Scribd
Upload
37 views

Lecture Note2

This document discusses techniques for selecting contingency cases for power system security analysis. It describes two sources of error in contingency selection and evaluates methods such as using a performance index to rank cases and the 1P1Q decoupled power flow method. It also introduces the concept of concentric relaxation to account for the limited geographic effect of outages.

Uploaded by

PRASHANT SINGH
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
37 views

Lecture Note2

This document discusses techniques for selecting contingency cases for power system security analysis. It describes two sources of error in contingency selection and evaluates methods such as using a performance index to rank cases and the 1P1Q decoupled power flow method. It also introduces the concept of concentric relaxation to account for the limited geographic effect of outages.

Uploaded by

PRASHANT SINGH
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 3

430 POWER SYSTEM SECURITY

the bad or likely trouble cases from the full outage case list is not an exact
procedure and has been the subject of intense research for the past 15 years.
Two sources of error can arise.
1. Placing too many cases on the short list: this is essentially the “conservative”
approach and simply leads to longer run times for the security analysis
procedure to execute.
2. Skipping cases: here, a case that would have shown a problem is not placed
on the short list and results in possibly having that outage take place and
cause trouble without the operators being warned.

11.3.4 Contingency Selection


We would like to get some measure as to how much a particular outage might
affect the power system. The idea of a performance index seems to fulfill this
need. The definition for the overload performance index (PI) is as follows:
2n
(11.8)
a11 branches
I

If n is a large number, the PI will be a small number if all flows are within
limit, and it will be large if one or more lines are overloaded. The problem then
is how to use this performance index.
Various techniques have been tried to obtain the value of PI when a branch
is taken out. These calculations can be made exactly if n = 1; that is, a table
of PI values, one for each line in the network, can be calculated quite quickly.
The selection procedure then involves ordering the PI table from largest value to
least. The lines corresponding to the top of the list are then the candidates for
the short list. One procedure simply ordered the PI table and then picked the
top N, entries from this list and placed them on the short list (see reference 8).
However when n = 1, the PI does not snap from near zero to near infinity
as the branch exceeds its limit. Instead, it rises as a quadratic function. A line
that is just below its limit contributes to PI almost equal to one that is just
over its limit. The result is a PI that may be large when many lines are loaded
just below their limit. Thus the PI’S ability to distinguish or detect bad cases
is limited when ti = 1. Ordering the PI values when n = 1 usually results in a
list that is not at all representative of one with the truly bad cases at the top.
Trying to develop an algorithm that can quickly calculate PI when n = 2 or
larger has proven extremely difficult.
One way to perform an outage case selection is to perform what has been
references 9 and 10). Here, a decoupled power flow
called the I P l Q tiic~t/~od(see
is used. As shown in Figure 11.10, the solution procedure is interrupted after
one iteration (one P - c) calculation and one Q - Vcalculation; thus, the name
1 P1 Q). With this procedure, the PI can use as large an n value as desired, say
n = 5. There appears to be sufficient information in the solution at the end of

BLOG FIEE http://fiee.zoomblog.com


C O N T I N G E N C Y ANALYSIS: D E T E C T I O N OF N E T W O R K PROBLEMS 431

.
Begin power flow solution

Build B' and B" matrices

Full outage
case list
J=====+ L
Model outage case
I

Solve the P-theta equation for


the AB's

Solve the Q-V equation for the


AIEI'S

Calculate flows and voltages for this case


then calculate the PI
.-

- Pick next outage case

FIG. 11.10 The 1 P1Q contingency selection procedure.


PI List
(one entry for
each outage
case)

the first iteration of the decoupled power flow to give a reasonable PI. Another
advantage to this procedure is the fact that the voltages can also be included
in the PI. Thus, a different PI can be used, such as:

n PIz = ) c ' ( + c ( AIEiI )zm


(11.9)
all branches all buses AIE("""
i i

BLOG FIEE http://fiee.zoomblog.com


432 POWER SYSTEM SECURITY

where AIEil is the difference between the voltage magnitude as solved at the
end of the l P l Q procedure and the base-case voltage magnitude. A/(E("""is a
value set by utility engineers indicating how much they wish to limit a bus
voltage from changing on one outage case.
To complete the security analysis, the PI list is sorted so that the largest PI
appears at the top. The security analysis can then start by executing full power
flows with the case which is at the top of the list, then solve the case which is
second, and so on down the list. This continues until either a fixed number of
cases is solved, or until a predetermined number of cases are solved which do
not have any alarms.

11.3.5 Concentric Relaxation


Another idea to enter the field of security analysis in power systems is that an
outage only has a limited geographical effect. The loss of a transmission line
does not cause much effect a thousand miles away; in fact, we might hope that
it doesn't cause much trouble beyond 20 miles from the outage, although if the
line were a heavily loaded, high-voltage line, its loss will most likely be felt
more than 20 miles away.
To realize any benefit from the limited geographical effect of an outage, the
power system must be divided into two parts: the affected part and the part
that is unaffected. To make this division, the buses at the end of the outaged
line are marked as layer zero. The buses that are one transmission line or
transformer from layer zero are then labeled layer one. This same process can
be carried out, layer by layer, until all the buses in the entire network are
included. Some arbitrary number of layers is chosen and all buses included in
that layer and lower-numbered layers are solved as a power flow with the outage
in place. The buses in the higher-numbered layers are kept as constant voltage
and phase angle (i.e., as reference buses).
This procedure can be used in two ways: either the solution of the layers
included becomes the final solution of that case and all overloads and voltage
violations are determined from this power flow, or the solution simply is used
to form a performance index for that outage. Figure 11.11 illustrates this
layering procedure.

FIG. 11.11 Layering of outage effects.

BLOG FIEE http://fiee.zoomblog.com

You might also like