Chapter 6.

Power Flow (Load Flow) Analysis (전력조류 해석)

- Steady-state circuital analysis of a power system in normal operation.
- Find V , I , S of all circuital elements (Generator, load, line).
- 발전기에서 생산된 전력은 송전 및 배전 선로를 통해 수용가에게 전송되어
소비되는데 이러한 전력의 흐름을 전력 조류 또는 조류라고 함.
- Backbone of power system analysis and design.
- Nonlinear static equations → Numerical technique

z Linear static (algebraic) equations

3x+4y=2 3 4 x 2
5x+6y=4 5 6 y 4

z Nonlinear static equations

3x2 +4y=2 # of variable = # of equation

2xy+ y =3
6.1 Gauss-Seidel Method of nonlinear algebraic equations.

z one variable nonlinear equation

f ( x) = 0 ⇒ x = g ( x)
f ( x*) = 0 ⇒ x* = g ( x*)

Assume initial x0
x1 = g ( x0 ) ⇒ x1 − x0 < ε ?
x2 = g ( x1 ) ⇒ x2 − x1 < ε ?
x3 = g ( x2 ) ⇒ x3 − x2 < ε ?
M M
xk +1 = g ( xk ) ⇒ xk +1 − xk < ε Yes!
→ x* = xk +1 ≈ xk

Update rule : xk +1 = g ( xk )
y y=x

x2 = g ( x1 )

x3 = g ( x2 )
x1 = g ( x0 ) y = g (x)

x
x1 x3 x* x4 x2 x0

y y=x

x1 = g ( x0 ) y = g (x)
x2 = g ( x1 )
x3 = g ( x2 )

x
x3 x2 x1 x0
 − 1 < g ( x)' < 1
g (x)

x3 x1 x0 x2

Ex 6.2 (Page 196)

f ( x) = x 3 − 6 x 2 + 9 x − 4 = 0 x* = ?

f ( x)' = 3( x 2 − 4 x + 9) = 3( x − 1)( x − 3) x'= 1,3

1 3 4

x* = 1,4
1 3

f (1) = 1 − 6 + 9 − 4 = 0
f (3) = 27 − 54 + 27 − 4 = −4

By Gauss – Seidel x0 = 2

x = g (x) 9x = − x3 + 6x 2 + 4
x3 6 2 4
x = − + x + = g ( x)
9 9 9

8 6 4
g ( x0 ) = g (2) = − + 4 + = 2.22222
9 9 9
g (2.22222) = 2.5173
g (2.5173) = 2.8966
3.3376
3.7398
3.9568
3.9988
 M

Fig 6.3 Page 197

z n-variable with n-equations

X = [ x1 , x2 , L , xn ]T

 f1 ( X ) = 0   f1 ( X ) 
   f ( X )
 f2 ( X ) = 0  Vector form
  F(X ) =  2 =0
M  M 
   
 f ( X ) = 0
 n   f n ( X )

 x1 = g1 ( X )   g1 ( X ) 
  Vector form
 g ( X )
 x2 = g 2 ( X ) 
  X = G ( X ), G ( X ) =  2 
M  M 
   
 x = g (X )
 n n   g n ( X )

z Update rule

X k +1 = G ( X k )

 x1( k +1) = g1 ( X k ) 
 
 x2( k +1) = g 2 ( X k ) 
x = g3 ( X k ) 
 3( k +1)  조금이라도 더 좋은 값으로 표현하기 위해서
 M 
 
 xn ( k +1) = g n ( X k )

→ Updated values are used immediately in subsequent equation.

z Stop condition
∆X k +1 ∆ ( X k +1 − X k ) → ∆X k +1 ≤ ε
n

Euclidean norm : ∆X k +1 = [∑ ( ∆xi ( k +1) ) ]

2 1/ 2

i =1

Sup norm : ∆X k +1 = max ∆xi ( k +1)

i
  f1 ( x1 , x2 ) = x13 + 10 x1 − x2 − 5 = 0
EX) 
 f ( x , x ) = x + x 3 − 10 x + 1 = 0
 2 1 2 1 2 2

X = [ x1 , x2 ]T , X 0 = [0.6,0.6]T

 f1 (0.6,0.6) = 0.616
Sol) With initial value →  , F ( X 0 ) = (0.616,−4.04)
T

 f 2 (0.6,0.6) = −4.04

 1 3 1 1
 x1 = − 10 x1 + 10 x2 + 2 = g1 ( x1 , x2 )

x = 1 x + 1 x 3 + 1 = g (x , x )
 2 10 1 10 2 10 2 1 2

Update rule: X k +1 = G ( X k )

 x1(1) = g1 (0.6,0.6) = 0.5384


 x2(1) = g 2 (0.5384,0.6) = 0.1754
∆X = max( 0.6 − 0.5384 , 0.6 − 0.1754 )
= max(0.0616,0.4246) = 0.4246

 x1( 2) = g1 (0.5384,0.1754) = 0.5019


 x2( 2) = g 2 (0.5019,0.1754) = 0.1507
∆X = max( 0.5384 − 0.5019 , 0.1754 − 0.1507 )
= max(0.0365,0.0247) = 0.0365
f1 (0.5019,0.1507) = −0.0053
f 2 (0.5019,0.1507) = 0.0176

 x1(3) = g1 (0.5019,0.1507) = 0.5024


 x2(3) = g 2 (0.5024,0.1507) = 0.1506
∆X = max( 0.5019 − 0.5024 , 0.1507 − 0.1506 )
= max(0.0005,0.0001) = 0.0005
f1 (0.5024,0.1506) = 0.0002
f 2 (0.5024,0.1506) = 0.0191

-> F ( X 3 ) = (0.0002,0.0191)
T
6.2 Newton – Raphson Method of nonlinear algebraic equations.

<Linearization of Nonlinear function>

z Taylor Series f (x) at x = a

∞ f ( n ) (a )
f ( x) = ∑ ( x − a) n
n=0 n!
( x − a) 2 ( x − a)3
= f (a) + ( x − a) f ′(a) + f ′′(a) + f ′′′(a) + L
2! 3!

f ( x, y ) at ( xo , yo )
∂f ∂f
f ( x, y) = f ( xo , yo ) + ( x − xo ) ( xo , yo ) + ( y − yo ) ( xo , yo ) + h.o.t.
∂x ∂y

Ex1) f ( x) = x + 2 x + 2 at x = 1
2

( x − 1) 2 ( x − 1) 3
f ( x) = f (1) + ( x − 1) f ′(1) + f ′′(1) + f ′′′(1) + L
2 6
f ′(1) = 2 x + 2 | x=1 = 4
f ′′(1) = 2 | x=1 = 2
f ′′′(1) = 0
( x − 1) 2
= 5 + ( x − 1) ⋅ 4 + ⋅2
2
= x2 + 2x + 2
→ 고차 미분 Term 이 Zero 가 아니면 무한대항까지!

Ex2) f (x) = x2 + 2 x + 2 at x = 2
( x − 2)2 ( x − 3) 2
f ( x) = f (2) + ( x − 2) f '(2) + f ''(2) + f ' ' ' ( 2 ) + ...
2 6
f ( 2) = 4 + 4 + 2 = 10
f ' ( 2) = 2 x + 2 | x = 2 = 6
f ' ' ( 2) = 2 | x = 2 = 2

( x − 2) 2
= 10 + ( x − 2)6 + x
x
= x2 + 2x + 2
z First order approximation with Taylor expansion

f ( x) ≅ f (a ) + ( x − a) f ' (a )
∂f ∂f
f ( x, y ) ≅ f ( x o , y o ) + ( x − x o ) | xo , y o + ( y − y o ) |x ,y
∂x ∂y o o

Ex3) f ( x ) = x 2 + 2 x + 2 at x = 2
f ( x ) = f ( 2) + ( x − 2) f ' ( 2)
= 6 x − 2 → Tangential line

Ex4) f ( x ) = x 2 + 2 x + 2 at x = 3
f ( x) = f (3) + ( x − 3) f ' (3) = 17 + ( x − 3)8
= 8x − 7

EX5) f ( x, y ) = 2 x 2 + 2 xy + y 2 at (1,2)
∂f ∂f
f ( x, y ) = f (1,2) + ( x − 1) (1, 2 ) + ( y − 2)
∂x ∂y
(1, 2 )

f (1,2) = 2 + 2 ⋅ 1 ⋅ 2 + 2 2 = 2 + 4 + 4 = 10
∂f
(1, 2 ) = 4 x + 2 y (1, 2 ) = 4 ⋅ 1 + 2 ⋅ 2 = 8
∂x
∂f
(1, 2 ) = 2 x + 2 y (1, 2 ) = 2 ⋅ 1 + 2 ⋅ 2 = 6
∂y
= 10 + 8( x − 1) + 6( y − 2)
> 가장 빨리 증가하는 방향 (접평면에 수직)
= 8 x + 6 y − 10 →Tangential plane

- linearization of a nonlinear function

비선형 함수의 선형화에 적용 !

< Newton-Raphson method>

z One variable nonlinear equation

f ( x) = c c : constant

Assume initial x0 and linearization at x0

 f ( x) ≅ f ( x0 ) + ( x − x0 ) f ′( x0 ) = c
1
( x − x0 ) = [c − f ( x0 )]
f ′( x0 )
1
x = x0 + [c − f ( x0 )] ∆ x1 , c − f ( x0 ) : error
f ′( x0 )

similarly,
1
x2 = x1 + [c − f ( x1 )]
f ′( x1 )

update rule
c − f ( xk )
xk +1 = xk +
f ′( xk )

x
x* x2 x1 x0
cf) limitation of Newton-Raphson

x
x* x1 x3 x2 x0

→ Initial point needs to be near the solution!

z Two variables nonlinear equations

f ( x, y ) = C1
g ( x, y ) = C2 with initial ( x0 , y0 )

Linearization at ( x0 , y0 )
∂f ∂f
f ( x, y ) ≅ f ( x0 , y0 ) + ( x − x0 ) + ( y − y0 ) = C1
∂x ∂y
( x0, y0 ) ( x0, y0 )

∂g ∂g
g ( x, y ) ≅ g ( x0 , y0 ) + ( x − x0 ) ( x0, y0 ) + ( y − y0 ) = C2
∂x ∂y
( x0, y0 )

∂f ∂f
( x − x0 ) + ( y − y0 ) = C1 − f ( x0 , y0 )
∂x ∂y
( x0, y0 ) ( x0, y0 )

∂g ∂g
( x − x0 ) ( x0, y0 ) + ( y − y0 ) = C1 − g ( x0 , y0 )
∂x ∂y
( x0, y0 )

 ∆x  ∆y

 ∂f ∂f 
 
 ∂x ( x0 , y 0 ) ∂y ( x  ∆x   c1 − f ( xk , yk ) 
0 , y0 ) 
   =  
 ∂g ∂g ∆
   2
y c − g ( x k k 
, y )
 ∂x ( x , y ) ∂y ( x , y ) 
 0 0 0 0 

= Jacobian matrix at ( x0 , y0 )
= J ( x0 , y0 )

 ∆x   C − f ( x0 , y 0 ) 
  = J −1 ( x0 , y0 ) 1 
∆
 
y  2
C − g ( x 0 0 
, y )

 x1   x0   C − f ( x0 , y 0 ) 
  =   + J −1 ( x0 , y0 ) 1 
 y1   y0   C 2 − g ( x0 , y 0 ) 
update rule
 xk +1   xk   c − f ( xk , yk ) 
  =   + J −1 ( xk , yk ) 1 
 yk +1   yk   c2 − g ( xk , yk ) 
z n-variables & n-equations

X = ( x1 x2 ⋅ ⋅ ⋅ ⋅ xn )

F(X ) = C X 0 = [x10 , x20, ⋅ ⋅ ⋅ ⋅ xn 0 ]

 f1 ( X ) = C1   f1 ( X )   C1 
     
 2
f ( X ) = C 2  f2 ( X )   C2 
 •   •  • 
  F(X ) =  , C = 
 •   •  • 
 •   •  • 
     
 f (X ) = C   f (X ) C 
 n n  n   n

update rule

X k +1 = X k + J −1 ( X k ){C − F ( X k )}
= X k − J −1 ( X k ){F ( X k ) − C}

 ∂f1 ∂f1 ∂f1 

 ⋅⋅⋅⋅⋅ 
 ∂x1 ∂x2 ∂xn 
 ∂f 2 ∂f 2 ∂f 2 
 ∂x ⋅⋅⋅⋅⋅
∂x2 ∂xn 
where, J ( X k ) =  1 
 M M ⋅⋅⋅⋅⋅ M 
 
 ∂f n ∂f n ∂f n 
 ⋅⋅⋅⋅⋅ 
 ∂x1 ∂x2 ∂xn 
X =Xk

z Stop condition
∆X k +1 ∆ ( X k +1 − X k ) → ∆X k +1 ≤ ε
n

Euclidean norm : ∆X k +1 = [∑ (∆xi ( k +1) ) 2 ]1/ 2

i =1

Sup norm : ∆X k +1 = max ∆xi ( k +1)

i

P. 205 Ex 6.5 6.6

Discussion) Why use the ∆X k +1 as a stop condition instead of F ( X k +1 ) − C ?

z Summary
Gauss – Seidel
・ Derivatives are not used → Relatively simple & fast calculation
・ Hard to converge & slow convergence
Newton-Raphson
・ Calculation of derivative & (n×n) matrix inversion
→ long time and not correct!
・ Good convergence & fast convergence
→ sparse matrix : A matrix has a few non zero elements.
・ sparse matrix technique with Newton- Raphson.

6.3. Y-Bus Matrix (Bus Admittance Matrix)

z 용어: Power System

= External Source + Transmission Network(Power Network)
- External Source: 발전기와 부하
- Transmission Network(Power Network): 송전선로로 이루어진 network부분

Transmission Network

External External
Source Source

External
Source

z Ii : 모선전류(Bus current)
: External Source (발전기/부하)에서 Transmission network으로 i- 모선을
통하여 유입되는 전류
- (발전기 혹은 부하)만 존재하는 모선
: (발전기 혹은 부하)에서 모선을 통하여 공급되는 전류
- 발전기와 부하가 동시에 존재하는 모선 : (발전기 전류+부하 전류)
- 발전기와 부하가 모두 없는 모선 : 모선전류는 zero.

ex1) a five-bus power system

 G2 L2 L5

2 I2 I5
5

I12 3
1 4
I1 I4
I13
G1 G4

I1 = I G1 ( = I12 + I13 )
I 2 = I G2 + I L2 ( = I 21 + I 23 )
I3 = 0 ( = I 32 + I 35 + I 51 )
I 4 = I G4 ( = I 45 )
I 5 = I L5 ( = I 53 + I 54 )

z Bus Matrix Equation

- A matrix equation which represents the bus current as a function of bus
voltage.
- Ohm’ s Law of entire power system.
- Nodal equation of a power system from KCL

I BUS = YBUSVBUS ,

 I1  V1 
I  V 
where, I BUS =  2 , VBUS =  2 ,
M M
   
I n  Vn 
n: number of bus
YBUS: Bus Admittance Matrix (n*n)

ex2) A 2-bus system with a short line

 I1 jX [Ω] I2
+ 1
+ I1 = (V1 − V2 )
jX
V1 V2
1
I2 = (V2 − V1 )
- - jX

 1 1   1 1 
−
 I1   jX jX  V1  − j X j
X  V1 
 = 1 
1  V2 
=  1 1  
 I 2  −  j −j  V2 
 jX jX   4X4244X3
1
Bus Admittance matrix Y BUS

ex3) A two bus system with a medium or long transmission line

① V1 V2 ②

G1 G2 L2

회로적으로 표현
I1 V1 V2 I2
ya

yb yb
G1 G2 L2

I1 = ybV1 + ya (V1 − V2 ) = ( ya + yb )V1 − yaV2

I 2 = ybV2 + ya (V2 − V1 ) = − yaV1 + ( ya + yb )V2

 I1   y a + y b − y a  V1 
  =   
 I 2   − yb y a + yb V2 
= YBUS
EX4) A three bus system

G1 G2
① ②

③
L3

I 1 V1 I2 V2
ya
G2
yα yb yb

yβ
G1 V3
I3
yβ
L3

KCL at bus ①,②,③

I1 = ybV1 + ya (V1 − V2 ) + y β V1 + yα (V1 − V3 )

= ( ya + yb + yα + y β )V1 − yaV2 − yα V3

I 2 = ybV2 + ya (V2 − V1 ) = − yaV1 + ( ya + yb )V2

I 3 = y β V3 + yα (V3 − V1 ) = − yα V1 + ( yα + y β )V2

 I1   ( y a + yb + yα + y β ) − ya − yα  V1 
    
 I2  =  − ya ( y a + yb ) 0  V2 
   − yα 0 ( yα + y β )  V3 
 I3  
z YBus of n-bus power system

 I1  Y11 Y12 L Y1n  V1 

    
 I 2  = Y21 Y22 L Y2n  V2 
M  M M L M  M 
    
 I n  Yn1 Yn 2 L Ynn  Vn 

n
I i = Yi1V1 + Yi 2V2 + L + YinVn = ∑ YijV j
j =1

Yii : Self-admittance of i-bus (diagonal)

sum of all admittance connected to i-bus
Yij : Mutual-admittance of i-j bus (off-diagonal)
Negative admittance between i-j bus

→ YBus is a symmetric matrix ( Yij = Y ji )

→ Yij = Y ji = 0 for no connection between i-j bus
(← impedance 는 ∞)
→ YBus has the information of network topology & transmission line parameters
→ Y-Bus matrix 와 모든 모선전압을 알면 임의의 모선 전류 표현 가능.

z Complex power of i-bus : Si

Si = Vi I i * = Pi + jQi
= SGi + S Li = ( PGi + PLi ) + j (QGi + QLi )

- Supplying power from i-bus to the network

- Sum of i-bus Generator & Load
- Zero for a bus has neither Generator nor load

z S1 in a two – bus system

2
S1 = P1 + jQ1 = V1 I1* = V1 ∑ (Y1 jV j )* = V1(Y11V1 + Y12V2 )*
j =1

Vi = Vi ∠δ i , δ ij ∆ δ i − δ j
Yij = Gij + jBij = Yij ∠θij
 { }*
2
= V1∠δ 1 ∑ (G1 j + jB1 j ) V j ∠δ j
j =1
2
= V1∠δ 1 ∑ (G1 j − jB1 j ) V j ∠ − δ j
j =1
2 2
= ∑V1V j (G1 j − jB1 j )∠(δ 1 − δ j ) = ∑ V1V j (G1 j − jB1 j )∠δ 1 j
j =1 j =1

{ }
2
= ∑V1V j (G1 j − jB1 j ) cos δ 1 j + j sin δ 1 j
j =1

{ }
2
= ∑V1V j G1 j cos δ 1 j + B1 j sin δ 1 j
j =1

{ }
2
+ j ∑V1V j G1 j sin δ 1 j − B1 j cos δ 1 j
j =1

2
P1 = ∑V1V j (G1 j cos δ 1 j + B1 j sin δ 1 j )
j =1
2
Q1 = ∑V1V j (G1 j sin δ 1 j − B1 j cos δ 1 j )
j =1

Similarly,
2
P2 = ∑V2V j (G2 j cosδ 2 j + B2 j sin δ 2 j )
j =1
2
Q2 = ∑V2V j (G2 j sin δ 2 j − B2 j cosδ 2 j )
j =1

→ Note : S 1 + S 2 ≠ 0 (∵transmission line)

z Si in n-bus power system (Power Balance Equation)

 
( )
n

 i ∑ i j ij
P = VV G cos δ ij + Bij sin δ ij 
 j =1
 i = 1, 2, 3, ⋅ ⋅ ⋅, n
 
( )
n

 Qi = ∑VV i j G ij sin δ ij − Bij cos δ ij 

 j = 1 

-> Power of External Source (발전기+부하) of i-Bus

= Power to the Network through i-Bus
 n

→ ∑S
i =1
i ≠0

cf) Power balance equation with Yij = Yij ∠θ ij

n
S i = Pi − jQi = (Vi I i ) * = Vi I i = Vi ∑ (Y V )
* * * *
ij j
j =1

{ }
n
= Vi ∠ − δ i ∑ Yij ∠θ ij V j ∠δ j
j =1
n n
= ∑ViYijV j ∠(θ ij − δ i + δ j ) = ∑ViYijV j ∠(θ ij − δ ij )
j =1 j =1
n
= ∑ViYijV j ∠(θ ij − δ ij )
j =1
n n
= ∑ViYijV j cos(θ ij − δ ij ) + j ∑ViYijV j sin(θ ij − δ ij )
j =1 j =1

 n n
 Pi = ∑ViYijV j cos(θ ij − δ ij ) = ∑ViYijV j cos(δ ij − θ ij )
j =1 j =1
 n n
i = 1,2,L, n
Qi = − ∑ViYijV j sin(θ ij − δ ij ) = ∑ViYijV j sin(δ ij − θ ij )
 j =1 j =1

Ex1) An elementary DC circuit.

4[V] 1[Ω] V 1[Ω] 5[V]

+ 1[A] 2[A] +
3[A]
4[V ] I1 1[Ω] I2 5[V ]

- -

Sol)
4 −V 5 −V V
Nodal: + =
1 1 1
9
9 − 2V = V , V = = 3 [V ]
3
  4 = I1 + ( I1 + I 2 ) = 2 I1 + I 2
Mesh: 
 5 = I 2 + ( I1 + I 2 ) = I1 + 2 I 2

I1 = 1, I 2 = 2

Supply power: I + V -
P4[V ] = 4 * 1 = 4[W ]
P5[V ] = 5 * 2 = 10[W ]
P1[Ω ]shunt = 3 * (−3) = −9[W ]
P1[ Ω]left = (4 − 3) * (−1) = −1[W ]
P1[Ω ]right = (5 − 3) * (−2) = −4[W ]

∑ Pi = 4 + 10 − 9 − 1 − 4 = 0

Ex2) Understanding of bus matrix eq. and power eq.

ex1) as a three bus DC power system (not a formal terminology)
with given VBUS = [4, 3, 5] , I BUS & Pi = ?
T

Bus1 Bus2 Bus3

1[S ] V2 1[S ] V3
V1
I1 I2 I3

Load
L2

 Y11 Y12 Y13   1 − 1 0 

   
YBUS =  Y21 Y22 Y23  =  − 1 2 − 1
Y   
 31 Y32 Y33   0 − 1 1 

From Bus matrix equation

 1 − 1 0   4  4 − 3   1 
       
I BUS = YBUS ∗ VBUS =  − 1 2 − 1 =  3  =  − 4 + 6 − 5  =  − 3 
 0 − 1 1   5  − 3 + 5   2 
       

From Power balance equation

 n n
Pi = ∑ViV j (Gi j cos δ ij + Bij sin δ ij ) = ∑ViV j (Gi j ⋅ 1 + 0 ⋅ 0)
j =1 j =1

in DC (δ ij = δ i − δ j = 0, θ ij = 0, Bij = 0)
n n
= ∑ViV jYij cos(δ ij − θ ij ) = ∑ViV j Yij cos(0 − 0)
j =1 j =1
n n
= ∑ViV j Gij = ∑ViV jYij
j =1 j =1

3
P1 = ∑ V1 V j Y1 j = V1 V1 Y11 + V1 V1 Y12 + V1 V1 Y13
j =1
= 4 ⋅ 4 ⋅ 1 + 4 ⋅ 3 ⋅ (−1) + 4 ⋅ 5 ⋅ 0 = 16 − 2 = 4 [W ]
3
P2 = ∑ V2 V j Y2 j = V2 {V1 Y21 + V2 Y22 + V3 Y33}
j =1
= 3{4 ⋅ (−1) + 3 ⋅ (2) + 5 ⋅ (−1)} = 3 ⋅ (−3) = −9 [W ]
3
P3 = ∑ V3 V j Y3 j = V3{V1 Y31 + +V2 Y32 + V3 Y33}
j =1
= 3{3 ⋅ 0 + 3 ⋅ (−1) + 5(1)} = 10 [W ]

⇒ P1 + P2 + P3 ≠ 0

transmission loss
PLEFT = y12 (V1 − V2 ) 2 = 1(4 − 3) 2 = 1 [W ]
PRIGHT = y23 (V2 − V3 ) 2 = 1(5 − 3) 2 = 4 [W ]

⇒ P1 + P2 + P3 + PL + PR = 4 − 9 + 10 − 1 − 4 = 0

(Conclusion)
* Y BUS → Network Topology + line parameters
* Given Y BUS & V BUS → I i , S i of all buses
→ Circuital analysis completed
* Finding out V BUS gives the complete circuital analysis.
6.4 Power flow (Load flow, 전력조류)

- steady-state circuital analysis of a power system

- nonlinear equations
load: S L = PL + jQL
generator: fixed PG & fixed VG = VG
← 발전기의 출력은 real power 와 전압의 크기.

Ex3) Given V1 = 4, P2 = -9, P3 = 10 of ex2)

find P1 = ?, V2 = ?, V3 = ?
← 어떤 모선은 전압, 어떤 모선은 전력이 주어짐!
← 단순한 nodal, mesh 로는 풀 수 없다!

V1 = 4 1[S ] V2 = ? 1[S ] V3 = ?

P1 = ? P2 = −9 P33 = −10

sol)

 1 −1 0 
YBus = − 1 2 − 1
 0 − 1 1 

Finding out V BUS gives the complete circuital analysis.

2 unknown voltage variables (V2, V3) ← 2 power equation at Bus 2 & 3

(power 를 알고 있는 모선에서 식을 만들 수 있다.)

Power equation at bus 2

3
P2 = V2 I 2 = V2 ∑ V jY2 j
j =1

= V2 (V1Y21 + V2Y22 + V3Y23 )

− 9 = V2 {4 ∗ (−1) + V2 ∗ (2) + V3 ∗ (−1)}
− 9 = V2 (2V2 − V3 − 4) -①

Power equation at bus 3

 3
P3 = V3 I 3 = V3 ∑ V jY3 j
j =1

= V3 (V1Y31 + V2Y32 + V3Y33 )

10 = V3{4 ∗ 0 + V2 ∗ (−1) + V3 ∗ (1)}
10 = V3 (V3 − V2 ) -②

①,② → Numerical technique <V2, V3>=<3,5>

←답을 알고 있기 때문에 확인 가능

VBUS = [4, 3, 5]T & YBUS give the other variables!

3
P1 = V1 I1 = 4 ∑ V1 Y j = 4(V1Y11 + V2Y12 + V3Y13 )
j =1

= 4{4 ⋅ 1 + 3(−1) + 5 ⋅ 0} = 4 [W ]

(4 − 3) 2
loss: P1[ Ω ]l = = 1 [W ]
R
(5 − 3) 2
P2[ Ω ]r = = 4 [W ]
R

Power conservation P1 + P2 + P3 + Pl + Pr = 4 − 9 + 10 − 1 − 4 = 0

- A circuital problem can be defined with given bus voltage and bus power.
← 어떤 모선은 전압, 어떤 모선은 전력이 주어짐.
-“ Power equations” are used to find unknown voltage variables.

EX4) Nonlinear eq.s does not always have the answer!

Given P1= 3[W] P2= -50[W] P3= 4[W]
Find V1= ? V2= ? V3= ?

sol) → Three power equations & 3 variables give <V1,V2,V3>

→ Power conservation can not hold because power losses are negative.

3 - 50 + 4 + loss < 0 (≠ 0)

P1 + P2 + P3 + Pl + Pr ≠ 0

→ No answer! (Numerical method is not converged!)

EX5) Nonlinear eq.s does not always have the answer!
Given P1 = 300[W] P2 = -50[W] P3 = 400[W]
Find V1 =? V2 =? V3=?
Sol)
A load 50[W]
Two generator 300+400=700[W]

→ three power equations & 3 variables

power balance gives
line loss = 700-50 = 650[W] !

→ Generally, no answer ! (Numerical method is not converged!)

How to avoid the situation of (EX4) & (EX5) ?

→ All generation power should not be given.
→ At least, one generator power should not be given: “ Slack generator”
→ Reasonable amount of generations considering the loads

Load flow is a Simulation of real situation to feed the loads.

EX6) Given P1 = 30[W] P2 = -50[W] P3 = ?

Find V1 =? V2 =? V3=10[V]

Slack generator P3 = 20(load) + loss

<Power flow formulation>

- Given YBUS & Load S L , and assumed Generator outputs ( PG & VG )

→ Network condition to feed the loads in the system.

- Firstly, find the voltage of all buses → the other variables.

z Type of buses

ⅰ) load bus (85%)

- a bus has no generator.
- given bus power S Li = PLi + jQLi
- find V i = Vi ∠δ i

cf) A bus has neither G nor L: S Li = 0

ⅱ) generator bus (voltage controlled bus,15%))

- a bus with generator.
- given PGi & VGi
- find QGi & δ Gi
cf) A bus has Generator and Load
S i = S Gi + S Li
= ( PGi + PLi ) + j (QG i + QLi )
cf) Note the sign of real power. (usually, positive convention)

ⅲ) slack bus (One bus in a power flow study)

- usually, the largest G bus
- Given V slack = 1∠0°
- Find S slack = Pslack + jQ slack
cf) V slack = 1∠0°
← per unit voltage base
← reference phase angle in a AC circuit problem

z Given Power balance equation

n
Pi = PGi + PLi = ∑ViV j (Gij cos δ ij + Bij sin δ ij )
j =1
n
Qi = QGi + QLi = ∑ ViV j (Gij sin δ ij − Bij cos δ ij ) , i = 1,L, n
j =1
z Problem formulation for n-bus power system

Given Unknown # of unknown

variables variables variables
Load Bus P, Q |V|, δ 2nL

Generator Bus P, |V| Q, δ 2nG

Slack Bus |V|=1, δ=0 P, Q 2

총모선수 : n = n g + nl + 1
where, nl : # of load bus
ng : (#-1) of G-bus

unknown voltage variables: (2* nl)+ ng

At load bus
n
PLi = ∑ ViV j (Gij cos δ ij + Bij sin δ ij )
j =1
n
QLi = ∑ ViV j (Gij sin δ ij − Bij cos δ ij ) i = 1, …, nl
i =1
→ 2* nl equations
At G-bus
n
Pi = PGi + PLi = ∑ ViV j (Gij cos δ ij + Bij sin δ ij ) i = 1, …, ng
j =1
→ ng equation

∴ ( 2nl + n g ) variables & ( 2nl + n g ) equations

→ All bus voltages V BUS by numerical technique

→ V BUS gives the complete circuital analysis.

z Other variables
→ Qi of G-bus
n
QGi + QLi = ∑ ViV j (Gij sin δ ij − Bij cos δ ij ) , i = 1, L , n g
j =1

→ S slack = Pslack + jQ slack

 n
PGi + PLi = ∑ ViV j (Gij cos δ ij + Bij sin δ ij )
j =1
n
QGi + QLi = ∑ ViV j (Gij sin δ ij − Bij cos δ ij )
j =1

If bus ① is slack
n
Pslack + PL1 = ∑V j {G1 j cos δ j + B1 j sin(−δ j )}
j =1
n
Qsalck + QL1 = ∑ V j {G1 j sin(−δ j ) − B1 j cos δ j }
j =1

→ I , S of transmission lines → YBus & VBus

Ex) Load flow analysis of following three bus power system

(Find PG1, QG1, δ2, QG2, |V3|, δ3 )

SG1 = PG1 + jQG1 PG 2 = 0.6661

V1 = 1∠0° V2 = 1.05∠δ 2

V3 = V3∠δ 3

S L3 = 2.8653 + j1.2244

− j19.98 j10 j10 

YBUS =  j10 − j19.98 j10 
 
 j10 j10 − j19.98

Sol) n = n g + nl + 1 (3 =1+1+1)
Unknown voltage variables : < δ 2 , V3 , δ 3 >
← ( 2nl + n g ) =(2 × 1+1)=3

Power Equation
 n n
PGi + PLi = ∑ViV j (Gij cosδ ij + Bij sin δ ij ) = ∑ViV j Bij sin δ ij
j =1 j =1
n n
QGi + QLi = ∑ViV j (Gij sin δ ij − Bij cosδ ij ) = − ∑ViV j Bij cosδ ij
j =1 j =1

There power equation with given P2 , P3 , Q3

3
・ PG 2 + PL 2 = ∑V2V j B2 j sin δ 2 j
j =1
= V2V1 B21 sin δ 21 + V2V2 B22 sin δ 22 + V2V3 B23 sin δ 23

P2 = 0.6661 + 0
= V2V1 B21 sin(δ 2 − δ1 ) + V2 2 B22 sin(0°) + V2V3 B23 sin(δ 2 − δ 3 )

= 1.05 * 10 * sin(δ 2 ) + 0 + 1.05 * V3 * 10 * sin(δ 2 − δ 3 ) - ①

3
・ PG 3 + PL3 = ∑V3V j B3 j sin δ 3 j
j =1
= V3V1 B31 sin δ 31 + V3V2 B32 sin δ 32 + V3V3 B33 sin δ 33
P3 = 0− 2.8653
= V3V1 B31 sin(δ 3 − δ1 ) + V3V2 B32 sin(δ 3 − δ 2 ) + 0
= V3 * 10 * sin(δ 3 ) + V3 * 1.05 * 10 * sin(δ 3 − δ 2 ) - ②

3
• QG 3 + QL 3 = − ∑V3VJ B3 j cos δ 3 j
j =1

= −(V3V1B31 cosδ 31 + V3V2 B32 cosδ 32 + V3V3 B33 cosδ 33 )

Q3 = 0− 1.2244
{
= − V3 * 1 * 10 cos δ 3 + V3 * 1.05 * 10 * cos(δ 3 − δ 2 ) + V3 2 * (−19.88) * 1 }
- ③

Three unknown variables ( δ 2 , V3 , δ 3 ) and Three equations

→ ( δ 2 , V3 , δ 3 )=(-3.0023°, 0.9502, -9.9924°) from numerical tech.

 1∠ 0° 
VBUS = 1.05 ∠ − 3.0023° 

 
0.9502 ∠ − 9.9924°
 The other variables
• PG1 = ?
3
PG1 + 0 = ∑V1V j B1 j sin δ1 j
j =1
=V1V1B1 sin δ11 + V1V2 B12 sin δ12 + V1V3 B13 sin δ13
= 0 + 1 * 1.05 * 10 * sin(3.0023°) + 1 * 0.9502 * 10 * sin(9.994°)
= 2.1987
• QG1 =?
3
QG1 + 0 = − ∑V1V j B1 j cos δ1 j
j =1
= − [ V1V1 B11 + V1V2 B12 cos δ12 + V1V3 B13 cos δ13 ]
= 0.1365

• QG 2 =?
3
QG 2 = − ∑V2 V j B2 j cos δ 2 j
j =1
= − 1.6395
Transmission loss = PG1 + PG 2 + PL 3
= 2.1987 + 0.6661 − 2.8653
= 0.0005
→ nearly zero (Q transmission line has no resistance )
I BUS = YBUS VBUS

<Procedure of power flow study>

1) Prepare Y Bus

2) Determine slack bus and Set V ( Slack ) = 1 ∠0°

3) Set S Li = PLi + QLi of Load buses

4) Set PGi , VGi of Generators ( i ≠ 1)

5) Apply G-S or N-R technique to find V Bus

→ Assume initial unknown voltage variable
 for numerical technique (usually Vi = 1∠0° )

6) Find the other circuital variables with V Bus and Y Bus

Discussion)
1. 수학적으로 정확한 식인가?
2. 답이 나오지 않거나 현실적으로 적절하지 않은 결과 (부하모선 전압, 1 번 발전기
출력)가 나오면?
< Power flow solution with Newton-Raphson >

n = ng + nl + 1
# of voltage variable : 2nl + n g

X ∆ [δ 2 , δ 3 ,⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅, δ n , V2 , V3 ,⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅⋅, Vn ]T
14442444 3 14442444 3
(nl + ng) (nl + ng)

2(nl + n g ) → 식 유도 후, 나중에 불필요한 변수 제외!

n
Pi = ∑ViV j (Gij cos δ ij + Bij sin δ ij ) ∆ Pi ( X )
j =1
n
= ∑ ViV jYij cos(δ ij − θ ij )
j =1
n
Qi = ∑ViV j (Gij sin δ ij − Bij cos δ ij ) ∆ Qi ( X )
j =1
n
= ∑ ViV jYij sin(δ ij − θ ij )
j =1

i = 2,3,⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅⋅, n

=> 2( nl + n g ) variables and 2( nl + n g ) equations

Given nonlinear equations

 P2 ( X ) − P2 = 0   ∆P2 ( X ) 
   
 P3 ( X ) − P3 = 0  ∆
 3P ( X ) 
M  M 
  vector form    ∆P( X ) 
 Pn ( X ) − Pn = 0     →  ∆Pn ( X )  ∆   ∆ F(X ) = 0

     ∆Q ( X ) 
∆
 Q2 ( X ) − Q 2 = 0   2Q ( X ) 
M  M 
Q ( X ) − Q = 0  ∆Q ( X ) 

 n n   n 

 ∆P( X ) (n − 1) × 1 
 
where,  ∆Q ( X ) (n − 1) × 1 
 F ( X ) 2(n − 1) ×1
 
update rule
X k +1 = X k − J −1 ( X k ){F ( X k ) − C}
= X k − J −1 ( X k ) F ( X k )
where,
 ∂P2 ( X ) ∂P2 ( X ) ∂P2 ( X ) ∂P2 ( X ) 
 L L 
 ∂ δ 2 ∂δ n ∂ V 2 ∂V n 
M M 
 
 n∂P ( X ) ∂P ( X ) ∂P ( X ) ∂P ( X ) 
 L n n
L n
∂F(X ) ∂δ 2 ∂δ n ∂V2 ∂Vn 
J(X ) = = 
∂X  ∂Q2 ( X ) L ∂Q2 ( X ) ∂Q2 ( X ) L ∂Q2 ( X ) 
 ∂δ 2 ∂δ n ∂V2 ∂Vn 
 
M M 
 ∂Qn ( X ) ∂Qn ( X ) ∂Qn ( X ) ∂Qn ( X ) 
 L L 
 ∂δ 2 ∂ δ n ∂V 2 ∂Vn 
 ∂P ∂P 
 | 
 ∂δ ∂V   J 11 J 12 
=  − | −  =  21 
 ∂Q ∂Q   J J 22 
 | 
 ∂δ ∂V 

∂P( X ) 12 ∂P( X ) 21 ∂Q( X ) 12 ∂Q( X )

J 11 = ,J = ,J = ,J = .
∂δ ∂V ∂δ ∂V

Let J pq designate the pq elements of J to avoid confusion ij in P( X ), Q( X ) .

∂Pp ( X )
11
i) J pq =
∂δ q
∂ n
( p ≠ q) = ∑V pV j (G pj cos δ pj + B pj sin δ pj )
∂δ q j =1

∂
= V pVq (G pq cos δ pq + B pq sin δ pq )
∂δ q {
= (δ p − δ q )

= V pVq {−G pq sin δ pq ⋅ (−1) + B pq cos δ pq ⋅ (−1)}

= V pVq (G pq sin δ pq − B pq cos δ pq )

∂Pp ( X ) ∂ n
( p = q) = = ∑V pV j (G pj cos δ pj + B pj sin δ pj )
∂δ p ∂δ p j =1
  V pV1 (G p1 cosδ p1 + B p1 sin δ p1 ) 
 O 
∂  
=  + V pV p (G pp cosδ pp + B pp sin δ pp ) 
∂δ p  
 O 
 + p n pn
V V (G cos δ pn + pn
B sin δ ) 
 pn 

 V pV1 (−G p1 sin δ p1 + B p1 cos δ p1 ) 

 O 
 
= +0 
 O 
 
 + V V
p n ( −G pn sin δ pn + B pn cos δ pn 
)

n
= ∑ V pV j (−G pj sin δ pj + B pj cosδ pj )
j =1, j ≠ p
n
= − ∑ V pV j (G pj sin δ pj − B pj cosδ pj ) Q p ( X )와 같은 form
j =1, j ≠ p

n
= − ∑ [V pV j (G pj sin δ pj − B pj cos δ pj ) + V pV p (G pp sin δ pp − B pp cos δ pp )]
j =1, j ≠ p 144444 42444444 3
=V pV p (G pp ⋅ 0 − B pp ⋅1)

+ V pV p (G pp ⋅ 0 − B pp ⋅ 1)

n
= −∑ V pV j (G pj sin δ pj − B pj cos δ pj ) − V p B pp
2

j =1

= −Q p ( X ) − V p B pp
2

∂Pp ( X )
12
ⅱ) J pq =
∂Vq
∂ n
( p ≠ q) = ∑V pV j (G pj cos δ pj + B pj sin δ pj )
∂Vq j =1
∂
= V pVq (G pq cos δ pq + B pq sin δ pq )
∂Vq
= V p (G pq cos δ pq + B pq sin δ pq )
∂Pp ( x) ∂ n
( p = q) = = ∑V pV j (G pj cos δ pj + B pj sin δ pj )
∂V p ∂V p j =1
 n
= ∑ V (G
j =1; j ≠ p
j pj cos δ pj + B pj sin δ pj )

∂
+ V pV p (G pp cos δ pp + B pp sin δ pp )
∂V p
n
= ∑ V j (G pj cos δ pj + B pj sin δ pj ) + 2V p (G pp )
j =1; j ≠ p
n
= ∑ V j (G pj cos δ pj + B pj sin δ pj ) + V p (G pp ⋅1 + 0) + V pG pp
j =1; j ≠ p
1 44444444424444444443
to complete ∑
n
= ∑ V j (G pj cos δ pj + B pj sin δ pj ) + V p G pp
j =1

1
=
Vp
∑ V V (G p j pj cos δ pj + B pj sin δ pj ) + V p G pp

Pp ( X )
= + V p G pp
Vp

∂Q p ( X )
21
iii) J pq =
∂δ q
∂ n
( p ≠ q) =
∂δ q
∑V V (G
j =1
p j pj sin δ pj − B pj cos δ pj )

∂
= V pVq (G pq sin δ pq − B pq cos δ pq )
∂δ q
= V pVq {G pq cos δ pq * (−1) + B pq sin δ pq * (−1)}
= −V pVq (G pq cos δ pq + B pq sin δ pq )
∂Q p ( X ) ∂Pp ( X )
cf) p ≠ q (off diagonal of J = −Vq
12
& J 21 ), ,
∂δ q ∂Vq
∂ n
( p = q) = ∑V pV j (G pj sin δ pj − B pj cos δ pj )
∂δ p j =1
n
= ∑ V V (G
j =1; j ≠ p
p j pj cos δ pj + B pj sin δ pj )
 ∂
+ V pV p (G pp sin δ pp + B pp cos δ pp )
∂δ p
144444424444443
=0
n
= ∑ V V (G cos δ pj + B pj sin δ pj ) + V pV p (G pp ⋅ 1 + 0) − V p G pp
2
p j pj
j =1; j ≠ p
n
= ∑ V pV j (G pj cos δ pj + B pj sin δ pj ) − V p G pp
2

j =1

= Pp ( x) − V p 2G pp

∂Q p ( X )
iv) J pq =
22
∂Vq
∂ n
( p ≠ q) = ∑V pV j (G pj sin δ pj − B pj cos δ pj )
∂Vg j =1
∂
= V pVq (G pq sin δ pq − B pq cos δ pq )
∂Vg
= V p (G pq sin δ pq − B pq cos δ pq )
∂ n
( p = q) = ∑V pV j (G pj sin δ pj − B pj cos δ pj )
∂V p j =1
n ∂
= ∑ V j (G pj sin δ pj − B pj cos δ pj ) + V p2 (0 − B pp )
j =1, j ≠ p ∂V p
14 42443
= −2V p B pp
n
= ∑ V j (G pj sin δ pj − B pj cos δ pj ) + V p (0 − B pp ) − V p B pp
j =1, j ≠ p
n
= ∑V j (G pj sin δ pj − B pj cos δ pj ) − V p B pp
j =1
Qp ( X )
= − V p B pp
Vp

cf) Jacobian with polar admittance

Yij = Gij + jBij = Yij ∠θ ij

 G pq = Y pq cosθ pq 
  → (6.55)~(6.62) P.234
 B pq = Y pq sin θ pq 
 
 ∂Pp ( X )
J 11 =
∂δ q
pq

( p ≠ q ) = V pVq (G pq sin δ pq − B pq cos δ pq )

= V pVq (Y pq cosθ pq sin δ pq − Y pq sin θ pq cos δ pq )

= V pVqY pq (cosθ pq sin δ pq − sin θ pq cos δ pq )

= V pVqY pq sin(δ pq − θ pq ) (6.56)

n
( p = q ) = − ∑V pV j (G pj sin δ pj − B pj cos δ pj ) − V p 2 B pp
j =1
n
= − ∑V pV j Y pj (cosθ pj sin δ pj − sin θ pj cos δ pj ) − V p2Y pp sin θ pp
j =1
n
= − ∑V pV j Y pj sin(δ pj − θ pj ) − V p2Y pp sin θ pp
j =1
n
= ∑V pV jY pj sin(θ pj − δ pj ) − V pV pY pp sin(θ pp − δ pp )
j =1
n
= ∑ V pV jY pj sin(θ pj − δ pj ) (6.55)
j =1, j ≠ p

Ex1) A Two Bus System,

I1 jX I2
+ +
V1 = V1 ∠ δ 1 V2 = V2∠ δ 2

- -

Sending Power
V1 − V2
S1 = V1 ⋅ I1* , I1 =
jX
*
 V ∠δ − V2 ∠δ 2 
= V1  1 1 
 jX 
  V1∠ − δ 1 − V2 ∠ − δ 2  V12 − V1V2 ∠(δ 1 − δ 2 )
= V1   =
 − jX  − jX

 V12 − V1V2 cos(δ 1 − δ 2 ) − jV1V2 sin(δ 1 − δ 2 ) 

= j 
 X 
 

V V sin(δ 1 − δ 2 )  V12 − V1V2 cos(δ 1 − δ 2 ) 

= 1 2 + j 
X  X 
 

V1V2 sin δ 12 V12 − V1V2 cos δ 12

= +j = P1 + jQ1
X X

Similary,
V2V1 sin δ 21 V2 2 − V2V1 cos δ 21
S 2 = V2 ⋅ I 2 *
= +j = P2 + jQ2
X X

∂P1 ∂  V1V2 sin δ 12  V1V2 cos δ 12

=  =
∂δ 1 ∂δ 1  X  X

∂P1 ∂  V1V2 sin δ 12  V V cos δ 12

=  =− 1 2
∂δ 2 ∂δ 2  X  X

∂P1 ∂  V1V2 sin δ 12  V2 sin δ 12

=  =
∂V1 ∂V1  X  X

Ex2) Result of ex1) from Jacobian.

 1 1 
 − j j
= X X 
1
YBUS
1
 j −j 
 X X

1 1 1 1
B11 = − , B12 = , B21 = , B22 = −
X X X X

∂P1 ( X )
= −Q p ( X ) − V p 2 B pp , ( p = 1)
∂δ 1
 2
= − ∑V1V j (G1 j sin δ1 j − B1 j cos δ1 j ) − V12 B11
j =1
2
= ∑V1V j B1 j cos δ1 j − V12 B11
j =1

= V1V1B11 cos δ11 + V1V2 B12 cos δ12 − V12 B11

123 {
=1 =1
X
V1V cosδ12
=
X

∂P1
= V pVq (G pq sin δ pq − B pq cos δ pq ), ( p = 1, q = 2)
∂δ 2

= −V1V2 B12 cos δ 12

V1V2 cos δ 12
=−
X

∂P1 ( X ) Pp ( X )
= + V p G pp , ( p = 1)
∂V1 Vp
2
= ∑V j (G1 j cos δ1 j + B1 j sin δ1 j ) + V1 G11
j =1

2
= ∑V j B1 j sin δ1 j
j =1

= V1 B11 sin δ 11 + V2 B12 sin δ 12

V2 sin δ 12
=
X
< Decoupled power flow >

z Jacobian matrix can be approximated simply.

 ∂P2 ( X ) ∂P2 ( X ) ∂P2 ( X ) ∂P2 ( X ) 

 L L 
 ∂δ 2 ∂δ n ∂V2 ∂Vn 
M M 
 
 ∂Pn ( X ) ∂Pn ( X ) ∂Pn ( X ) ∂Pn ( X ) 
L L
∂ F ( X )  ∂δ 2 ∂δ n ∂V2 ∂Vn   J
11
J 12 
J(X ) = =  =  21 
∂X  ∂Q2 ( X ) L ∂Q2 ( X ) ∂Q2 ( X ) L ∂Q2 ( X )   J J 22 
 ∂δ 2 ∂δ n ∂V2 ∂Vn 
 
M M 
 ∂Q ( X ) ∂Q ( X ) ∂Q ( X ) ∂Q ( X ) 
 n L n n
L n 
 ∂δ 2 ∂δ n ∂V 2 ∂Vn 
 J 11 0 
≅  
 0 J 22 

i) Transmission line is mostly reactive.

Gij ≅ 0, ( Bij ≠ 0)

ii) Voltage angle difference is small

δ ij << 10° → sin δ ij ≅ 0 (cos δ ij ≠ 0)

∂Pp ( x)
J 12
pq = = V p (G pq cosδ pq + B pq sin δ pq ) ≅ 0
∂Vq
∂Pp ( x) 1
= = ∑ V pV j (G pj cos δ pj + B pj sin δ pj ) + V p G pp ≅ 0
∂V p Vp

∂Q p ( X )
J 21 = = −V pVq (G pq cos δ pq + B pq sin δ pq ) ≅ 0
∂δ q
pq

∂Q p ( X ) n
= = ∑ V pV j (G pj cos δ pj + B pj sin δ pj ) − V p 2 G pp ≅ 0
∂δ p j =1

z Update rule can be decoupled.

X k +1 = X k − J −1 ( X k ){F ( X k ) − C}
= X k − J −1 ( X k ) F ( X k )
 where, X = [δ12 , δ 3 ,⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅, δ n , V2 , V3 ,⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅⋅, Vn ]
4442444 3 14442444 3
(nl + ng) (nl + ng)

-> X δ = [δ 2 , δ 3 ,⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅, δ n ] , X V = [V2 , V3 ,⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅⋅, Vn ]

T T

X δ ,k +1 = X δ ,k − ( J 11 ) −1 ( X δ ,k ) Fδ ( X δ ,k )
X V ,k +1 = X V ,k − ( J 22 ) −1 ( X V ,k ) FV ( X V ,k )

-> The decoupled power flow require more iteration, but less time per iteration.

z Analysis problem of Power system can be decoupled also !

P depends on δ dominantly.
Q depends on V dominantly.