al size=-1 color=black>

Validation of Out-of-Step Protection With a Real Time Digital Simulator 1
Validation of Out-of-Step Protection With a
Real Time Digital Simulator
Frank Plumptre, BC Hydro, Stephan Brettschneider, Cegertec, Allen Hiebert, BC Transmission Corporation
Michael Thompson and Mangapathirao Venkat Mynam, Schweitzer Engineering Laboratories, Inc.
AbstractThis paper describes the use of a real time digital
simulator with dynamic machine models to validate out-of-step
tripping and blocking elements in a new protective relaying sys-
tem being installed on the BC Hydro 500 kV power system. The
technique has also been used to study and validate a generation
shedding remedial action scheme. This unique approach has
many advantages over traditional methods of studying the effect
of power swings on protection systems. Traditional methods for
studying power swings are limited in their ability to predict the
response of protective elements due to the fact that they model
the power system in the positive-sequence network only. A real
time digital simulator can represent the power system under
more realistic conditions so that the response of the protective
system can be tested under conditions that nearly match actual
field conditions. Case studies are discussed in the paper showing
the importance of this new approach.
I. I
NTRODUCTION
Protective relays, especially distance elements used for de-
tecting and isolating faulted sections of the power system, can
respond to power system swings and out-of-step conditions.
BC Hydro uses the principle that distance elements will trip
and generally accepts natural tripping of transmission lines for
out-of-step conditions. It is recognized that such tripping may
be at a nonoptimal angle across the tripped breakers; however,
transmission line breakers can be specified for such a switch-
ing duty. In situations where protective relaying will not re-
spond to power swings, dedicated out-of-step tripping protec-
tion may be applied. For example, where the expected swing
is within a transmission line that is protected by only current-
based protection (i.e., line current differential or phase com-
parison) or within power transformers.
There are situations however, where the natural response of
distance elements to out-of-step tripping on power system
swings and out-of-step conditions is undesirable. Tripping
may open critical paths of the transmission system grid while
the system is under a stressed condition, which could poten-
tially make the situation worse and could lead to complete
collapse of the power system. Most multifunction distance
relays today include some form of power swing blocking and
out-of-step tripping elements. Power swing blocking elements
prevent the undesired tripping of critical transmission paths
during a power swing, and out-of-step tripping elements allow
intentional opening of transmission paths to aid in creating
grid islands.
Traditional methods of studying dynamic power system
stability do a good job of predicting what conditions of load
flow and system contingencies can cause a system to go out of
step. These methods can also identify specific line terminals
that are susceptible to tripping during power swings. They are
also used to design remedial action schemes to shed load or
generation under certain conditions and to keep the system
from going unstable. The information provided by dynamic
stability studies is also invaluable to the protection engineer
who is charged with implementing blocking and/or tripping
elements in the protective relays to improve the robustness of
the protection system during these power swing disturbances.
However, most dynamic-stability programs assume bal-
anced conditions and model the system in the positive-
sequence only. Distance elements are complex devices that do
not simply measure the V/I = Z [1]. Modern numerical dis-
tance elements also often include many supervisory checks
that must be satisfied before they issue a trip. Simply plotting
the apparent positive-sequence trajectory on an RX diagram
versus the distance elements characteristic will not fully pre-
dict the response of the element during a real power swing
disturbance.
Modern real time digital simulators (RTDS) can model the
power system under both balanced and unbalanced conditions.
The dynamic machine data models (generator, excitation con-
trols, governor controls, stabilizer controls, etc.) from popular
dynamic stability programs can be used to build dynamic ma-
chine models that can run in real time on the simulator. The
actual protective relays on the system can be connected to the
simulator and their response during power swings can be ex-
amined and adjusted to ensure that the desired operation is
achieved.
In this paper, we will provide two examples of using this
new tool. The first example describes validating the design
and settings for an out-of-step blocking and tripping scheme.
The second example describes validating a generation shed-
ding remedial action scheme. But first, we will provide some
background on the subject of out-of-step relaying.
II. B
ACKGROUND
A. Traditional Analysis
A system planner uses a dynamic-stability program to simu-
late the power systems response following possible distur-
bance events, especially those that could stress the system to,
or beyond, its dynamic stability limit. The power system dis-
turbance events of special interest to the system planner and
also the protection engineer are those that show a marginally
stable or even unstable result. These simulations, with varia-
tions on the severity of the system loading and the distur- 2
bance, can be used by the system planner to provide stability
study results for a variety of out-of-step events.
When assisting the protection engineer with out-of-step
protection settings on a particular transmission line, the sys-
tem planner uses the same power flow and stability model that
is already available from the planners dynamic stability stud-
ies of the same region. To show the out-of-step condition, it is
usually necessary to assume the disturbance has a longer than
normal fault duration or assume that the system is loaded be-
yond its operating limits. A range of out-of-step simulation
cases can be providedfrom one that almost slips, to one that
is just marginally unstable (a slow-developing slip), to one
that shows a fast slip such as a line reclose when the generat-
ing station is already isolated and has a frequency difference
of several hertz.
The protection engineer typically requests the transmission
lines apparent impedance during the power swing and possi-
bly also the voltages and power flows at the line terminals.
The apparent impedance trajectory can be shown along with
the line protections distance element characteristic. Because
the system planners usual stability model uses only the posi-
tive-sequence network, these simulation results are for a bal-
anced system only. Attachments 1 and 2 show the impedance
trajectories overlaid on a mho characteristic for stable and
unstable power swings respectively. Attachment 1 shows the
impedance trajectory (stable swing) entering the Zone 2 mho
characteristic for the fault and then swings out as the fault is
cleared. The impedance re-enters the Zone 3 characteristic
during the swing and then swings out as the swing dampens
out.
B. Power Swing Detection
Power swing detection methods are based on the fact that
the change in apparent impedance, seen by the relay due to a
power swing, is gradual compared to a step change that occurs
when the system is faulted. Traditional techniques used double
blinders, concentric polygons, or concentric circles to detect
power swings [2] [3]. Fig. 1 shows an example of an out-of-
step element that uses concentric polygons.
Zone 2
Zone 6
Zone 1
Right
Inner-Blinder
Left
Inner-Blinder
Z1L1
(R1L6)
(R1L7)
(R1LB)
(R1RB)
(R1R6)
R1
X1
(X1T7)
(R1R7)
Z1
Trajectory
(X1B7)
(X1B6)
(X1T6)
Fig. 1. OOS Characteristic Using Concentric Polygons
The detection and decision of the nature of the power swing
is derived from the travel time of the positive-sequence im-
pedance, which is measured between the moment that it enters
the outer characteristic and the moment that it enters the inner
characteristic. If the measured impedance stays between the
characteris