Al-Balqa Applied

University

Dr audih 1
Load flow ( power flow ): Is the solution obtained for the power
system under steady state conditions of operation .
In the operation of a power system there are three major
problems encountered which are listed as below:-
1. Load flow problem
2. Optimal Load scheduling problem
3. Systems control problem
Why the load flow studies are important ?
The load flow investigation is important in order to determine :
1. The line flows (active & reactive power flow) .
2. The bus voltages and system voltage profile .
3. The phase angles of load bus voltages , reactive power at
generator bus .

Dr Audih alfaouri 2
4. For new circuits on system loading and load distributions .
5. Economic system operation and minimizing losses .
6. Possible improvements to an existing system by change
of conductor sizes and system voltages .

A single line diagram representation of the power is used

since the system is generally considered balanced.
The analysis are of analogue type, or for computer
changed from analogue to digital for simulation for large
number of algorithms ( large power systems) solution .

- The generators buses are provided with voltage magnitude

and phase angle controls.
- The loads buses are provided (or measuring) by voltages,
currents, and powers.

Dr Audih alfaouri 3
- The loads are generally represented by constant powers.
- In the network at each bus or node, there are four
variables: (voltage magnitude , voltage phase angle , real
power , reactive power ) .
Two of them are specified at each bus and the remaining two
must be calculate using the load flow solution.
Since the real and reactive power losses in lines are not
known till the end of the power flow solution, the generator bus
is selected as, called slack or swing bus.

For slack bus the generator voltage magnitude and phase

angle are specified and the other quantities are unknown .
For load buses (P) and the (Q) load demands are specified
and two unknown quantities to be determined δi and |Vi|.

Dr Audih alfaouri 4
 Bus Name Specified Variables Computed Variables

Voltage magnitude and its phase

Slack-bus Real and reactive powers
angle (δi and |Vi)|).

Generator bus(PV-bus or voltage Magnitude of bus voltages and real Voltage phase angle and reactive
controlled bus) powers Pgi and |Vi|. power

Magnitude of phase angle of bus

Load bus (PQ) Real and reactive powers
voltages

At each load bus, both power generated Pgi and Qgi are
zero, and the real power demand Pdi and reactive power
demand Qdi drawn from the system by the load (negative
inputs into the system).
 The scheduled values Pi,sch = Pgi - Pdi and Qi,sch = Qgi - Qdi
are known and mismatches ΔPi and Δqi can be defined.

Dr Audih alfaouri 5
Voltage-controlled buses ( generator bus):-
Any bus of the system at which the voltage magnitude is
kept constant is said to be voltage controlled.
Pgi and |Vi| are specify . With Pdi also known, to define
mismatch Δpi (losses).
Generator reactive power Qgi required supporting the
scheduled voltage |Vi| which is can’t be known .
And mismatch ΔQi is not defined after the power flow
problem is solved, Qi can be calculated.

**Note: ( At each bus to which there is a generator connected the MW

generation can be controlled by adjusting the prime mover, and the
voltage magnitude can be controlled by adjusting the generator
excitation).

Dr Audih alfaouri 6
1- Bus Admittance
Consider the transmission system shown in figure . The line
impedances at buses 1, 2, and 3 are z12, z23 and z13 .
y20
The corresponding line admittances G1 G2

are y12, y23 and y13. z12

The total capacitance susceptances
at the buses represented by y10, y20 z13 z23
and y30. Applying KCL at each bus
I 1  V 1 . y 10  (V 1  V 2 ). y 12  (V 1  V 3 ). y 13  G3
y30

 ( y 10  y 12  y 13 )V 1  (  y 12 )V 2  (  y 13 )V 3 1
yik 
Same for I 2 V 2 .y 20  (V 2 V1 ).y 21  (V 2 V 3 ).y 23 and I 3 zik
I3 V3.y30  (V3 V1).y31  (V3 V2 ).y32
Dr Audih alfaouri 7
In matrix form  I1   Y 11 Y 12 Y 13  V1 
 I   Y   
 2  21 Y 22 Y 23   V 2 
 I 3  Y Y 32 Y 33  V 
 31   3
For our example 3 buses ,we have:

I1   y10  y12  y13 y12 y13  V1 

I      
 2   y12  y 20  y12  y23   y23   V2
 
I3   y13 y23  y30  y13  y23 V3
where Y  y  y  y
11  10 12 13 
Y 22   y 20  y 21  y 23  are the self admittances forming the
diagonal terms
Y33   y30  y32  y31
Dr Audih alfaouri 8
 And
Y12  Y21   y12 
 are the mutual admittances forming
Y23  Y 23   y 23  the off-diagonal elements of the bus
Y13  Y31   y13  admittance matrix.

The injected current as a function of complex power and

voltage for bus 2 is:

 P2  jQ 2   P2  jQ 2 
* *
S2 
I2     *
 *
 Y 12 .V 1  Y 22 .V 2  Y 23 V
. 3
V
 2 V 2 V 2

The voltage of Bus 2 is:

1  P2  jQ2 
V2   *
 Y12 .V1  Y23 .V3  
Y22  V2 
Dr Audih alfaouri 9
 Example of Ybus
Determine the bus admittance matrix for the network shown
below, assuming the current injection at each bus i is
Ii = Igi - Idi :where Igi is the current injection into the bus from
the generator and Idi is the current flowing into the load

Dr Audih alfaouri 10
By KCL at bus 1we have
I1  Ig1  Id1
V1 V2 V1 V3 1
I1  I12  I13   , (with yj  )
ZA ZB Zj
I1  (V1 V2)yA (V1 V3)yB V1yA V2 yA V1yB V3 yB 
 (yA yB)V1  yA V2  yB V3
Similarly
I2  I21  I23  I24
yAV1  (yA yC  yD)V2  yC V3  yDV4

Dr Audih alfaouri 11
We can get similar relationships for buses 3 and 4
The results can then be expressed in matrix form
I  YbusV
 Y11
  Y12
 Y13
Y14

 I1   yA  yB   yA   yB   0  V1 
I   Y21  
 2     Y22
   Y23
 V2 
Y24
.
I3    yA   yA  yC  yD   yC   D  V3 
 y 
    y   yC   yB  yC   
 0  V4 
I
 4  B
  0  yD   0   yD 
*Note that (Y12 = Y21)
Dr Audih alfaouri 12
2- Modeling Shunts in the Ybus
IL=ΔV/Zk

IC

(V i V j ) Vi Y kc
Since I ij  I L  IC    (V i V j )Y k V i
Zk  Z kc / 2  2

1 1 R k  jX k R k  jX k
Note Y k    2
Z k R k  jX k R k  jX k R k  X k2

Dr Audih alfaouri 13
Two Bus System Example

(V V 2 ) Y
I1  1 V 1 c , y 1 0  y 2 0  j 0 .1
Z 2
1
y 12   1 2  j 1 6 , Y 11  y 10  y 12
0 . 0 3  j 0 .0 4
 I1   1 2  j (1 6  0 .1) 12  j 16  V 1 
I    V 
 2  12  j 16 1 2  j 1 5 .9   2 
Dr Audih alfaouri 14
1- Solving for Bus Currents
For example, in previous case for Ybus , assume V1  10o
and V 2  0.82  15.5o ,then :
 1.0  12  j 15.9 12  j 16 
V  and  
 0.8  j 0.2   12  j 16 12  j 15.9 
12  j 15.9 12  j 16   1.0   5.60  j 0.70 
 II1   
 2   12  j 16 12  j 15.9 0.8  j 0.2  5.58  j 0.88
Therefore the power injected at bus 1 is
S1 V 1I 1*  1.0  (5.60  j 0.70)  5.60  j 0.70 pu .
S 2 V 2I 2*  (0.8  j 0.2)  (5.58  j 0.88)  4.64  j 0.41 pu .
*
N o t e : 1 2  j 1 5 .9   (1 .0 )  (   1 2  j 1 6   ( 0 .8  j 0 .2 ) )  5 .6 0  j 0 .7 0
Dr Audih alfaouri 15
 2- Solving for Bus Voltages
For example, in previous Ybus case assume
 5.0 
I   , (the conjugate is not important since I  complex,then
4.8
1
V1  12  j15.9 12  j16  5.0   0.0738  j0.902 
V   12  j16 12  j15.9 4.8  0.0738  j1.098
 2      
Therefore the power injected is
S1 V1I1*  (0.0738  j0.902) 5  0.37  j4.51
S2 V2I2*  (0.0738  j1.098) (4.8)  0.35  j5.27

Dr Audih alfaouri 16
The variables and parameters associated with bus i and bus
k are represented in the usual notation as follows :
j o i
Vi  Vi    Vi ei
o
 Vi (cos  i o  j sin  io )
jiko
Yik  Yik   Yik e  Yik (cosiko  j sin iko )
o
ik
The Complex power (apparent power)is
*
  n n
S i  Pi  jQi  V I  Vi   Yik Vk   Vi  Yik*Vk*
*
i i
 k 1  k 1

The bus current is given by Ibus  Ybus .Vbus

Dr Audih alfaouri 17
Using the indexes (g) for generation and (d) for load
demand, (i) for bus :
Pi = Pgi - Pdi = Re Vi Ii*  Qi = Qgi  Qdi = Im Vi Ii* 
For an n-bus system:
*
 Si  n Pi  jQi n
For current I i  
V   Y ikV k  *
. i  y ik .V k
Y ii V
 i  k 1 Vi k 1
k i

1 Pi  jQ i n
For voltage V i  [ *
 Y ik V
. k ]
Y ii Vi k 1
k i

n n
For power S*i  P  jQ Vi*Ii Vi*YikVk Vi*Y ikVk
k 1 k 1
Dr Audih alfaouri 18
Further, In the polar form
n n
Pi  jQi  Vi  Yik .Vk   Vi .Vk .Yik e j (  i  k  ik )

k 1 k 1
k i
n
So that
Pi  k 1
V i .V k .Yik co s( i   k   ik )
n
Qi   V .V
k 1
i k .Yik sin ( i   k   ik )

Note: i =1,2,…,n i ≠ slack bus

Finally, the powers at the slack bus may be computed from
which the losses and all other line flows. Also one we select
the impedance matrix the equation rewrite as :
1
V Y .I  Z .I
Dr Audih alfaouri 19
There are a number of different iterative methods
we can use. We'll consider two: Gauss and Newton.
With the Gauss method we need to rewrite our
equation in form: x = h(x).
(0)
To iterate we first make an initial guess of x, x ,
and then iteratively solve x(v +1)  h (x (v ) ) until we
ˆ such that xˆ  h (x).
find a "fixed point", x, ˆ
Dr Audih alfaouri 20
Gauss Iteration mathematics example

Example: Solve x - x 1  0  x  x 1 or

 x (v 1)  1 x (v ) ,v is iteration nummber

Let k = 0 and arbitrarily guess x(0)  1 and solve
k x (v ) k x (v )
0  x (v 1)  1 0(v )  1(flat ) 5  2.61185
1  x (v 1)  1  1(v )  2 6  2.61612
2  x (v 1)  1 2(v )  2.41421 7  2.61744
3  x (v 1)  1  2.41421(v )  2.55538 8  2.61785
4  x (v 1)  1 2.55538(v )  2.59805 9  2.61798
Dr Audih alfaouri 21
The solution is iteration 7  2.617
x (7)  1  x (6)
x  2.617  1  1.617  1  2.617
Stopping Criteria
A key problem to address is when to stop the
iteration. With the Guass iteration we stop when
x ( v )   with x ( v )  x (v 1)  x ( v )
Where  is tolerance value

‫(المعطاة‬tolerances) ‫نقف عندما يكون الفرق ما بين الخطوة السابقة والجديدة اقل من قيمة‬

Dr Audih alfaouri 22
Dr Audih alfaouri 23
In this method:
The voltages at all buses except at the slack bus (since its
reference) are assumed ( etc..V2,V3,..)
The voltage at the slack bus is specified and remains fixed at
that value. The (n-1) bus voltage relations

1 Pi  jQ i n
Vi  [ *
  Yik .V k ]
Yii Vi k 1
k i

These equations are solved simultaneously for an improved

solution. The bus voltage equation for the (m+1) iteration
may then be written as:

1 P  jQ n n
Vi  [ (m)*  Yik .Vk  YikVk ]
(m1) i i (m1) (m)

Yii Vi k1 ki1

ki
Dr Audih alfaouri 24
Because of the loose mathematical coupling between the
buses is slowly .The rate of convergence of the process can
be increased by using acceleration factors to the solution
obtained after each iteration.
A fixed acceleration factor α (1≤ α ≤ 2) is normally applied to
each voltage change:
 S i*
 Vi   *
V i Y ii
(optional) :alternatively, different acceleration factors may be
used for real and imaginary parts of the voltage.

Vi i   Vi i
(m) (m)

Dr Audih alfaouri 25
Gauss Two Bus Power Flow Example

A 100 MW, 50 Mvar load is connected to a generator through

a line with z = 0.02 + j0.06 p.u. and line charging of 5 Mvar
on each end (100 MVA base). Also, there is a 25 Mvar
capacitor at bus 2. If the generator voltage is 1.0 p.u., what
is V2?
The charging is 5Mvar on base of 100=0.05pu
Q 0.05
and v=1pu  y 10 = = 2 =0.05 
V2 1
2
V 2
jy 10  j 0.05 and Q=  V .yC
xC

j0.05 pu line charging j0.05 pu line charging

SLoad = 1.0 + j0.5 p.u.

Dr Audih alfaouri 26
Since V1 is slack bus ,the value is 10o pu and the
unknown is the complex load voltage V2 .
To determine V2 we need to know the Ybus .
Y 11  y12  y10  5  j15  j 0.05  5  j (15  0.05)  5  j 14.95
1
Y 12 Y 21  y 12    5  j 15
0.02  j 0.06
Y 22  y 21  y 20  5  j15  ( j 0.05  j 0.25)  5  j (15  0.05  0.25)  5  j14.70

5  j 14.95 5  j 15 
Hence Ybus   
 5  j 15 5  j 14.70 

Dr Audih alfaouri 27
 At bus 2 there is no generation then , Pg2 =0 Pi  Pgi  Pdi  Pdi

1   P2 Q2  n 
V2     Y ikV k
 , where P2 andQ2 are100MW and 50Mvar load
Y 22  V 2* 

k 1,k i

on base of 100MW  P2  Pd 2 1 and Q2  Qd 2   j 0.5

  1  j 0.5 
V2 
1
  (5  j 15)(1.00)   Y12 Y21 5 j15

5  j 14.70  V2 * 
 Carefully for rounding
numbers decimal 4
digits
Guess V 2(0)  1.00 (this is known as a flat start(intial V2 )
v V 2(v ) v V 2(v )
0 1.000  j 0.000  flat star  3 0.9622  j 0.0556
1 0.9671  j 0.0568 4 0.9622  j 0.0556
2 0.9624  j 0.0553 3  4  0 then 4 is the solution
Dr Audih alfaouri 28
The solution is :V 21m  V 2 4  0.9622  j 0.0556
 0.9638 3.3
V 2 4  0.9638 and 2 4  3.3
Once the voltages are known all other values can be determined,
such as the generator powers and the line flows
S1* V1*.I1 V1*(Y 11V1 Y 12V 2 )  1.023  j 0.239
On base 100MW the actual units(S=P+jQ) gives (generation)
P1  102.3 MW, and Q1  23.9 Mvar
2 2
The capacitor is supplying Q= V . yC  0.9638  0.25  0.232 pu.
or in base of 100MW  Q=0.232 100=23.2Mvar
Dr Audih alfaouri 29
In this example the active power losses is Pg-PL
P  Pg  Pd 102.3100  2.3MW

and reactive power losses is :

Q  Qg  Qd
Qg  Qgeneration  Qcapacitor  Qcharging from line
Q  Qg  Qd  (23.9  23.2  5)  50  2.1MVAr
The total power losses is
S  P  jQ  2.3  j2.1MVA

Dr Audih alfaouri 30
Example:

Dr Audih alfaouri 31
 PLi=Pdi

Dr Audih alfaouri 32
Dr Audih alfaouri 33
Dr Audih alfaouri 34
 P=m iteration number

Dr Audih alfaouri 35
Dr Audih alfaouri 36
Dr Audih alfaouri 37
Dr Audih alfaouri 38
For second iteration

Dr Audih alfaouri 39
 3

V
3
Pi   V i .V k .Y ik cos( i   k   ik ) Qi  i .V k .Y ik sin( i   k   ik )
k 1 k 1
i 1 i 1

P1  V 1.V 1.Y 11 cos(11  1  1 )  V 1.V 2 .Y 12 cos(12  1  2 )  V 1.V 3.Y 13 cos(13  1  3 )

 V 12 Y 11 cos(11 )  V 1.V 2 .Y 12 cos(12  1  2 )  V 1V
. 3.Y 13 cos(13  1  3 ).
Same for Q but with sin and minus for Q since S  P  jQ

Dr Audih alfaouri 40
Dr Audih alfaouri 41
Dr Audih alfaouri 42
Dr Audih alfaouri 43
Dr Audih alfaouri 44
 P12 P21
∆P Line

Dr Audih alfaouri 45
Dr Audih alfaouri 46
Dr Audih alfaouri 47
N-R method is an algorithm which transforms nonlinear
equations into the procedure of iterative linear equations.
Consider the set of nonlinear equations as:
f i ( x1 , x2 ,......, x n )  yi i  1,2,...., n
Let x(0) be the initial value of the above equation. Assume the
real solution x is close to x(0)
xi  xi ( 0 )   xi ( 0 )
where ∆xi are the corrections to xi (i=1,2,…..,n) . When Δx is
known, the solution x can be calculated , the corrected set of
equations then are :

fi (x1 x1, x2 x2,......, xn xn )  yi

(0) (0) (0)

Dr Audih alfaouri 48
or expressed in matrix form, J . X   Y
where :
∆Y is a column vector determined from formula before as
y i  f i ( x 1( 0 ) , x 2( 0 ) ,......, x n( 0 ) )
∆x- is the column vector of correction terms ∆xi .
J- is the Jacobian matrix for the function (fi) given by the first
order partial derivatives evaluated… at xi(0). The corrected
solution is obtained as
xi(1) =xi(0) + ∆xi
The square Jacobin matrix J is defined by:
f i
J ik 
x k
Dr Audih alfaouri 49
Since most of the time in N-R iteration is dealing with
the Jacobian, speed up the iterations is calculate (and
factorize) the Jacobian occasionally known as the
“Dis-honest” Newton-Raphson method,

Honest: x(v 1)  x(v ) - J(x(v ) )-1 f (x(v ) )

Dis-honest: x(v 1)  x(v ) - J(x(0) )-1 f (x(v ) )
Stopping criterion f (x(v ) )   used in both cases.
an extreme example is to only calculate the Jacobian
for the first iteration is:

Dr Audih alfaouri 50
Example :
 x1 
S o lv e fo r x    s u c h th a t f ( x )  0 w h e re
 x2 
f 1 ( x )  2 x 12  x 22  8  0
f 2 ( x )  x 12  x 22  x 1 x 2  4  0
F irs t s y m b o lic a lly d e te rm in e th e J a c o b ia n
  f1 ( x )  f1 ( x ) 
 x x2 
J (x)   
1

 f2 ( x) f2 ( x) 
 x  x 2 
 1

Dr Audih alfaouri 51
  4 x1 2 x2  f i
J ( x)   J ik 
 2 x1  x2 x1  2 x2  x k

Then
1
 x1   x   4 x1(v )
2 x2   f1 ( x) 
 x    x 
1

 2    2 x1  x2
2
(v)
x1  2 x2   f 2 ( x) 
1
Arbitrarily guess x   
(0)
X(0) applied in
1 matrix
Stopping criterion
1
1  4 2   5  2.1 when f ( x(v ) )  
x (1)       
1  3 1  3 1.3 
X(0) applied in ( fx)
Dr Audih alfaouri 52
 1
2.1 8.20 2.60  2.51 1.8284
x  
(2)
    
1.3
   5.50  0.50  1.45  1.2122 
Each iteration we check f(x) to see if it is below
our specified tolerance  Appling this value in f(x) for x1
and x2 we get;

 2  (1.8284) 2
 (1.2122) 2
8 
f (x )  
(2)

(1.8284)  (1.2122)  (1.8284 1.2122)  4
2 2

0.1556
  , If  = 0.2 then we would be done.
0.0900
Dr Audih alfaouri 53
 The polar coordinate is used for solution of the load flow
using Newton-Raphson . The real and reactive power at bus
(i) are indicated before as:
n
Pi   V .V
k 1
i k .Yik cos(  i   k   ik )
n
Qi   V .V
k 1
i k .Yik sin(  i   k   ik )

The Jacobin is then formulated in terms of Δ|Vi | and δ.

Equation takes the form

 Pi   j1 j2     i 
 Q    j   V 
j4  
 i  3 i
Dr Audih alfaouri 54
For the two power relations (P&Q) at each bus and the line.
Equations can be written as:

  P2  P2  P2  P2 
  ..
 n V2 Vn 
 2

  Pi   j1 j2   
   P  Pn  Pn  Pn    
   n  i 
    2  n V2 Vn   
     
      
Qi   Q 2 Q2 Q2 Q2    Vi 
     
    2  n V2 Vn   
   j3 j4   
 
 Qn Qn Qn Qn 
   2  n V2  V n 

Dr Audih alfaouri 55
The off-diagonal and diagonal elements of matrix at bus (i)
and (k) are:

Pi
 ViV k Yik sin(  i   k   ik )
k
For j1 Pi n
  Vi .Vk .Yik sin( i   k ik )
 i k 1

 Pi
 V i Y ik cos(  i   k   ik )
 Vk
For j2
Pi n
 ViYii cos  ii   Vk .Yik cos( i   k   ik )
 Vi k 1
k i

Dr Audih alfaouri 56
 Q i
  V iV k Y ik cos(  i   k   ik )
k
For j3
Qi n

 i
  V .V
k 1
i k .Yik cos(  i   k   ik )
k i

Q i
 V i Y ik sin(  i   k   ik )
 Vk
and j4
Qi n

i ii sinii   Vk .Yik sin(i  k  ik )

  VY
 Vi k 1
k i

Dr Audih alfaouri 57
Decoupled power flow method is use for large-scale power
transmission systems to improving computational efficiency
and reducing computer storage requirements . The principle
is based on two observations :
(1)Change in voltage angle δ at a bus primarily affects the
flow of real power P in the transmission lines and leaves
the flow of reactive power Q respectively unchanged.

(2)Change in voltage magnitude at a bus primarily affects

the flow of reactive power Q in the transmission lines
and leaves the flow of real power P respectively
unchanged.

Dr Audih alfaouri 58
All power systems exhibit in the steady state a strong
interdependence between active powers and bus voltage
angles ( P ,  ) and between reactive powers and voltage
magnitude (Q , V ) .
The coupling between real power and bus voltage
magnitude and between reactive powers and voltage angles
are both relatively weak then (j2 and j3 are zero). This weak
coupling is utilized in the development of the so called
decoupled methods. Recalling equation

  P   j1 0    
 Q    0   
j4    V / V 
or
  
 P  j1 .  and  Q  j4 . V
Dr Audih alfaouri 59
Step 0: Formulate and assemble Ybus in per unit
Step 1: For load buses, where Pisch and Qisch are specified,
voltage magnitudes and phase angles are set equal to the
slack bus values or 1.0 and 0.0 i.e. (Assign initial guesses to
unknown voltage magnitudes and angles for a flat start
|V|=1.0, δ=0.0)
Step 2: For load buses, Pi(k) and Qi(k) are calculated. Also,
ΔPi(k) and ΔQi(k) are calculated.
Step 3: For voltage-controlled buses, Pi(k) and ΔPi(k) are
calculated
Step 4: Determine the element of Jacobian Matrix J for
Iteration k

Dr Audih alfaouri 60
Step 5: The linear simultaneous equation is solved directly.
Step 6: The new voltage magnitudes and phase angles are
computed.
Step7: The process is continued until the residuals ΔPi(k)
and ΔQi(k) are less than the specified accuracy, ΔPi(k) ≤ ε and
ΔQi(k)≤ ε

Dr Audih alfaouri 61
 n
Pi  k 1
V i .V k .Y ik co s(  i   k   ik )

2 n
Qi  
k 1
V i .V k .Y i k s i n (  i   k   ik )

Dr Audih alfaouri 62
Dr Audih alfaouri 63
Dr Audih alfaouri 64
Dr Audih alfaouri 65
Dr Audih alfaouri 66
Dr Audih alfaouri 67
Dr Audih alfaouri 68
Dr Audih alfaouri 69
Dr Audih alfaouri 70
Dr Audih alfaouri 71
Dr Audih alfaouri 72
all….

Dr Audih alfaouri 73