝐴𝑘 1 𝑥 1+ ⋯ + 𝐴𝑘 ,𝑘 −1 𝑥𝑘 −1 +⋯ + 𝐴𝑘 ,𝑘+1 𝑥𝑘 +1 +⋯ + 𝐴𝑘𝑁 𝑥 𝑁 ) ]
𝐴𝑘𝑘 𝑘

[ ]
𝑘 −1 𝑁
1
𝑥 𝑘= 𝑦 𝑘 − ∑ 𝐴𝑘𝑛 𝑥𝑛 − ∑ 𝐴𝑘𝑛 𝑥 𝑛 Eq. 1
𝐴𝑘𝑘 𝑛=1 𝑛=𝑘 +1

• Equation 1 suggest that can be found if all other values of are available.
ITERATIVE SOLUTION TO LINEAR ALGERAIC
EQUATIONS:
THE JACOBI METHOD
• Equation 1 is the basis of the Jacobi method by repeatedly using “old” values
of on the right side to generate the “new” value on the left side, where is the
iteration. Rewriting Equation 1,

[ ]
𝑘 −1 𝑁
1
𝑥 𝑘 ( 𝑖+1 ) = 𝑦 𝑘 − ∑ 𝐴𝑘𝑛 𝑥𝑛 ( 𝑖 ) − ∑ 𝐴𝑘𝑛 𝑥𝑛 ( 𝑖 ) 𝑘=1 ,2 , …, 𝑁 Eq. 2
𝐴𝑘𝑘 𝑛=1 𝑛=𝑘+1

• The iteration continues until the following stopping condition is satisfied,


as
| 𝑥𝑘 ( 𝑖) |
𝑥𝑘 ( 𝑖+1 ) − 𝑥 𝑘 ( 𝑖 )
< 𝜀 𝑓𝑜𝑟 𝑎𝑙𝑙 𝑘=1 , 2 ,… , 𝑁 Eq. 3

where is the kth component of vector and is a specified tolerance level.


ITERATIVE SOLUTION TO LINEAR ALGERAIC
EQUATIONS:
THE JACOBI METHOD
• The Jacobi method can be expressed in matrix format as follows,
𝐱 ( 𝑖+1 ) =𝐌𝐱 ( 𝑖 ) + 𝐃− 1 𝐲 Eq. 4
where,
𝐌= 𝐃−1 ( 𝐃 − 𝐀 )
and

[ ]
𝐴 11 0 ⋯ 0
𝐃= 0 𝐴 22 ⋯ 0
⋮ ⋮ ⋮ 0
0 0 0 𝐴 𝑁𝑁

is the diagonal elements of the A matrix.


ITERATIVE SOLUTION TO LINEAR ALGERAIC
EQUATIONS:
THE GAUSS-SEIDEL METHOD
• Another technique called the Gauss-Seidel method uses a similar approach as
Jacobi, except that during each iteration, the “new” values, , for are used on
the right side to generate the “new” value on the left side.

[ ]
𝑘 −1 𝑁
1
𝑥 𝑘 ( 𝑖+1 ) = 𝑦 𝑘 − ∑ 𝐴𝑘𝑛 𝑥𝑛 ( 𝑖+1 ) − ∑ 𝐴𝑘𝑛 𝑥𝑛 ( 𝑖 ) 𝑘=1 ,2 , …, 𝑁 Eq. 5
𝐴𝑘𝑘 𝑛=1 𝑛=𝑘+1
ITERATIVE SOLUTION TO LINEAR ALGERAIC
EQUATIONS:
THE GAUSS-SEIDEL METHOD
• The Gauss-Seidel method can also be written in matrix format,
𝐱 ( 𝑖+1 ) =𝐌𝐱 ( 𝑖 ) + 𝐃− 1 𝐲
where,
𝐌= 𝐃−1 ( 𝐃 − 𝐀 )

[ ]
𝐴11 0 0 ⋯ 0
𝐴21 𝐴22 0 ⋯ 0
𝐃= ⋮ ⋱
⋮ ⋮ ⋮
𝐴𝑁1 𝐴𝑁2 ⋯ ⋯ 𝐴 𝑁𝑁

is the lower triangular portion of A.


§ Example 1.1
Solve the linear algebraic equations below using the Jacobi method. Start with and
and continue until convergence criterion .

[ 10
2
5
9 ][ ] [ ]
𝑥1
𝑥2
=
6
3
10 𝑥1 +5 𝑥 2=6
2 𝑥1 +9 𝑥 2=3
§ Example 1.1-Solution
Using Eq. 2,

[ ]
𝑘 −1 𝑁
1
𝑥 𝑘 ( 𝑖+1 ) = 𝑦 𝑘 − ∑ 𝐴𝑘𝑛 𝑥𝑛 ( 𝑖 ) − ∑ 𝐴𝑘𝑛 𝑥𝑛 ( 𝑖 ) 𝑘=1 ,2
𝐴𝑘𝑘 𝑛=1 𝑛=𝑘+1

𝑘=1 :

[ ]
𝟐
𝟏 𝟏 𝟏
𝒙 𝟏 ( 𝒊+𝟏 )= 𝒚 𝟏 − ∑ 𝑨𝟏 𝒏 𝒙𝒏 ( 𝒊 ) = [ 𝒚 𝟏 − 𝑨 𝟏 𝟐 𝒙 𝟐 ( 𝒊 ) ]= [ 𝟔− 𝟓 𝒙𝟐 ( 𝒊 ) ]
𝑨 𝟏𝟏 𝒏=𝟐 𝑨 𝟏𝟏 𝟏𝟎
𝑘=2:

[ ]
𝟏
𝟏 𝟏 𝟏
𝒙 𝟐 ( 𝒊+𝟏 )= 𝒚 𝟐 − ∑ 𝑨𝟐 𝒏 𝒙𝒏 ( 𝒊 ) = [ 𝒚 𝟐 − 𝑨 𝟐𝟏 𝒙 𝟏 ( 𝒊 ) ]= [ 𝟑−𝟐 𝒙 𝟏 ( 𝒊 ) ]
𝑨 𝟐𝟐 𝒏=𝟏 𝑨 𝟐𝟐 𝟗
§ Example 1.1-Solution
Using Eq. 4 to form the matrix iteration formula,
𝐱 ( 𝑖+1 ) =𝐌𝐱 ( 𝑖 ) + 𝐃− 1 𝐲

where, .

[ ]
𝐴 11 0 ⋯ 0
𝐃= 0

𝐴 22 ⋯
⋮ ⋮
0
0
=
10
0 [ 0
9 ]
0 0 0 𝐴 𝑁𝑁

[ ]
1
0
[0 ] [0 ] [ ] [ ]
−1
𝐃− 1 = 1 0 0
=
10
𝐃 − 𝐀 = 10 0

10 5
=
0 −5
9 1 9 2 9 −2 0
0
9
§ Example 1.1-Solution

[ ] [ ]
1 5
0 0 −
𝐌= 𝐃−1 ( 𝐃 − 𝐀 ) = 10
0
1 [ 0
−2
−5
0
=
] −
2
10
0
9 9
𝐱 ( 𝑖+1 ) =𝐌𝐱 ( 𝑖 ) + 𝐃− 1 𝐲

] [ ][ [ ]
5 1
0 − 0
[ 𝑥1 (𝑖 + 1)
𝑥2 (𝑖 + 2 )
=

2
9
10
0
]
𝑥 1( 𝑖)
𝑥 2 (𝑖)
+
10
0
1
9
[ ]
6
3

Expanding,
§ Example 1.1-Solution
Performing the iteration starting w/ and at ,
𝑘=1 :
𝟏
𝒙 𝟏 ( 𝒊+ 𝟏 )=
𝟏𝟎
[ 𝟔 −𝟓 𝒙 𝟐 ( 𝒊 ) ]

𝑘=2:
𝟏
𝒙 𝟐 ( 𝒊+ 𝟏 )=
𝟗
[ 𝟑− 𝟐 𝒙 𝟏 ( 𝒊 ) ]

|
𝜀 𝑥 1=
𝑥1( 1) ||
𝑥1 ( 2 ) − 𝑥1 ( 1 )
=
0.6 |
0.4333− 0.6
=0.27783 |
𝜀 𝑥 2=
𝑥2 1( ) ||
𝑥2( 2) − 𝑥2( 1)
=
0.2 − 0.3333
0.3333 |
=0.4
§ Example 1.1-Solution
Continuing iteration will reveal that convergence criterion is satisfied starting at i =
10 for both as shown by the table below.

i 0 1 2 3 4 5 6 7 8 9 10
x1(i) 0.0000 0.6000 0.4333 0.5000 0.4815 0.4889 0.4868 0.4877 0.4874 0.4875 0.4875
x2(i) 0.0000 0.3333 0.2000 0.2370 0.2222 0.2263 0.2247 0.2251 0.2250 0.2250 0.2250
ϵx1 x ∞ 0.2778 0.1538 0.0370 0.0154 0.0042 0.0017 0.0005 0.0002 0.0001
ϵx2 x ∞ 0.4000 0.1852 0.0625 0.0185 0.0073 0.0020 0.0008 0.0002 0.0001

Hence,
𝒙 𝟏=𝟎 . 𝟒𝟖𝟕𝟓
𝒙 𝟐=𝟎 . 𝟐𝟐𝟓𝟎
§ Example 1.2
Rework example 1.1 using the Gauss-Seidel method. [ 10
2
5
9 ][ ] [ ]
𝑥1
𝑥2
=
6
3
Solution:
Using Eq. 5,

[ ]
𝑘 −1 𝑁
1
𝑥 𝑘 ( 𝑖+1 ) = 𝑦 𝑘 − ∑ 𝐴𝑘𝑛 𝑥𝑛 ( 𝑖+1 ) − ∑ 𝐴𝑘𝑛 𝑥𝑛 ( 𝑖 ) 𝑘=1 ,2 , …, 𝑁
𝐴𝑘𝑘 𝑛=1 𝑛=𝑘+1

1 1
𝑥1 ( 𝑖+1 ) =
𝐴 11
[ 𝑦 1 − 𝐴 12 𝑥 2 𝑖 ] =
( )
10
[ 6 − 5 𝑥2 ( 𝑖) ]

1 1
𝑥 2 ( 𝑖+1 ) =
𝐴 22
[ 𝑦 2 − 𝐴 21 𝑥1 𝑖+1 ] = [ 3 − 2 𝑥 1 ( 𝑖+1 ) ]
( )
9

𝑥 2 ( 𝑖+1 ) =
1
9 [
3−
2
10
[ 6 −5 𝑥 2 ( 𝑖 ) ] =
] [
1 9
9 5
+ 𝑥2 ( 𝑖)
]
§ Example 1.2
Forming the iteration formula using matrix,
𝐀 = [ 120 5
]
𝐱 ( 𝑖+1 ) =𝐌𝐱 ( 𝑖 ) + 𝐃− 1 𝐲
9

[2 ]
[] ] [ ]
𝐃= 1 0 0
0 −
1 1
0
9
[ 𝑥1 (𝑖+ 1)
] 2
[ ]
𝑥 1 (𝑖) 10
[]
6

[
= +
1 𝑥2 (𝑖+ 1 𝑥 2 (𝑖) 2 1 3
0 2) 0 −
[ ]
−1
10 0 10 9 90 9
𝐃 −1
= =
2 9 2 1

90 9
1
𝑥1 ( 𝑖+1 ) = [ 6 − 5 𝑥 2 (𝑖 ) ]

[ ]
10
1
0 −
2
𝐌= 𝐃−1 ( 𝐃 − 𝐀 ) =
0
1
9
𝑥 2 ( 𝑖+1 ) =
[
1 9
9 5
+ 𝑥2( 𝑖 )
]
§ Example 1.2-Solution
Performing the iteration starting w/ and at ,

|
𝜀 𝑥 1=
𝑥1( 1) ||
𝑥1 ( 2 ) − 𝑥1 ( 1 )
= |
0.5− 0.6
0.6
=0.1667 |
𝜀 𝑥 2=
𝑥2 ( 1) ||
𝑥2( 2) − 𝑥2( 1)
=
0.2 |
0.2222 − 0.2
=0.1111
§ Example 1.2-Solution
i 0 1 2 3 4 5 6 7
x1(i) 0.0000 0.6000 0.5000 0.4889 0.4877 0.4875 0.4875 0.4875
x2(i) 0.0000 0.2000 0.2222 0.2247 0.2250 0.2250 0.2250 0.2250
ϵx1 X ∞ 0.1667 0.0222 0.0025 0.0003 0.0000 0.0000
ϵx2 X ∞ 0.1111 0.0111 0.0012 0.0001 0.0000 0.0000

𝒙 𝟏=𝟎 . 𝟒𝟖𝟕𝟓
𝒙 𝟐=𝟎 . 𝟐𝟐𝟓𝟎

• As shown, the convergence rate is faster using Gauss-Seidel.


• This is not always the case as some matrices converge faster with Jacobi.
• In some cases, one method converge while the other diverges.
ITERATIVE SOLUTION TO NONLINEAR ALGERAIC
EQUATIONS:
THE NEWTON-RAPHSON METHOD
• A set of nonlinear algebraic equations in matrix form can be written as,

[ ]
𝑓 1(𝐱)
𝐟 ( 𝐱 )= 𝑓 2 ( 𝐱 ) = 𝐲 Eq. 6

𝑓 𝑁(𝐱)

where y and x are N vectors and f(x) is an N vector of functions.

• The objective is to solve for the vector x given the vectors y and f(x).
ITERATIVE SOLUTION TO NONLINEAR ALGERAIC
EQUATIONS:
THE NEWTON-RAPHSON METHOD
• An example of these system of function would be,

𝑓 1 ( 𝑥 ) = 2 𝑥 12+ 𝑥 2= 1
2
𝑓 2 ( 𝑥 ) =− 𝑥1 + 𝑥 2 =0
𝐟 ( 𝐱 )=
[ ] [
𝑓 1 ( 𝑥)
𝑓 2 ( 𝑥)
=
]
2 𝑥12 + 𝑥 2
− 𝑥1 + 𝑥 22
= 𝐲 []
= 1
0
ITERATIVE SOLUTION TO NONLINEAR ALGERAIC
EQUATIONS:
THE NEWTON-RAPHSON METHOD
• To solve x, we manipulate Eq. 6,
0= 𝐲 − 𝐟 ( 𝐱 )
• Adding Dx to both sides where D is a square N x N invertible matrix,
𝐃 𝐱 =𝐃𝐱 + 𝐲 − 𝐟 ( 𝐱 )

• Solving for x,
𝐱 = 𝐱 + 𝑫− 𝟏 [ 𝐲 − 𝐟 ( 𝐱 ) ]

• This will now be the basis for iteration, by rewriting the last equation as,

𝐱 ( 𝑖+1)= 𝐱 (𝑖)+ 𝐃−𝟏 { 𝒚 − 𝒇 [ 𝒙 ( 𝑖) ] }

where D must be specified for nonlinear equations.


ITERATIVE SOLUTION TO NONLINEAR ALGERAIC
EQUATIONS:
THE NEWTON-RAPHSON METHOD
• A method for specifying D is by expanding f(x) by its Taylor series about ,

| |
2
𝑑𝐟 ( 𝐱 − 𝐱𝟎) 𝑑 𝐟 2
( 𝐱 − 𝐱 𝟎)
𝐲 = 𝐟 ( 𝐱 ) =𝐟 ( 𝐱 0 ) + + +…
𝑑𝐱 𝐱= 𝐱𝟎 1! 𝑑𝐱
2
𝐱 = 𝐱𝟎
2!
• Neglecting higher order terms in the series, and solving for x,

[ | ]
−𝟏
𝑑𝐟
𝐱 = 𝐱𝟎 +
𝑑𝐱 𝐱= 𝐱𝟎
[ 𝐲 − 𝐟 ( 𝐱0) ]
ITERATIVE SOLUTION TO NONLINEAR ALGERAIC
EQUATIONS:
THE NEWTON-RAPHSON METHOD
• The Newton-Raphson replaces by the old values x(i) and x by the new value
x(i+1). The last equation can be rewritten as,
𝐱 ( 𝑖+1)= 𝐱 (𝑖)+ 𝐉− 1 (𝑖) { 𝐲 − 𝐟 [ 𝐱 ( 𝑖) ] } Eq. 7

[ ]
where,
𝜕𝑓1 𝜕𝑓1 𝜕𝑓1

𝜕 𝑥1 𝜕 𝑥2 𝜕 𝑥𝑁
𝜕𝑓 2 𝜕𝑓2 𝜕𝑓1
𝐉 (𝑖)=
𝑑𝐟
|
𝑑 𝐱 𝐱 = 𝐱 (𝑖 )
= 𝜕𝑥

1 𝜕 𝑥2
⋮ ⋱

𝜕 𝑥𝑁

𝜕𝑓 𝑁 𝜕𝑓𝑁 𝜕𝑓𝑁

𝜕 𝑥1 𝜕 𝑥2 𝜕𝑥𝑁 𝐱 = 𝐱 (𝑖 )

• J(i) w/c is an N x N matrix of partial derivatives is called the Jacobian matrix.


§ Example 1.3
Solve the scalar equation , where and . Starting with , use (a) Newton-Raphson and
(b) extended Gauss-Seidel with D = 3 until . Compare the two methods.

Solution:
a. Using Newton-Raphson. With ,

𝐉 ( 𝑖 )=
𝑑𝐟
𝑑𝐱 |
𝐱 =𝐱 (𝑖 ) [ ]
=
𝜕𝑓 1
𝜕 𝑥1 𝑥=𝑥 ( 𝑖 )
= [ ]
𝜕𝑓
𝜕𝑥 𝑥 =𝑥 ( 𝑖 )
= [ 𝜕 ( 2)
𝜕𝑥
𝑥 ] 𝑥= 𝑥 ( 𝑖 )
=[ 2 𝑥 ] 𝑥=𝑥 ( 𝑖 )=2 𝑥 (𝑖)

The inverse of J(i) is,


−𝟏 1
𝐉 ( 𝑖 )= [ 2 𝑥 (𝑖) ]
−1
=
2 𝑥( 𝑖)
§ Example 1.3-Solution
Using Eq. 7,
𝐱 ( 𝑖+1)= 𝐱 (𝑖)+ 𝐉− 1 (𝑖) { 𝐲 − 𝐟 [ 𝐱 ( 𝑖) ] }

Since there is only one unknown,


𝑥 (𝑖+1)= 𝑥 (𝑖)+ 𝐉 −1 (𝑖) { 𝑦 − 𝑓 [ 𝑥 (𝑖) ] }

1
𝑥 (𝑖+1)=𝑥 (𝑖)+ [ 9 − 𝑥2 ( 𝑖) ]
2 𝑥 (𝑖)
With , successive iteration using the Newton-Raphson is shown in the following
table with satisfied after 5 iterations.

i 0 1 2 3 4 5 6 7
x(i) 1.0000 5.0000 3.4000 3.0235 3.0001 3.0000 3.0000 3.0000
ϵx X 8.00E-01 4.71E-01 1.25E-01 7.81E-03 3.05E-05 4.66E-10 0.00E+00
§ Example 1.3-Solution

b. Using extended Gauss-Seidel with D = 3.

Using the formula ,

1
𝑥 ( 𝑖+1 ) = 𝑥 ( 𝑖 ) +
3
[ 2
9 − 𝑥 (𝑖) ]
i 0 1 2 3 4 5 6 7 8 9 10
x(i) 1.0000 3.6667 2.1852 3.5935 2.2891 3.5425 2.3595 3.5038 2.4116 3.4730 2.4525
ϵx X 7.27E-01 6.78E-01 3.92E-01 5.70E-01 3.54E-01 5.01E-01 3.27E-01 4.53E-01 3.06E-01 4.16E-01

As shown, Gauss-Seidel oscillates about the solution, slowly converging. Here


we assume that the matrix inverse exists. Thus, the initial value of should be
avoided.
§ Example 1.4
Solve the system of nonlinear algebraic equation,

[ ]
Solution: 𝜕𝑓 1 𝜕 𝑓1
𝑓 1 ( 𝑥 ) =
𝐉 ( 𝑖 ) =𝑥𝜕 𝑥1 +
1
𝜕 𝑥2 𝑥
𝜕𝑓 2 𝜕𝑓2
2 = 1
𝑥2 [ 1
𝑥1 ] 𝐱 =𝐱 (𝑖 )
=
[ 1
𝑥 2 (𝑖)
1
𝑥1 (𝑖) ]
𝑓 2 ( 𝑥 ) =𝑥 𝜕 𝑥 2𝑥
𝜕 𝑥1 1 𝐱 = 𝐱 ( 𝑖2
)

[ ] | [| ] [ ]
−1
1 𝑥 1 (𝑖) −1 1 𝑥 1 (𝑖) −1
𝐉 −1
= 1 1 = =
𝑥 2 (𝑖) 𝑥1 (𝑖) 1 1 − 𝑥 2 (𝑖) 1 𝑥 1 ( 𝑖 ) − 𝑥 2( 𝑖) − 𝑥 2 (𝑖) 1
𝑥 2 (𝑖) 𝑥1 (𝑖)
§ Example 1.4-Solution
𝐱 ( 𝑖+1)= 𝐱 (𝑖)+ 𝐉− 1 (𝑖) { 𝐲 − 𝐟 [ 𝐱 ( 𝑖) ] }

[ 𝑥1 (𝑖+ 1)
𝑥2 (𝑖+ 2) ][
𝑥 (𝑖)
= 1
𝑥 2( 𝑖)
+
] 1
[ 𝑥 1 (𝑖)
𝑥 1 ( 𝑖 ) − 𝑥 2 (𝑖) − 𝑥 2 (𝑖)
−1
1 ][ 1 5 − 𝑥 1( 𝑖)+ 𝑥 2 (𝑖)
50 − 𝑥 1( 𝑖) 𝑥2 (𝑖) ]
Expanding the equations,
𝑥 1 ( 𝑖 ) [ 1 5 − 𝑥1 ( 𝑖 ) + 𝑥 2 ( 𝑖 ) ] − [ 50 − 𝑥1 (𝑖) 𝑥 2 (𝑖) ]
𝑥1 ( 𝑖+1 ) =𝑥 1 ( 𝑖 ) +
𝑥 1 ( 𝑖 ) − 𝑥 2 (𝑖)

− 𝑥 2 ( 𝑖 ) [ 15 − 𝑥1 ( 𝑖 ) + 𝑥2 ( 𝑖 ) ] + [ 50 − 𝑥1 (𝑖) 𝑥 2 (𝑖) ]
𝑥 2 ( 𝑖+1 ) = 𝑥2 ( 𝑖 ) +
𝑥 1 ( 𝑖 ) − 𝑥 2 (𝑖)
§ Example 1.4-Solution

Successive calculations of these equations with and will yield the table below.

i 0 1 2 3 4 5
x1(i) 4.0000 5.2000 4.9913 5.0000 5.0000 5.0000
x2(i) 9.0000 9.8000 10.0087 10.0000 10.0000 10.0000
ITERATIVE SOLUTION TO NONLINEAR ALGERAIC
EQUATIONS:
THE NEWTON-RAPHSON METHOD
• Equation 7 can be rewritten as follows,
𝐱 ( 𝑖+1 ) − 𝐱 (𝑖)= 𝐉 −1 (𝑖) { 𝐲 − 𝐟 [ 𝐱 (𝑖) ] }

Letting and ,
−1
∆ 𝐱 (𝑖)= 𝐉 ( 𝑖) ∆ 𝐲 (𝑖)

Multiplying both sides by


𝐉 (𝑖)∆ 𝐱 (𝑖)=∆ 𝐲 (𝑖)
ITERATIVE SOLUTION TO NONLINEAR ALGERAIC
EQUATIONS:
THE NEWTON-RAPHSON METHOD
• Each iteration can then be summarized into the following steps,

Step 1: Compute .
Step 2: Compute .
Step 3: Using Gauss elimination and back substitution, solve from .
Step 4: Compute from .

• Experience from power flow studies has shown that Newton-Raphson


converges in many cases where Jacobi and Gauss-Seidel diverge.
§ Example 1.5
Redo Example 1.4 by completing the 4 step process.

Solution:

Step 1:
[ ] [
∆ 𝐲 ( 0 ) = 𝒚 − 𝐟 [ 𝐱 ( 0 ) ]= 1 5 −
50
4 +9
] [ ]
( 4) ( 9 )
= 2
14

Step 2:
𝐉 (0) =
[ 1
𝑥 2 (0)
1
] [
𝑥 1 (0)
=
1
9
1
4 ]
§ Example 1.5-Solution
Step 3:
.

[ 1
9
1
4 ][ ] [
∆ 𝑥1 ( 0 )
∆ 𝑥2 ( 0 )
=
2
14 ]
By Gauss elimination:

[ 1
0
1
−5 ][ ] [
∆ 𝑥 1 (0 )
∆ 𝑥 2 (0 )
=
2
−4 ]
4
∆ 𝑥 2 ( 0 )= − = 0.8
−5
∆ 𝑥 1 ( 0 )= 2− 0.8=1.2
§ Example 1.5-Solution
Step 4:
[ 9 ] [ 0.8 ] [ 9.8 ]
𝐱 ( 1 )= 𝐱 ( 0 ) + ∆ 𝐱 ( 0 ) = 4 + 1.2 = 5 .2

𝒙 𝟏 ( 𝟏 ) =𝟓 .𝟐

𝒙 𝟐 ( 𝟏 ) =𝟗 .𝟖
-end-
𝜕𝑓 1 𝜕𝑓 1
=2 =2 𝑥 2
𝜕 𝑥1 𝜕 𝑥2

𝜕𝑓 2 𝜕𝑓 2
=2 𝑥 1 + 𝑥2 =−2 𝑥2 + 𝑥 1
𝜕 𝑥1 𝜕 𝑥2

You might also like