UNIT III POWER FLOW STUDIES

3.1 Introduction
The fast decoupled power flow method is a very fast and efficient method of obtaining power flow
problem solution. In this method, both, the speeds as well as the sparsity are exploited. This is actually an
extension of Newton-Raphson method formulated in polar coordinates with certain approximations which result
into a fast algorithm for power flow solution.
This method exploits the property of the power system where in MW flow-voltage angle and MVAR
flow-voltage magnitude are loosely coupled. In other words a small change in the magnitude of the bus voltage
does not affect the real power flow at the bus and similarly a small change in phase angle of the bus voltage has
hardly any effect on reactive power flow.

Because of this loose physical interaction between MW and MVAR flows in a power system, the MW-
δ and MVAR-V calculations can be decoupled. This decoupling results in a very simple, fast and reliable
algorithm. As we know, the sparsity feature of admittance matrix minimizes the computer memory
requirements and results in faster computations. The accuracy is comparable to that of the N-R method.

Load flow analysis is an important tool used by power engineers for planning and determining the
steady state operation of a power system. Power flow studies provide a systematic mathematical approach to
determine the various bus voltages, phase angles, active and reactive power flows through different branches,
generators, transformer settings and load under steady state conditions. The power system is modeled by an
electric circuit which consists of generators, transmission network and distribution network.

The admittance matrix of a power system is an abstract mathematical model of the system. It consists of
admittance values of both lines and buses. The Y-bus is a square matrix with dimensions equal to the number of
buses. This matrix is symmetrical along the diagonal. The values of diagonal elements (Y ii) are equal to the sum
of the admittances connected to bus i. The off-diagonal elements (Y ij) are equal to the negative of the
admittance connecting the two buses i and j.

Load flow studies are required for deciding the economic operation of the power system. They are also
required in transient stability studies. Hence, load flow studies play a vital role in power system studies. Thus
the load flow problem consists of finding the power flows (real and reactive) and voltages of a network for
given bus conditions.

The following expression is used to determine the reactive power of the system.
N
Qi cal = ∑ |V i||Y ij||V j|sin(θ ¿ ¿ij +δ j ¿−δ i)¿ ¿
j=1

For PV bus, If the value of reactive power limit violates, the following expression will be used to determine the
reactive power.

If Qi(min) ¿ Qi ¿ Qi(max), calculate Pi cal

If Qi cal ¿ Qi(min) , then Qi(spec) = Qi(min)

If Qi cal ¿ Qi(max), then Qi(spec) = Qi(max), the PV bus will act as PQ bus.

1
3.1.1 Sensitivity Analysis
Sensitivity analysis determines how different values of an independent variable affect a particular
dependent variable under a given set of assumptions. In other words, sensitivity analyses study how various
sources of uncertainty in a mathematical model contribute to the model's overall uncertainty.
Power systems exhibit dynamic behaviour which is governed by a mix of constrained continuous-time
dynamic phenomena. Analysis of this dynamic behaviour is vital in system design and operation. However the
nonlinear non smooth nature of dynamic behaviour generally precludes the use of simple systematic analysis
techniques. Power system operation therefore rely largely on simulation. The advantage of simulation is that it
is applicable for arbitrarily complicated models. A disadvantage is that it provides information about a single
scenario. Generally it is not possible to confidently extrapolate results, even for small changes in system
conditions. Each change to the system requires another simulation. For large systems, such as power systems,
this often involves a large computational cost. Trajectory sensitivity analysis offers some relief from the rigors
of repetitive simulation. The approach is based upon linearizing the system around a nominal trajectory, rather
than around an equilibrium point. It is therefore possible to determine directly the change in a trajectory due to a
(small) change in initial conditions and/or parameters. The analysis is straightforward for smooth systems, but
can also be applied to systems which contain discontinuities.
3.1.2 Decoupled Load Flow Methods (DLF):

An important characteristic of any practical electric power transmission system operating in steady state
is the strong interdependence between real powers and bus voltages angles and between reactive powers and
voltage magnitudes. This interesting property of weak coupling between P-δ and Q-V variables gave the
necessary motivation in developing the Decoupled Load Flow methods (DLF), in which P-δ and Q-V problems
are solved separately.

3.1.3 Decoupled Newton Methods:

In any conventional Newton method, half of the elements of the Jacobian matrix represent the weak
coupling referred to above, and therefore may be ignored. Any such approximation reduces the true quadratic
convergence to geometric one, but there are compensating computational benefits. A large number of decoupled
algorithms have been developed in the field of power flow analysis. However, only the most popular decoupled
Newton version is presented here.

The elements to be neglected are submatrices [N] and [J] are shown as follows.

The resulting decoupled linear Newton equations become

2
 [ ∆ P] = [ H ] [ ∆ δ ]

[ ∆ δ ] = [ L] [ ]
∆|V |
|V |

where it can be shown that

Hij = Lij = |V i| |V j| (Cij sin δ ij – Bij cos δ ij)

i≠j

and
2 2
Hii = - Bii |V i| −Qi , Lii = - Bii |V i| +Qi

Equations ΔP and ΔQ can be constructed and solved simultaneously with each other at each iteration,
updating the [H] and [L] matrices in each iteration using H ij, Hii and Lii. A better approach is to conduct each
iteration by first solving ΔP for Δδ, and use the updated δ in constructing and then solving ΔQ for Δ|V|. This
will result in faster convergence than in the simultaneous mode.

The main advantage of the Decoupled Load Flow Methods (DLF) as compared to the NR method is its
reduced memory requirements in storing the Jacobean. There is not much of an advantage from the point of
view of speed since the time per iteration of the DLF is almost the same as that of NR method and it always
takes more number of iterations to converge because of the approximation.

3.1.4 Fast Decoupled Load Flow (FDLF):

Further physically justifiable simplifications may be carried out to achieve some speed advantage without much
loss in accuracy of solution using the DLF model described in the previous subsection. This effort culminated in
the development of the Fast Decoupled Load Flow (FDLF) method by B. Stott in 1974. The assumptions which
are valid in normal power system operation are made as follows:

The earlier equation of power flow studies using N-R method can be written in polar coordinates as –

][ ]
∆δ
[ ] [
∆P
∆Q
=
H
M
N
L
∆V
V

where H, N, M and L are the elements (viz., J1, J2, J3 and J4) of the Jacobian matrix.

Since changes in real power (i.e., ΔP) are less sensitive to the changes in voltage magnitude (i.e., ΔV) and
changes in reactive power (i.e., ΔQ) are less sensitive to the changes in phase angle of voltage (i.e., Δδ), Eq. can
be reduced to

][ ]
∆δ
[ ] [
∆P
∆Q
=
H
0
0
L
∆V
V

3
The Equation is the decoupled equation and can be expanded as
∆P =H∆δ
∆V
∆Q = L
V
The next step in deriving the algorithm is to make suitable assumptions in deriving the expressions for H and L
as follows.
Cos δ ij ≈ 1;
Sin δ ij ≈ 0;
Gij Sin δ ij ≪ Bij
2
Qi ≪ Bii |V i|

With these assumptions, the entries of the [H] and [L] submatrices will become considerably simplified and are
given by
Hij = Lij = -|V i||V j|Bij for i ≠ j
and
2
Hii = Lii = - Bii |V i| for i ¿ j

Matrices [H] and [L] are square matrics with dimension (nPQ+nPV) and nPQ respectively.
Equations for ΔP and ΔQ can now be written as

[ ∆ P ] = [|V i||V j||B' ij|] [ ∆ δ ]

[ ∆ Q ] = [|V i||V j||B

''
|]
ij
[ ]
∆|V |
|V |

where B′ij, B″ij are elements of [- B] matrix.

Further decoupling and logical simplification of the Fast Decoupled Load Flow FDLF algorithm is achieved by:
Omitting from [B′] the representation of those network elements that predominantly affect reactive power
flows, i.e., shunt reactances and transformer off-nominal in-phase taps;
Neglecting from [B”] the angle shifting effects of phase shifters;
Dividing each of the Eqs. ΔP and ΔQ by |Vi| and setting |Vj| = 1 pu in the equations;
Ignoring series resistance in calculating the elements of [B′] which then becomes the dc approximation power
flow matrix.
With the above modifications, the resultant simplified FDLF equations become

[| |] [
∆P
V
= B' ][ ∆ δ ]

[| |] [
∆Q
V
= B '' ] [ ∆|V |]

4
 ∆P ∆Q
In Eqs. [ ] and [ ], both [B’] and [B”] are real, sparse and have the structures of [H] and [L], respectively.
|V | |V |
Since they contain only admittances, they are constant and need to be inverted only once at the beginning of the
study. If phase shifters are not present, both [B’] and [B”] are always symmetrical, and their constant sparse
upper triangular factors are calculated and stored only once at the beginning of the solution.
∆P ∆Q
Equations [ ] and [ ], are solved alternatively always employing the most recent voltage values. One
|V | |V |
iteration implies one solution for [Δδ], to update [δ] and then one solution for [Δ|V|] to update [|V|] to be called
1-δ and 1-V iteration. Separate convergence tests are applied for the real and reactive power mismatches as
follows:
Max [∆ P ] ≤ ε P;
Max |∆ Q|≤ ε Q
where ε P and ε Q are the tolerances.
In power flow analysis the decoupling strategy helps in strengthening the P–δ and Q–V relationships
based on line flow and voltage magnitude measurements in order to obtain a reliable convergence and higher
computational speed. The real and reactive set of measurements are combined using simple multiplying factors
such that the modified set is decoupled into two set of equations without making any assumption on r/x ratios.
The load flow equations are nonlinear algebraic equations and so explicit solution is not possible. The solution
of nonlinear equations can be obtained only by iterative numerical techniques. Direct current load flow gives
estimations of lines power flows on AC power systems. Direct current load flow looks only at active power
flows and neglects reactive power flows. This method is non-iterative and absolutely convergent but less
accurate than AC Load Flow solutions.
The fast decoupled power flow method is a very fast and efficient method of obtaining power flow problem
solution. The convergence of the fast decoupled load flow is unaffected by the starting voltage. Thus the flat
voltage start, that is, 1.0 p.u. volt for the load nodes and zero degree for all the angles is a good starting point in
the fast decoupled load flow solution. However, the method is very sensitive to the R/X ratio of transmission
lines.
3.2 POWER FLOW MODEL IN COMPLEX AND POLAR VARIABLE FORM.
The power flow or load flow model in complex form is obtained by writing one complex power matching
equation at each bus for the figure shown below.

Net power injected into the bus i.

Si = SGi – SDi

5
 = PGi + jQGi – (PDi + jQDi)
= Pi + jQi
We know,
Pi + jQi = Vi Ii*
Consider two bus system with I1 and I2 as net current entering into bus 1 and 2.

[ I ]= [ Y ] [ V ]

[][
I1
I2
= ][ ]
Y 11 Y 12 V 1
Y 21 Y 22 V 2

Y 11 = y 10 + y 12
Y 22 = y 20 + y 21
Y 12 = Y 21 = - y 21
In general, Yij = |Y ij|∠ θij
I1 = Y 11 V 1+ Y 12 V 2
I 2 = Y 21 V 1 + Y 22 V 2
In general , the net current entering into ith bus
N
Ii = Yi1V1 + Yi2V2+………………………………..+ YiNVN = ∑ Y ij V j
j=1

Substituting the value of Ii in power flow eqn we get.

Si = Pi + jQi = Vi Ii*
Si = Pi - jQi = Vi Ii*
N
Pi - jQi = Vi*∑ Y ij V j where i = 1,2,3…………..N
j=1

There are N complex variable equations for which the N unknown complex variables V 1,V2…….VN can be
determined.
Substituting Yij from the above eqn, we get.
N
Pi - jQi = Vi*∑ |Y ij|∠ θij V j
j=1

Where Vi = |V i|∠ δ i, Vi*= |V i|∠−δ i,

Vj = |V j|∠ δ j
Therefore net power equations can be written as
N
Pi - jQi = ∑ |V i||Y ij||V j|∠(θ ¿ ¿ij +δ j−δ i )¿
j=1

6
Equating real and reactive parts,
N
Pi = ∑ |V i||Y ij||V j|cos (θ ¿ ¿ ij+δ j −δ i ) ¿
j=1
N
Qi = - ∑ |V i||Y ij||V j|sin(θ ¿ ¿ij +δ j−δ i )¿
j=1

We can write the above equation as

N
Pi = |V i| |Y ii|cos θii ∑ |V i||Y ij||V j|cos (θ ¿ ¿ ij+ δ j −δ i)¿
2

j=1
N
Qi = - |V i| |Y ii|sin θii ∑ |V i||Y ij||V j| cos(θ ¿ ¿ ij+δ j−δ i )¿
2

j=1

The above equations are called as polar form of the power flow equations.

3.3 STEP-BY-STEP ALGORITHM TO SOLVE THE LOAD FLOW PROBLEM USING FAST
DECOUPLED METHOD.
Step 1 : Form Y-Bus and then compute bus susceptance matrices B’ and B’’
Step 2 : Assume flat start for starting voltage solution
0
δ i = 0, for i = 1…..N ( for all bus except slack bus )

|V 0i |=1.0, for i = M+1……….N (for all PQ buses)

|V i| = |V i|spec for all PV bus and slack bus.

Step 3 : for load buses, calculate Pi cal and Qi cal
N
Pi cal = ∑ |V i||Y ij||V j|cos (θ ¿ ¿ ij+δ j ¿−δ i )¿ ¿
j=1

N
Qi cal = ∑ |V i||Y ij||V j|sin(θ ¿ ¿ij +δ j ¿−δ i)¿ ¿
j=1

Step 4 : for PV bus, Check for Q limit violation

If Qi(min) ¿ Qi ¿ Qi(max), calculate Pi cal
If Qi cal ¿ Qi(min) , then Qi(spec) = Qi(min)
If Qi cal ¿ Qi(max), then Qi(spec) = Qi(max), the PV bus will act as PQ bus.
Step 5 : Compute mismatch vector using.
∆ Pi = Pi(spec) - Pi cal
∆ Q i = Qi(spec) - Qi cal
Step 6 : Compute ∆ Pi(max) = max |∆ P i| i = 1,2……….N ; ≠ except slack

∆ Qi(max) = max |∆ Qi| i = M+1……….N

7
 [ ]
∂ Pi ∂ Pi
∂δ ∂|V |
Step 7 : compute jacobian matrix using J =
∂ Qi ∂Q i
∂δ ∂|V |

Step 8 : Calculate ∆ δ∧∆ V using

[| |]
[ ∆ δi ] = -[ B' ]-1 .
∆ Pi
Vi

] [
| |]
∆ Qi
[ ∆ V i ] = -[ B' ' -1 . Vi

Step 9 : Update state vector using .

Vnew = Vold +∆ V
δ new = δ old + ∆ δ
Step 10 : This procedure is continued until
|∆ P i| ¿ ε and |∆ Qi| ¿ ε ,
Otherwise go to step 3.

8
9
Example 1: The specification of the buses is given in the table 1 and line impedances in table 2. Determine
bus admittance matrix. Assume that all the buses other than slack bus are PQ type. Draw the required
bus figure and by taking a flat voltage profile, determine the bus voltages at the end of iteration.
Table 1
Bus P Q V Remarks
1 - - 1.05∟0o Slack bus
2 0.5 -0.2 - PQ bus
3 -1.0 0.5 - PQ bus
4 0.3 -0.1 - PQ bus

Table 2

Line R (in p.u) X (in p.u)

1-2 0.05 0.15

1-3 0.10 0.30

1-4 0.20 0.40

2-4 0.10 0.30

3-4 0.05 0.15

Solution:

Convert the line impedances to line admittances. The line admittances are given by the inverse of the line
impedances.

Z12 = 0.05+j0.15 p.u

Z13 = 0.10+j0.30 p.u

Z14 = 0.20+j0.40 p.u

Z24 = 0.10+j0.30 p.u

Z34 = 0.05+j0.15 p.u

10
 [ ]
y 11 y 12 y 13 y 14
y 21 y 22 y 23 y 24
Y bus = y 31 y 32 y 33 y 34
y 41 y 42 y 43 y 44

[ ]
4− j 11 −2+ j 6 −1+ j 3 −1+ j 2
−2+ j 6 3− j 9 0 −1+ j 3
=
−1+ j 3 0 3− j 9 −2+ j 6
−1+ j 2 −1+ j 3 −2+ j 6 4− j 11

The initial value of load bus voltages can be assumed as 1+j0 p.u
V20 =1+j0 ; V30 = 1+j0 ; V40 = 1+j0
The slack bus (bus 1) voltage is not modified in any iteration
V10 = V11 = ……. = V1n= V1 = 1.05+j0 p.u
The load bus voltages can be calculated using the following equation for the first iteration k=0. Here the table
9.1 provides the net power and so the values of P and Q are used as suchin the following equation.

Vpk+1 =
Y pp [
1 P p− jQ p
Vp k*
−∑ Y pq Vq k+1−∑ Y pq Vpq k
]
V10 = V11=1.05+j0 (bus 1 is a slack bus)

V21 = [
1 P 2− jQ 2
Y 22 V 2 o*
−Y 21V 1 1−Y 23 V 3 0 −Y 24 V 4 0
]
1 0.5+ j 0.2
= [ −2+ j 6 ( 1.05+ j0 )−0 × ( 1+ j0 )−1+ j 3 ( 1+ j 0 ) ]
3− j9 1− j 0
= 1.03163.15
= 1.0300+ j0.0567 p.u

V31 = [
1 P 3− jQ 3
Y 33 V 3 o*
−Y 31 V 11−Y 32 V 2 0 −Y 34 V 4 0
]
1 −1− j 0.5
= [ −(−1+ j3 )( 1.05+ j0 )−( 0 × )−(−2+ j6 )( 1+ j0 ) ]
3− j9 1− j 0
=1.0399-6.44
= 1.0333-j0.1166 p.u

V41 = [
1 P 4− jQ 4
Y 44 V 4 o*
−Y 41 V 11−Y 42V 2 1 −Y 43 V 3 1
]
1 0.3+ j 0.1
= [ −(−1+ j 2 ) (1.05+ j 0 )−(−1+ j3 )( 1.03+ j0.0567 )−(−2+ j6 ) (1.0333− j 0.1166 ) ]
4− j 11 1− j 0
= 1.0352-1.12
= 1.0350 – j0.0202 p.u
The bus voltages at the end of Gauss-Seidal iteration are

11
 V11=1.05+j0 = 1.050
V21 = 1.0300+ j0.0567 p.u = 1.03163.15
V31= 1.0333-j0.1166 p.u = 1.0399-6.44
V41=1.0350 – j0.0202 p.u = 1.0352-1.12
Jacobian matrix
The matrix formed from the first derivatives of load flow equation is called jacobian
matrix and it is denoted by J.
The elements of jacobian matrix will change in every iteration. The elements of the jacobian matrix are
obtain matrix are obtained by partially differentiating the load flow equation with respect to a unknown
variable and then evaluating the first derivative as using the solution of previous iteration.
Example 2: Using fast decoupled load flow method, determine bus voltage for the system as shown in fig.

perform one iteration.

Bus No. PL QL
1 90 20
2 30 10

Solution:
Step -1 : Form Ybus matrix

Ybus = [−0.3044+ j1.88 0.3043− j 1.817 ]

0.3043− j1.817 −0.3044 + j 1.88

=[
1.904 1.7314 1.842−1.405 ]
1.842−1.405 1.904 1.7314
Ybus {Note: Use in rad mode}

Step – 2:Initialize bus voltages

V1old = 1.050 p.u. (slack bus)

V2old = 1.00 p.u. (P-V bus)
Step – 3: Check for Q-limit violation.
P2cal = |V 2|{|V 1||Y 21|cos ( 21+1−2 )+|V 2|2|Y 22|cos ⁡(22) }

12
 = 1.02[1.05 X 1.904 cos(1.7314) + 1.022 X 1.842 cos(-1.405)]
P2cal = -0.0098 p.u.
Q2cal = - {|V 2||V 1||Y 21|sin ( 21+ 1−2 ) +|V 2|2|Y 22|sin ( 22+2−2 ) }
= - 1.02[1.05 X 1.904 sin(1.7314)+ 1.022 X 1.842 sin(-1.405)]
Q2cal = - 0.1228 p.u.
−20 100
←0.1228<
100 100
Q2(min) ¿Q2 < Q2(max)
Bus 2 acts as PV bus.
Step – 4: Compute P and Q.
30
P2(spec) = PG2 – PL2 = 0- = -0.3 p.u.
100
P2 = P2(spec) – P2cal = -0.3 – (-0. 0098) = - 0.2902
10
Q2(spec) = QG2 – QL2 = 0- = -0.1 p.u.
100
Q2 = Q2(spec) – Q2cal = 0.1 – (-0.1228) = 0.0228

Step – 5: Bus susceptance matrix

[B] = [-1.817]
1
[B]-1 = = -0.5504
−1.817
[B] = -1.817
1
[B]-1 = = -0.5504
−1.817
NOTE:
B matrix is the imaginary part of Ybus for the buses except slack bus.
Step – 6: Calculate 

[ ] [ ]
[ 2 ] =−[ B ] -1 P 2 = -(-0.5504) −0.2902 = -0.1597
|V 2| 1.0

] [
| |] [ ]
[ V 2 ] =−[ B -1 Q 2 = -(-0.5504) 0.0228 = 0.0125 p.u.
V2 1.0
21 = 20 + 2 = 0+(-0.1597) = -0.1597rad = -9.150
V21 = V20 + V2
= 1.0 + 0.0125 = 1.0125 p.u
V2new = 1.0125-9.150

13
Example 3: Perform two iterations of decoupled load flow method and determine the power flow solution
for the system as shown in Fig. Take base MVA as 100.

Solution:
Step -1 : Form Ybus matrix

Ybus = [−0.3044+ j1.88 0.3043− j 1.817 ]

0.3043− j1.817 −0.3044 + j 1.88

=[
1.904 1.7314 1.842−1.405 ]
1.842−1.405 1.904 1.7314
Ybus {Note: Use in rad mode}

Step – 2:Initialize bus voltages

V1old = 1.050 p.u. (slack bus)

V2old = 1.020 p.u. (P-V bus)
Step – 3: Check for Q-limit violation.
Q2cal = - {|V 2||V 1||Y 21|sin ( 21+ 1−2 ) +|V 2|2|Y 22|sin ( 22+2−2 ) }
= - 1.02[1.05 X 1.904 sin(1.7314)+ 1.022 X 1.842 sin(-1.405)]
Q2cal = - 0.1228 p.u.
10
Q2(min) = 10 MVAR = p.u. =0.1 p.u.
100
Q2 < Q2(min); Q2 = Q2(min) = 0.1 p.u.
and the bus 2 will acts as load bus, V2 = 10 p.u.
Step – 4: Compute P and Q.
P2cal = |V 2|{|V 1||Y 21|cos ( 21+1−2 )+|V 2|2|Y 22|cos ⁡(22+2−2) }
= 1.0[1.05 X 1.904 cos(1.7314) + 1.02 X 1.842 cos(-1.405)]
= -0.0157 p.u.
60
P2(spec) = PG2 – PL2 = = 0.6 p.u.
100
P2 = P2(spec) – P2cal = 0.6 – (-0.0157)= - 0.6157
Q2 = Q2(spec) – Q2cal = 0.1 – (-0.1) = 0.2

14
Step – 5: Bus susceptance matrix
2
[B] = 2 [1.817]
1
[B]-1 = = -0.5504
2 −1.817
[B] = 2 [1.817], [B]-1 = -0.5504
NOTE:
B matrix is the imaginary part of Ybus for the buses except slack bus.
B matrix is the imaginary part of Ybus for the load buses.
Step – 6: Calculate  and V.

[ ] [ ]
[ 2 ] =−[ B ] -1 P 2 = -(-0.5504) 0.6157 = 0.34
|V 2| 1.0

] [
| |] [ ]
[ V 2 ] =−[ B -1 Q 2 = -(-0.5504) 0.2 = 0.1226 p.u.
V2 1.0
21 = 20 + 2 = 0+0.34 = 0.34 rad
V21 = V20 + V2 = 1.0 + 0.1226 = 1.1226 p.u.
V21 = 1.12260.34
Iteration – 2: Check for Q-limit
Q2cal = - [1.1226 X 1.05 X 1.904 sin(1.7314 – 0.34)+ 1.12262 X 1.842 sin(-1.405)]
= - (-0.0812) =0.0812
0.0812 < 0.1, Q2cal<Q2(min)
Q2 = Q2(min) = 0.1 p.u. MVAR
Bus 2 again act as load bus. V2old = 1.12260.34

P2cal = 1.1226 X 1.05 X 1.904 cos(1.7314 – 0.34) + 1.12262 X 1.842 cos(-1.405)

= 0.7836 p.u.
60
P2(spec) = PG2 – PL2 = = 0.6 p.u.
100
P2 = P2(spec) – P2cal = 0.6 – (0.7836)= - 0.1836
Q2 = Q2(spec) – Q2cal = 0.1 – 0.0812 = 0.0188
Step – 5: Bus susceptance matrix
[B] = [B] = [1.817]
[B]-1 = [B]-1 = -0.5504
Step – 6: Calculate  and V.

[ ] [
[ 2 ] =−[ B ] -1 P 2 = -(-0.5504) −0.1836 = - 0.09
|V 2| 1.1226 ]
15
 [ ] [ ]
[ V 2 ] =−[ B ] -1 Q 2 = -(-0.5504) 0.0188 = 0.0092
|V 2| 1.1226
22 = 21 + 2 = 0.34+(-0.09) = 0.25 rad =14
V22 = V21 + V2 = 1.1226 + 0.0092 = 1.1318
V21 = 1.131814

Example 4: Obtain the power flow solution(one iteration) for the system shown in Fig. The line
admittances are in per unit on a 100MVA base. Use fast decoupled load flow method.

Solution:
Step -1 : Form Ybus matrix

[ ]
Y 12+Y 13 −Y 12 −Y 13
Ybus = −Y 12 Y 12+Y 23 −Y 23
−Y 13 −Y 23 Y 13+ Y 23

[ ]
− j 20+(− j 10) j 20 j10
Ybus = j20 − j 20+ j10 j10
j10 j 10 − j 10+(− j10)

[ ]
30−1.57 20 1.57 10 1.57
Ybus = 201.57 30−1.57 10 1.57 {Note: Use in rad mode}
101.57 10 1.57 20−1.57

Step – 2: Initialize bus voltages

V1old = 1.050 p.u. (slack bus)

`V2old = 1.020 p.u. (P-V bus)
V3old = 1.00 p.u. (P-Q bus)
Step – 3: Check for Q-limit violation for Q bus.

16
Q2cal = - {|V 2||V 1||Y 21|sin ( 21+ 1−2 ) +|V 2|2|Y 22|sin ( 22 ) +|V 2||V 3||Y 23|sin ( 23+3−2 ) }
= - 1.02[1.05 X 20 sin(1.57-0+0)+ 1.022 X 30 sin(-1.57)+1.02 X 1.0 X 10 X sin(1.57-0+0)]
Q2cal = - 0.408 p.u.
−50 100
< -0.408 <
100 100
Q2(min) < Q2 < Q2(max), Bus 2 acts as P-V bus.
Step – 4: Compute P and Q.
300
P2(spec) = PG2 – PL2 = = 3 p.u.
100
400
P3(spec) = PG3 – PL3 = 0 - = -4 p.u.
100
200
Q3(spec) = QG3 – QL3 = 0 - = - 2 p.u.
100
P2cal = |V 2|{|V 1||Y 21|cos ( 21+1−2 )+|V 2|2|Y 22|cos ( 22 ) +|V 2||V 3||Y 23|cos ( 23+ 3−2 ) }
= 1.02 X 1.05 X 20 cos(1.57-0+0) + 1.022 X 30 cos(-1.57) + 1.02 X 1.0 X 10 X cos(1.57 – 0 + 0)]
P2cal = 0.05
P3cal ={|V 3||V 1||Y 31|cos ( 31+1−3 )+|V 3||V 2||Y 32|cos ( 32 - 3+ 2 ) +|V 3||V 3||Y 33|cos ( 33 ) }
= 1.0 X 1.05 X 10 cos(1.57-0+0) + 1.0 X 1.02 X 10 cos(1.57– 0 + 0) + 1.02 X 20 X cos(-1.57)]
P3cal = 0.0324

Q3cal = - {|V 3||V 1||Y 31|sin ( 31+ 1−3 ) +|V 3||V 2||Y 32|sin ( 32 - 3 + 2 )+|V 3||V 3||Y 33| sin (33 ) }
= - 1.0 X 1.05 X 10 sin(1.57-0+0) + 1.0 X1.02X10 sin(1.57 -0 +0) + 1.02 X20Xcos(1.57 – 0 + 0)]
Q3cal = - 0.7
P2 = P2(spec) – P2cal = 3 – 0.05= 2.95
P3 = P3(spec) – P3cal = - 4 – 0.0324 = - 4.0324
Q3 = Q3(spec) – Q3cal = - 2 – (-0.7) = - 1.3
Step – 5: Bus susceptance matrix

[B] = [−30
10 −20 ]
3
10

500 [−10 −30 ]

-1 1 −20 −10
[B] =
3

[B] = [
−0.02 −0.06 ]
-1 −0.04 −0.02

[B] = 3 [-20]

17
 1
[B]-1 = = - 0.05
−20
[B]-1 = - 0.05

Note
B matrix is the imaginary part of Ybus matrix for the buses except slack bus.
B matrix is the imaginary part of Ybus matrix for the P - V buses only.
Step – 6: Calculate  and V.

[] [ ]
P2 2.95

[]
2
3
=−[ B ] -1
|V 2|
P3
=- [
−0.04 −0.02 1.02
−0.02 −0.06 −4.0324
= ]
−0.04 −0.02 2.892
−0.02 −0.06 −4.0324[ ][ ] [ 23]=[−0.184
0.035
]
|V 3| 1.0

21 = 20 + 2 = 0+0.035 = 0.035 rad

31 = 30 + 3 = 0+(-0.184) = - 0.184 rad

[ ] [ ]
[ V 3 ] =−[ B ] -1 Q 3 = -(-0.05) −1.3 = - 0.065 p.u.
|V 3| 1.0
V31 = V30 + V3 = 1.0 + ( -0.065) = 0.935 p.u.
V31 = 0.935 p.u.
Example 5: Using fast decoupled load flow method, determine bus voltages for the system as shown in fig.
Perform one iteration.

Solution.

Step -1 : Form Ybus matrix

Ybus = [−0.3044+
0.3044− j 1.816
j1.88
−0.3044 + j 1.88
0.3044− j 1.816 ]
18
 Ybus = [ 1.842−1.405
1.904 1.7314
1.904 1.7314
1.842−1.405 ]
{Note: Use calculator in rad mode}

Step – 2:Initialize bus voltages

V1old = 1.050 p.u. (slack bus)

V2old = 1.020 p.u. (P-V bus)
Step – 3: Check for Q-limit violation.
Q2cal = - {|V 2||V 1||Y 21|sin ( 21−1+2 ) +|V 2|2|Y 22|sin ( 22 ) }
= - {1.02 x 1.05 x 1.904 x sin(1.7314)+ 1.022 X 1.842 sin(-1.405)]
Q2cal = - 0.1228 p.u.
Q2(min)¿Q2¿ Q2(max)

Bus 2 will acts as P-V bus

Step – 4: Compute P.

60
P2(spec) = PG2 – PL2 = = 0.6 p.u.
100
Q2cal = - {|V 2||V 1||Y 21|cos ( 21−1+2 )+|V 2|2|Y 22|cos ( 22 ) }
= - {1.02 x 1.05 x 1.904 x cos(1.7314)+ 1.022 X 1.842 cos(-1.405)]
Q2cal = - 0.0098 p.u.
P2 = P2(spec) – P2cal = 0.6 – (-0.0096)= 0.6098
Step – 5: Bus susceptance matrix
[B] = - 1.817
1
[B]-1 = = -0.5504
−1.817
[B]-1 = -0.5504

NOTE.
B matrix is the imaginary part of Ybus for the buses except slack bus.

Step – 6: Calculate  and V.

[ ] [ ]
[ 2 ] =−[ B ] -1 P 2 = -(-0.5504) 0.6157 = 0.34
|V 2| 1.0

19
 [ ] [ ]
[ V 2 ] =−[ B ] -1 Q 2 = -(-0.5504) 0.2 = 0.1226 p.u.
|V 2| 1.0
21 = 20 + 2 = 0+0.34 = 0.34 rad
V21 = V20 + V2 = 1.0 + 0.1226 = 1.1226 p.u.
V21 = 1.12260.34
Step – 6: Calculate .

[ ] [ ]
[ 2 ] =−[ B ] -1 P 2 = -(-0.5504) 0.6098 = 0.329 rad
|V 2| 1.02
21 = 20 + 2 = 0+0.329 = 0.329 rad = 18.85o
V1 = 1.050o
V2 = 1.0218.85o

Example 6: Obtain the power flow solution (one iteration) for the system shown in figure. The line
admittances are in per unit on a 100 MVA base. Use fast decoupled load flow method.

Q2(min) = -10MVAR
Q2(max) = 100MVAR

Solution:

Step 1: Form Ybus matrix.

20
Step 2: Initialize bus voltages.

(slack bus)

(P-V bus)

(P-Q bus)

Step 3: Check for Q limit violation for P – V bus

and the bus 2 will act as load bus (P - Q). [V2 =1.0o p.u]

Step 4: Calculate and .

21
 =-1.5
Step 5: Bus susceptance matrix

Step 6: Calculate and

22
 `

3.4 DC POWER-FLOW
Direct current load flow gives estimations of lines power flows on AC power systems. Direct current load
flow looks only at active power flows and neglects reactive power flows. This method is non-iterative and
absolutely convergent but less accurate than AC Load Flow solutions. Direct current load flow is used wherever
repetitive and fast load flow estimations are required.
Direct Current Load Flow (DCLF) gives estimations of lines power flows on AC power systems. DCLF
looks only at active power flows and neglects reactive power flows. This method is non-iterative and absolutely
convergent but less accurate than AC Load Flow (ACLF) solutions. DCLF is used wherever repetitive and fast
load flow estimations are required.
In DCLF, nonlinear model of the AC system is simplified to a linear form through these assumptions
 Line resistances (active power losses) are negligible i.e. R ≪ X.
 Voltage angle differences are assumed to be small i.e. sin(h) = h and cos(h) = 1.
 Magnitudes of bus voltages are set to 1.0 per unit (flat voltage profile).
 Tap settings are ignored.
Based on the above assumptions, voltage angles and active power injections are the variables of DCLF.
Active power injections are known in advance. Therefore for each bus i in the system, real power can be
converted into

23
 N
Pi = ∑ Bij (θi −¿ θ j )¿
j=1

in which Bij is the reciprocal of the reactance between bus i and bus j. As mentioned earlier, B ij is the imaginary
part of Yij. As a result, active power flow through transmission line i, between buses s and r, can be calculated
from
1
Pli = ( θ −θ )
X Li S R
where XLi is the reactance of line i.
DC power flow equations in the matrix form and the corresponding matrix relation for flows through
branches are represented as
θ= [ B ]−1 P
PL= (b x A)θ
where
P = Nx1 vector of bus active power injections for buses 1, …, N
B = NxN admittance matrix with R = 0
h = Nx1 vector of bus voltage angles for buses 1, …, N
PL = Mx1 vector of branch flows (M is the number of branches)
b = MxM matrix (bkk is equal to the susceptance of line k and non-diagonal elements are zero)
A = MxN bus-branch incidence matrix
Each diagonal element of B (i.e. B ii) is the sum of the reciprocal of the lines reactances connected to bus
i. The off-diagonal element (i.e. B ij) is the negative sum of the reciprocal of the lines reactances between bus i
and bus j. A is a connection matrix in which a ij is 1, if a line exists from bus i to bus j; otherwise zero.
Moreover, for the starting and the ending buses, the elements are 1 and -1, respectively.
Example 7: A simple example is used to illustrate the points discussed above. A three bus system is
considered. This system is shown in the below figure with the details given in the table 1 and 2. With the
base apparent power equal to 100 MVA, B and P are calculated as follows.

Table 1 : Loads and Generation.

Bus Number Bus Type PD (MW) QD (MVAr) PG (MW)

24
 1 Slack 0 0 Unknown

2 PV 10 5 63

3 PQ 90 30 0

Table 2 : Branches

Line Number From Bus To Bus X(p.u) Rating (MVA)

1 1 2 0.0576 250

2 2 3 0.092 250

3 1 3 0.17 150

[ ]
23.2435 −17.3611 −5.8824
B = −17.3611 28.2307 −10.8696
−5.8824 −10.8696 16.7519

[ ]
Unknown
P= 0.53
−0.9
As bus 1 is considered as slack, the firs row of P and the first row and column of B are disregarded. θ 2 and θ 3
are then calculated as follows.

[][ ] [−0.9 ]
−1
θ2 28.2307 −10.8696 0.53
=
θ3 −10.8696 16.7519

= [−0.0025
−0.0554 ]
Radian

[ ]
0
−0.1460
=
−3.17300
The values of A and b can be calculated as

[ ]
1 −1 0
A = 0 1 −1
1 0 −1

[ ]
17.3611 0 0
B= 0 10.8696 0
0 0 5.8824
Therefore he transmission flows are calculated as

25
[]
PL 1
PL 2 = Base MVA x b x A x θ
PL 3

[ ][ ][ ]
17.3611 0 0 1 −1 0 0
= 100 x 0 10.8696 0 x 0 1 −1 x −0.0025
0 0 5.8824 1 0 −1 −0.0554

[ ]
4.4243
= 57.4243 MW
32.5757

Example 8: Determine a DC load flow solution for the 5 bus system with B marix and P vector are
specified.

[ ]
−30 0 10 20
0 −100 100 0
B =
10 100 −150 40
20 0 40 −110

[ ] [ ] [ ]
−8.0 −0.3263 −18.70
4.4 0.0091 0.5214
P= ; δ = -B-1P ; δ = radians = degrees.
0 −0.0349 −2.000
0 −0.0720 −4.125

3.5 Comparison of load flow study.

[Link] GSLF NRLF FDLF

Require large number of Require less number of Require more number of

iterations to reach iterations to reach iterations than N.R
1
convergence. convergence. method.

Computation time per Computation time per Computation time per iteration
2 iteration is less iteration is more is less

It has linear convergence It has quadratic convergence

No convergency
3 characteristics characteristics
The number of iterations
The number of iterations are The number of iterations are
required for convergence
independent of the size of the does not dependent of the
4 increases with size of the
system size of the system
system
Less memory requirements than
Less memory requirements. More memory requirements.
5 N.R. method.

26
 Results in High Accuracy with Results in High Accuracy with Less accuracy than N.R.
6
maximum number of iteration minimum number of iteration method.

27