EE 369
POWER SYSTEM ANALYSIS
Lecture 13
Newton-Raphson Power
Flow
Tom Overbye and Ross Baldick
1
Announcements
Homework 10 is: 3.49, 3.55, 3.57, 6.2, 6.9,
6.13, 6.14, 6.18, 6.19, 6.20; due 11/17.
(Use infinity norm and epsilon = 0.01 for
any problems where norm or stopping
criterion not specified.)
Homework 11 is 6.24, 6.26, 6.28, 6.30,
6.38, 6.42, 6.43, 6.46, 6.49, 6.50; due
Tuesday 11/22. Note that HW is due on
Tuesday because Thanksgiving is on
Thursday.
2
Dishonest Newton-Raphson
Since most of the time in the Newton-Raphson
iteration is spent dealing with the Jacobian, one
way to speed up the iterations is to only calculate
(and factorize) the Jacobian occasionally:
known as the Dishonest Newton-Raphson or
Shamanskii method,
an extreme example is to only calculate the Jacobian
for the first iteration, which is called the completely
dishonest Newton-Raphson or chord method.
Honest:
x ( v 1) x ( v ) - J ( x ( v ) )-1 f ( x ( v ) )
Dishonest: x ( v 1) x ( v ) - J ( x (0) )-1 f ( x ( v ) )
Stopping criterion f ( x
(v )
) used in both cases.
Dishonest Newton-Raphson
Example
Use the Dishonest Newton-Raphson (chord method)
to solve f ( x ) 0, where:
2
f ( x) x - 2
x
(v)
x ( v )
x ( v 1)
df (0)
(x )
f ( x(v) )
dx
(v) 2
((
x
) - 2)
(0)
2 x
1
(v)
(v) 2
x
((
x
) - 2)
2 x (0)
4
Dishonest N-R Example,
contd
1
( v 1)
(v )
(( x ( v ) )2 - 2)
2 x (0)
Guess x (0) 1. Iteratively solving we get
0
1
2
3
4
x ( v ) (honest)
1
1.5
1.41667
1.41422
1.41422
x ( v ) (dishonest)
1
1.5
1.375
1.429
1.408
We pay a price
in increased
iterations, but
with decreased
computation
per iteration
5
Two Bus Dishonest ROC
Region of convergence for different initial
guesses for the 2 bus case using the
Dishonest N-R
Red region
converges
to the high
voltage
solution,
while the
yellow region
converges
to the low
voltage
solution
Honest N-R Region of
Convergence
Maximum
of 15
iterations
Decoupled Power Flow
The completely Dishonest NewtonRaphson (chord), where we only calculate
the Jacobian once, is not usually used for
power flow analysis. However several
approximations of the Jacobian matrix are
used that result in a similar approximation.
One common method is the decoupled
power flow. In this approach
approximations are used to decouple the
real and reactive power equations.
Coupled Newton-Raphson
Update
Standard form of the Newton-Raphson update:
P ( v )
Q ( v )
P ( v )
V
( v )
Q
( v )
P( x ( v ) )
(v)
f
(
x
)
(v)
(v)
Q
(
x
P2 ( x ( v ) ) PD 2 PG 2
(v )
where P( x )
M
.
P (x( v ) ) P P
n
Dn
Gn
Note that changes in angle and voltage magnitude
both affect (couple to) real and reactive power.
Decoupling Approximation
P ( v )
Q ( v )
Usually the off-diagonal matrices,
and
V
are small. Therefore we approximate them as zero:
P ( v )
0
( v )
P( x ( v ) )
(v)
f
(
x
)
(v)
(
v
)
(
v
)
Q V
Q( x )
V
Then the update can be decoupled into two separate updates:
(v)
( v ) 1
P(x
(v)
),
(v)
1
(
v
)
Q(x ( v ) ).
10
Off-diagonal Jacobian Terms
So, angle and real power are coupled closely, and
voltage magnitude and reactive power are coupled closely.
Justification for Jacobian approximations:
1. Usually r = x, therefore Gij = Bij
2. Usually ij is small so sin ij 0
Therefore
Pi
Vi Gij cos ij Bij sin ij
Vj
Qi
j
Vi V j Gij cosij Bij sin ij
0
11
Decoupled N-R Region of
Convergence
12
Fast Decoupled Power Flow
By further approximating the Jacobian we
obtain a typically reasonable approximation
that is independent of the voltage
magnitudes/angles.
This means the Jacobian need only be built and
factorized once.
This approach is known as the fast decoupled
power flow (FDPF)
FDPF uses the same mismatch equations as
standard power flow so it should have same
solution if it converges
The FDPF is widely used, particularly when we
only need an approximate solution.
13
FDPF Approximations
The FDPF makes the following approximations:
1.
Gij 0
2.
Vi
3.
sin ij 0
1 (for some occurrences),
cos ij 1
Then: ( v ) B 1diag{| V |( v ) }1 P( x ( v ) ),
V
(v )
B 1diag{| V |( v ) }1 Q( x ( v ) )
Where B is just the imaginary part of the Ybus G bus jB bus ,
except the slack bus row/column are omitted. That is,
B is B bus , but with the slack bus row and column deleted.
Sometimes approximate diag{| V |( v ) } by identity.
14
FDPF Three Bus Example
Use the FDPF to solve the following three bus
Line Z = j0.07
One
Two
Line Z = j0.05
Three
Line Z = j0.1
200 MW
100 MVR
1.000 pu
200 MW
100 MVR
Ybus
20
34.3 14.3
j 14.3 24.3 10
10
30
20
15
FDPF Three Bus Example,
contd
Ybus
B 1
20
34.3 14.3
24.3 10
j 14.3 24.3 10 B
10
30
10
30
20
0.0477 0.0159
0.0159
0.0389
Iteratively solve, starting with an initial voltage guess
2
3
(0)
2
3
(1)
0
0
V 2
V
3
(0)
0 0.0477 0.0159
0
0.0159
0.0389
1
1
0.1272
2
0.1091
16
FDPF Three Bus Example,
contd
V 2
V
3
(1)
0.9364
1 0.0477 0.0159
1
1 0.0159 0.0389
1 0.9455
Pi ( x ) n
P PGi
Vk (Gik cos ik Bik sin ik ) Di
Vi
Vi
k 1
2
3
V 2
V
3
(2)
0.1272
0.1091
(2)
0.0477 0.0159
0.0159 0.0389
0.151 0.1361
0.107
0.1156
0.924
0.936
0.1384
Actual solution:
0.1171
0.9224
V
0.9338
17
FDPF Region of
Convergence
18
DC Power Flow
The DC power flow makes the most severe
approximations:
completely ignore reactive power, assume all the voltages are
always 1.0 per unit, ignore line conductance
This makes the power flow a linear set of equations, which
can be solved directly:
where B is the imaginary part of the bus admittance matrix
with the row and column corresponding to the slack bus
deleted, and, similarly, and P omit the slack bus.
B 1 P
19
DC Power Flow Example
20
DC Power Flow 5 Bus Example
One
360 MW
0 Mvar
Five
Four
MVA
Three
MVA
520 MW
MVA
0 Mvar
slack
1.000 pu
0.000 Deg
1.000 pu
-4.125 Deg
MVA
MVA
1.000 pu
-18.695 Deg
1.000 pu
-1.997 Deg
80 MW
0 Mvar
1.000 pu
0.524 Deg
Two
800 MW
0 Mvar
Notice with the dc power flow all of the voltage magnitudes are
1 per unit.
21
Power System Control
A major problem with power system operation
is the limited capacity of the transmission
system
lines/transformers have limits (usually thermal)
no direct way of controlling flow down a
transmission line (e.g., there are no low cost valves
to close to limit flow, except on and off)
open transmission system access associated with
industry restructuring is stressing the system in new
ways
We need to indirectly control transmission line
flow by changing the generator outputs.
22
Indirect Transmission Line
Control
What we would like to determine is how a cha
generation at bus k affects the power flow on
from bus i to bus j.
The assumption is
that the change in
generation at bus
is matched by an
opposite change a
the slack bus.
23
Power Flow Simulation Before
One way to determine the impact of a generator
change is to compare a before/after power flow.
For example below is a three bus case with an
overload.
131.9 MW
124%
One
200.0 MW
71.0 MVR
Two
68.1 MW
68.1 MW
200 MW
100 MVR
Z for all lines = j0.1
Three
1.000 pu
0 MW
64 MVR
24
Power Flow Simulation After
Increasing the generation at bus 3 by
95 MW (and hence decreasing
generation at the slack bus 1 by a
corresponding amount), results in a
31.3 MW drop in the MW flow on the
100%
line from bus 1 to 2.
101.6 MW
One
105.0 MW
64.3 MVR
Two
3.4 MW
Z for all lines = j0.1
Limit for all lines = 150 MVA
Three
98.4 MW
200 MW
100 MVR
92%
1.000 pu
95 MW
64 MVR
25
Analytic Calculation of
Calculating Sensitivities
control sensitivities by repeated power
flow solutions is tedious and would require many
power flow solutions.
An alternative approach is to analytically calculate
these values
The power flow from bus i to bus j is
Pij
Vi V j
So Pij
X ij
i j
sin( i j )
X ij
i j
X ij
We just need to get
ij
PGk
26
Analytic Sensitivities
From the fast decoupled power flow we know: B P(x ).
Sign convention in definition of P( x ) is that entry in P( x )
is negative if change in net injection (generation) is positive.
So to get the change in due to a change of generation at
bus k , just set P( x ) equal to all zeros except a minus one
at position k :
0
M
1 For 1MW increase in generation at bus k
0
M
27
Three Bus Sensitivity
Example
For the previous three bus case with Z j0.1
line
20 10 10
20 10
Ybus j 10 20 10 B
10
20
10 10 20
Hence for a change of generation at bus 3
2
20 10
10
20
0.0333
0
1 0.0667
0.0667 0
Changes in line flows are: P3 to 1
0.667 pu
0.1
P3 to 2 0.333 pu
P 2 to 1 0.333 pu
28
