Power System State Estimation
Power System State Estimation
1.0 Introduction
In these notes, we explore two very practical
and related issues in regards to state estimation:
- use of pseudo-measurements
- network observability
1
The most common “exact” pseudo-measurement
is the bus injection at a substation that has no
generation and serves no load. Figure 1 below
illustrates.
Bus p
Fig. 1
In Fig. 1, bus p has no generation or load. We
therefore know the real and reactive power
injection of this bus with precision; it is 0. And
so we can add two more measurements to the
measurements that we actually have:
zi hi ( x) i (1)
zi 1 hi 1 ( x) i 1 (2)
2
where:
zi Pp ,inj 0 (3)
“Measurements”
zi 1 Qp ,inj 0 (4)
hi ( x) Pp ,inj V p Vk G pk cos( p k ) B pk sin( p k ) 0
n
k 1
(5)
hi 1 ( x) Qp ,inj V p Vk G pk sin( p k ) B pk cos( p k ) 0
n
k 1
(6)
We recognize the summations of eqs. (5) and
(6) as the power flow equations for real and
reactive power injection, respectively.
3
3. Observability
Recall our very first example at the beginning of
the first set of state estimation notes. It is below.
I1 I2
R2 R3
A2
● Node 3
Fig. 2
4
To determine the state of the circuit (v1, v2, and
e), we wrote each one of the measurements in
terms of the states. We then expressed these
four equations in matrix form:
1 1 1 1.0
1 0 0 v1 3.2
v 2
0 1 0 0.8
e (7)
0 0 1 1.1
Let’s denote terms in eq. (7) as A, x, and b, so:
Ax b (8)
We solved eq. (8) using:
x A A A b G A b A b
T 1 T 1 T I
(9)
First, the gain matrix is given as
1 1 1
1 1 0 0 2 1 1
1 0 0
G A A 1 0 1 0 1 2 1
T
0 1 0 (10)
1 0 0 1 1 1 2
0 0 1
The inverse of the gain matrix is then found
from Matlab as
3 1 1
G 1 3 1
1 1
4 (11)
1 1 3
5
The pseudo-inverse is then
3 1 1 1 1 0 0 1 3 1 1
A G A 1 3 1 1 0 1 0 1 1 3 1
I 1 T 1 1
4 4 (12)
1 1 3 1 0 0 1 1 1 1 3
1 0 0 v1 3.2
0 1 0 v 0.8
2 (14)
0 0 1 e 1.1
6
Actually, here the matrix is 3×3 and therefore
we can solve exactly as:
v1 3.2
v 0.8
2 (15)
e 1.1
But let’s go ahead and use eq. (9) to see what
happens.
7
1 0 0 3.2 3.2
x A b 0 1 0 0.8 0.8
I
(19)
0 0 1 1.1 1.1
Compared to the solution of eq. (13), our
estimate can be assumed to be less accurate
(since it is based on fewer measurements), but at
least we still did obtain a solution.
8
The matrix is once again non-square, so we
must use our least-squares procedure.
9
Example:
Consider the system in Fig. 3. This is the same
system considered in our previous set of state
estimation notes and in your homework
assignment. There are real power measurements
taken at P12, P13, and P32. But we assume that the
measurements P13 and P32 fail so that we only
have the measurement P12=0.62 pu. As in
previous examples, assume all voltages are 1.0
per unit, all measurement devices have σ=0.01,
and that the bus 3 angle is reference. The state
vector is x=[θ1 θ2]T as before.
Bus 1 Bus 2
P12
P13
P32 Bus 3
Fig. 3
10
a) Determine the vector of measurement
expressions h(x), the derivative
h( x )
expressions H , and the weighting
x
matrix R.
b) Develop A H T ( x)R 1 H ( x) , b H T ( x) R 1z h( x)
and comment on our ability to solve
AΔx=b for Δx.
Solution:
(a)
h1 ( x) P12 V1 g12 V1 V2 g12 cos(1 2 ) V1 V2 b12 sin(1 2 ) 5 sin1` 2
2
Now get b.
11
5 cos(1 2 )
b H T ( x) R 1 z h( x) 100000.62 5 sin 1` 2
5 cos(1 2
)
3.1cos(1 2 ) 25 cos(1 2 ) sin 1` 2
10000
3.1cos(1 2 ) 25 cos(1 2 ) sin 1` 2
12
The matrix is singular and we cannot solve the
equation.
13
Most state estimators will perform an
observability analysis. If the network is
unobservable, it may be the case that some
pockets or islands of the network are still
observable. Thus we also need to be able to
identify observable parts of the system from
unobservable parts of the system. Doing so will
will enable us to determine from which part of
the network we need to obtain additional
measurements.
4. Approximate pseudo-measurements
14
When the state estimator detects that the
network is unobservable, it can make use of
approximate pseudo-measurements. Examples
of such approximate pseudo-measurements
include:
Information obtained from plant operators over
phone or e-mail.
Information obtained from previous
measurements.
Information obtained from a load flow
calculation.
In using approximate pseudo-measurements, it
is generally good practice to pair it with a
relatively large variance in the weighting
matrix, given that it is in fact “approximate.”
15
Clearly, if the state estimate is very poor, we
will not want to use it even if it is the “best.”
16
[1] W. Brogan, “Modern Control Theory,” Prentice-Hall, 1982.
[2] G. Strang, Linear Algebra and its applications,” Harcourt-Brace, 1988.
[3] A. Monticelli, “State estimation in electric power systems, a generalized
approach,” Kluwer, 1999.
17