Human Factors Analysis of Power System Visualizations
Thomas J. Overbye
Dept. Elect. & Comp. Eng.
Overbye@ece.uiuc.edu
Douglas A. Wiegmann
Institute of Aviation
dwiegman@uiuc.edu
Aaron M. Rich
Institute of Aviation
am-rich@uiuc.edu
Yan Sun
Dept. Elect. & Comp. Eng
yansun@uiuc.edu
University of Illinois at Urbana-Champaign
Urbana, IL 61801 USA
Abstract
This paper describes an experimental approach to formally
testing the usability of different power system
visualizations. In particular, the ability of participants to
assess and correct power system voltage problems was
tested. Participants were divided into three groups: the
first group only saw tabular data, the second group oneline data, while the third group saw one-line data and a
color voltage contour.
The time to acknowledge the
voltage violations and the time to correct the violations
were assessed.
1. Introduction
One area in need of new research is the visualization of
electric power system operation and analysis information.
For the most part the visualization tools encountered by
many power system operators and engineers have evolved
little beyond tabular displays and the one-line diagram. In
a typical utility control center system quantities such as
power flows and voltages are usually represented either as
analog fields on a set one-line diagram consisting of all the
utility’s substations, or as entries in tabular list displays.
An overview of the system is usually only available on a
static map board with the only dynamic data shown using
different colored lights.
This needs to change. Restructuring is resulting in the
creation of much larger markets. Rather then transacting
with their neighbors, utilities are buying and selling in a
transcontinental market. Restructuring is resulting in the
creation of much larger markets under the control of a
single system operator. This will result in even more buses
and other devices to monitor and control. Simultaneously,
the entry of many new players into the market and the
increase in power transfers will result in even more data to
manage. Finally, electric grid operators will come under
increased scrutiny since their decisions, such as whether to
curtail particular transactions, can have a tremendous
financial impact on market participants. Power system
analysis will need to be modified in a number of different
ways to handle these new challenges.
One such
modification is in how system information is presented to
the user
Earlier work has, of course, been done in the area of
developing visualization techniques to aid in interpreting
power system data, although most have not yet actually
made it into the hands of practicing engineers. Several
recent examples are described in [1]-[9]. The purpose of
the present paper is slightly different, but we believe
complimentary to these earlier works.
Rather than
presenting new visualization techniques, our goal is to
describe initial experiments designed to provide at least a
preliminary formal assessment of whether these newer
visualization techniques are actually effective.
2. Experimental Setup
In attempting to formally assess visualization
effectiveness at least three issues must be considered. The
first issue is determination of the task. That is, what is the
user trying to accomplish. Of course in the area of power
system operations and analysis there are a wide variety of
different tasks. The tasks of the system planner are
substantially different from the tasks of the system
operator, which are in turn quite different from those of the
marketer or the engineer making presentations to
regulators. In this paper the task we focused on was
monitoring and control of bus voltage levels from either an
operator or an operational planning perspective.
The second issue is how to assess whether the
visualization helped the user to better perform a task or a
set of tasks. Effective task assessment requires a rather
specific knowledge of the ultimate goal. In our area of
voltage monitoring and control, we defined the goal to be
maintaining the bus voltage magnitudes above a specified
threshold. More specifically, following a disturbance we
were interested in how quickly the user was able to restore
the bus voltages to acceptable levels.
The third issue is how closely the experimental setup
matches what actually occurs in practice. The goal in the
setup of this experiment was to provide results that would
help to design the next generation displays for use by
either electric utility operators or operational planning
engineers. Therefore in the contextual setup of the
experiment our goal was to at least roughly approximate a
utility control environment in which an operator responds
to a system voltage disturbance, such as the one faced by
the Delmarva Power and Light (DPL) operators on July 6th
1999 during which their system experienced a near voltage
collapse following the loss of the Indian River 2 generator
[10].
To completely replicate such a situation, and then test
the response of the operators to various system
visualizations would obviously be quite difficult. In the
initial stages of the research project reported here our
experimental setup was admittedly a far cry from a control
center environment, in which experienced operators or
engineers monitor and control a system of which they are
quite familiar using displays they’ve used for years.
Nevertheless, we believe our experimental setup to be
viable and a good platform for future enhancements.
For test participants we used thirty University of
Illinois upper level undergraduates or graduate students
majoring in electrical and computer engineering. The
students had a good familiarity with basic electrical
engineering concepts, such as voltage and impedance, but
for the most part were not well versed in power system
operations. The experiment was performed using a small
30 bus power system modeled with a modified version of
PowerWorld Simulator [11].
During the course of the experiment the 30 bus system
was subjected to a variety of different contingencies, each
causing low voltage (defined as being below 0.95 per unit)
at one or more buses. All but one of these contingencies
involved line outages; the remaining contingency was a
generator outage.
Following each contingency the
participants needed to perform two tasks: 1) acknowledge
all the low voltage violations, and 2) switch in one or more
capacitors in order to correct the low voltage violations.
The system had a total of seven capacitors, all initially outof-service, and was designed so each contingency could be
corrected with the insertion of a single capacitor. The
students were told to assume that the impedance of each
transmission line was proportional to its length on the
system one-line diagram, and that the capacitor that was
electrically closest was always the correct choice.
Each participant saw one of three different
visualizations of the system, Tabular, Oneline or Contour.
As shown in Figure 1, the Tabular group only saw a
tabular visualization of the system voltages and
capacitors. The leftmost column on the display showed the
30 bus voltage magnitudes, sorted by bus number, while
the column immediately to its right showed the status of
the seven capacitors, again sorted by bus number. The
remaining column showed the line(s) outaged by the
contingency (this data was needed for the students to
correctly determine the electrically closest capacitor in the
post-contingent system).
Voltages below the 0.95
threshold were shown using a red font, while those above
this threshold were shown in blue. Following each
contingencies the participants first had to acknowledge all
the low voltage violations by clicking anywhere in the
bus’s row.
For each of the groups the computer
continually “beeped” until all the voltage violations were
acknowledged. Then they had to correct the voltage
violation by clicking on one or more capacitors to change
the capacitor’s status. The tabular group did have access
to a static system one-line diagram shown on a piece of
paper taped to their desk. This diagram was identical to
the one-line seen by the Oneline and Contour groups
except it was obviously not dynamic and hence did not
show the bus voltage magnitudes.
Figure 1: Tabular Group View of the 30 Bus System
As shown in Figure 2, the Oneline group only saw a
dynamic one-line visualization of the system, with the bus
voltage magnitude shown immediately to the right of the
bus on the one-line. Voltages below the 0.95 threshold
were drawn using a red font, while those above the
threshold were drawn in black. Again, following each
contingency the participants first had to acknowledge all
the low voltage violations. This was done by clicking on
the voltage magnitude field on the one-line. Then they had
to correct the voltage violation by clicking on the
capacitor’s circuit breaker symbol to change its status.
Lines outaged by each contingency were drawn on the oneline using a dashed line (e.g., line 22 to 26 in Figure 2).
As shown in Figure 3, the Contour group saw the exact
same representation as the Oneline group except in
addition to seeing the voltage magnitude fields on the oneline, their one-line also contained a color contour of the
voltage magnitudes [12]. The contour color was chosen so
buses with voltages below the threshold had a dark blue
contour, while those above the threshold were contoured
using a light cyan or yellow.
The low voltage
acknowledgement and capacitor switching procedures
were identical to the Oneline group.
Figure 2: Oneline Group View of the 30 Bus System
ignore values of others), the benefit of the high-proximity
display will likely be reduced, if not reversed [14].
This reversed-benefit of integrated displays was
evidence in the present study by participants’ response
latencies to the low-voltage alarms.
In general,
participants in the Tabular group had quicker
acknowledgement times than participants in the both the
Oneline and Contour groups. These differences are
depicted in Figure 4 and were confirmed by a one-way
between-groups ANOVA that revealed a significant effect
for display condition, F (2,21) = 4.10, p <. 05. Apparently,
the process of detecting and acknowledging low voltage
alarms required focused attention on one dimension of the
display which was facilitated by the tabular format and
somewhat hindered by the highly integrated one-line
diagrams.
3
2.5
2
1.5
1
0.5
0
Tabular
Figure 3: Contour Group View of the 30 Bus System
3. Results and Discussion
In the present experiment, the task of identifying and
resolving low-voltage violations required several
procedural steps and mental operations (albeit in simplified
form compared to “real world” situations).
Such
operations and procedures included the detection and
acknowledgement of the low bus voltages, as well as the
identification and closure of the electrically closest
capacitor to the low voltage buses that would resolve the
incident.
According to the well-established proximity
compatibility principle, these procedures and mental
operations are likely to be facilitated by one-line diagrams
and color contouring when information needs to be
integrated to perform a task [13]. That is, high display
proximity, or highly integrated information, helps in tasks
with high mental proximity. However, to the extent that
information must be treated separately (low mental
proximity, as when one must focus on one variable and
Oneline
Contour
Figure 4: Time in Seconds to Acknowledge All
Violations
In contrast to low voltage acknowledgement times,
however, participants in the Oneline and Contour groups
were more than twice as fast in solving low voltage
problems than participants in the Tabular group, F (2,21) =
41.30, p < .001. The average amount of time required to
resolve low voltages after all violations had been
acknowledge is depicted in Figure 5. As expected, the oneline diagrams effectively integrated the relative location of
buses and system capacitors, allowing for the electrically
closest capacitors to the low voltage buses to be quickly
identified. This finding corroborates the results of previous
research on other integrated displays (e.g., engine
parameter displays) showing that high display proximity
produces an emergent feature or Gestalt that directly serves
the integration of task requirements and facilitates the
parallel processing of information [15].
12
10
8
6
4
2
0
Tabular
Oneline
Contour
Figure 5: Time in Seconds to Completely Correct Low
Voltages
1.4
1.35
1.3
1.25
1.2
1.15
1.1
1.05
1
Tabular
Oneline
Contour
Figure 6: Mean Number of Capacitors Closed
In addition to facilitating quicker solutions to voltage
violations, the one-line diagrams also reduced the number
errors or capacitors closed when attempting to correct the
low voltages, F (2,21) = 7.56, p < .05. As illustrated in
Figure 6, participants in the Tabular group closed
significantly more capacitors when attempting to correct
the voltage problems. This finding is noteworthy because
previous reaction-time research has shown a speedaccuracy trade-off. As people respond more rapidly, they
tend to make more errors [16]. This reciprocity between
response latency and errors, however, was not evident in
the present study. This finding indicates that participants in
the Oneline and Contour groups were not just quickly and
haphazardly selecting capacitors. Rather, they were more
able to identify and closed the correct capacitor on their
first attempt to remedy voltage violations than were
participants in the Tabular group.
A final comment should be made about the apparent
lack of additional benefit afforded by the color contouring
used in the present experiment. The performance of
participants in the Contour group did not differ markedly
from the performance of participants in the Oneline group.
One reason for this observation may simply be that color
contouring was not needed to perform the tasks used in this
experiment. Although the color contours increased the
saliency of both the location and severity of the low
voltages, this saliency made not have been necessary in the
small 30 bus testing scenario used here.
Another issue that may have impacted the results was
the increased time necessary to refresh the contour display.
Following each voltage acknowledgement the displays
needed to be refreshed. For the Tabular and Oneline group
this refresh was quite fast, taking just 25 ms for the
Tabular display and 50 ms for the Oneline display. For the
Contour group the refresh was also relatively fast but there
was a very perceptible latency of about 200 ms between
the time the user acknowledged the violation and the time
the display was refreshed. This delay was simply due to
the time it took to update and render the contour bitmap on
the screen. This latency could certainly account for the
time variations shown in Figure 4 and at least some of the
response latency difference in Figure 5 between the
Oneline and Contour groups.
1 0 92 MW
2 55 M VR
D C 1
1 0 92 MW
2 61 M VR
N G
U
D C 2
S U
N S E T L K
M I L
U G
N
R D
E X A C O 5
T
3
R E Y B O D
L
C 1 0
D
0 M W
0 M V R
L A S G O W
G
P CH B TM T P
C LR A TP 0 3
C N A S T N
O E
5588
%%
C ON A ST O N
GR A CE T ON
G R A C
E . - 1
D 3
C
D C 1
5588
%%
C ON A ST O N
T P
U
N G
R D
E X A C O 5
T
3
R E Y B O D
L
C 1 0
D
0 M W
0 M V R
L A S G O W
G
P CH B TM T P
C LR A TP 0 3
C N A S T N
O E
L A N S
D
A N D I S
L
N G
U
D C 2
S U
N S E T L K
M I L
N U G
5 5 M W
1 1 M V R
CL R AT P 01
GR A CE T ON
G R A E
C . - 1
D 3
C
N U G
5 5 M W
1 5 M V R
CL R AT P 01
L A N S
D
T P
A N D I S
L
F I V E . F R
O
F I V E . F R
O
F A C E R K 1
F A C E R K 1
U M S P O N D
L
U M S P O N D
L
F A C E R K 2
F A C E R K 2
H O PE C RK
CE C IL
CO L O RA 1
CO L O RA 2
C ON O WG 0 3 C O NO W G0 1
H O PE C RK
CE C IL
CO L O RA 1
CO L O RA 2
C ON O WG 0 3 C O NO W G0 1
10 3 1 M W
37 3 M V R
10 3 1 M W
35 6 M V R
M T P L S N T
M T P L S N T
R O C R
K R
O K
E
G
C R
2 G E 1
R O C R
K O
R C
E
K
G R
2 G E 1
T P L S N P
M
T
T P L S N T
M
P
M N O T L
A
M N O T A
L
1
D R
O O T H Y
. P I P E
C
. P I P E
C
OT T R P T 0
J
C O
L . G A S
S AL E M
O T TR PT 1
S AL E M
O T TR PT 1
S H R A
M N # 1
S H R A
M N # 1
R AP H AE L
R AP H AE L
C AR D IF F
PE R RY 3 61 P ER R Y3 6 0
H A R F
R D 5 A
11
06
6 M
MV
WR
11 0 6 M W
35
3 5 6 M VR
H A
R F R D 6 A
P E
R R Y M A N
H A R R
F D 5 A
M Y
A S C 5 7 8
T E X A 5
S 0 9
E W I S
L
E D G D
W 6 2 0
CU M B # 1
C U M B # 1
C R A N E 9
5 2
1
C U M B # 1
J O P P 6
A 2 1
J O P P 6
A 2 1
T E X A
S 5 1 5
C UM B # 3
B E R D N 1 4
A
E D G W 6
D 2 1
G W Y N
CU M B # 1
13
17
03
6 M
MV
W1
06 MW
R1
3 7 3 M VR
H A
R F R D 6 A
P E
R R Y M A N
G L N
E A R M 1
L E N A R M 2
G
S A
H
S W
A A
W N
A 1
N 8
7
1
T . W A S H 1
M
. T O W S 7
E
N
M T . W A S
H 2
E . T O S
W N 8
B 9
S U M M E 1
R 8
M A
Y S C 5 7 9
N . C L F
I
U F
S
N 8
M
M
. E
C R
L F
7
1
I F 7
E W I S
L
E D G W
D 6 2 0
C AR D IF F
PE R RY 3 61 P ER R Y3 6 0
B
R U
K D
8 R
D K 9 E
C O L D
S CP O
G 2
L D S G
P 1
C UM B # 3
B E R D N 1 4
A
E D G W 6
D 2 1
1
C R A N E 9
5 2
T E X A 5
S 0 9
D O
R O T H Y
C . I
P P E 1 2
OT T R P T 0
J
C O
L . G A S
G L N
E A R M 1
L E N A R M 2
G
S W
A A
W N
A 1
N 8
7
1
S A
H
. T O W S 7
E
N
T . W A S H 1
M
M T . W A S
H 2
E . T O S
W N 8
B 9
S U M M E R
1 8
U F
S
N 8
M
. E
M
C R
L 7
F
I F 7
1
M A
Y S C 5 7 9
N . C L I
F
G W Y N
M A
Y S C 5 7 8
1
C . P P
I E 1 1
C . P I
P E 1 1
C . I
P P E 1 2
S
B
R U
K D
8 R
D K 9 E
C O L D
S C
P O
G L
2 D S G
P 1
J O
P P A 6 2 0
T E X A
S 5 1 5
J O
P P A 6 2 0
W N
I D Y E W
D 1
I
N D Y E D 2
W I
N D Y E W
D I
N
1 D Y E D 2
G U N O
P W D 6
G U P
N O W D 3
G U N P
O W D 6
G U N
P O W D 3
NE A ST 3 17
P H
I L A 5 A
P 5
H
O 7
I
N L
U A
M A
6
5 8
E N T R 5 5 1 M O N U 5
C
M
C E N
T R 5 5 2
H H
I I
G G
H H
L L
A N
6 5
A
C O N C
R D 6 4
NE A ST 3 17
U N I N
O
C ED A R C K
T O
W N S E N D
M I L
F U L R T N 2
A Z E L W
H
D
4
O
P L D N T 3
P
O N E Y 5
H
H 8
O 1
E
N Y 5 8 2
H 5
Z E L W D G
A
6 O L D N T G
E . P O I T
N 3
E .
S C
P H
K E
2 S . K
P
. 1
E
P O I N 1
T
N . E A T
S
H L
I L C
E H
N T
8
E P
. O I N T 4
H I L L E T
N 7
. R I V R 2 A
M
R A Y M N 1
G
R
H I L E N 0 2
R D M N 5 5
E
0
H I L E N G
0 R
4 A M
Y N G
R 4
R A M
Y N R 2
S T S
L D 9 8 3
R D M N 5 6
E
0
E R D M 5
N
S T L S D 8
9 4
E 0
R 7
D M N 0
5 8
# 1
M I
L L
# 2
I A D U
C T 7
R I V E R F
D
S M
E
M .
R I V R 8 3
P H
I L A 5 A
E N T R 5 5 1 M O N U 5
C
M
P 5
O
H 7
N
I U
L M
A 5
A
6 8
A
C E T
N R 5 5 2
H H
I I
G G
H H
L L
A N
6 5
C O N R
C D 6 4
S U
C L L # 2
M A S S R
Y E A
K E
N N D Y V L
M A S S E Y
M I L
D 7
C L A Y
T O N
M L
I L
# 2
R I V E R F
D
S M
E
P U MP H R EY
S H P
I L Y 2 2
P U M P
H R E Y
S U
C L L # 2
M A S S R
Y E A
S U B P
K E
N N D Y V L
M A S S E Y
E L A N E Y
D
C L A T
Y O N
L Y N C H
R E E
G
5
G N
R 5
E 3
N
E 5 5 4
W S P T 2
5 1
W S P 5
T 2 5
W S P T 2
5 4
6
A D C
U T 7
B E
L
B L
E
J O N S
E
J O N S
E
L I P I N 3
S 5
C O R O
S N
L I P I N S
3 5
HW K PT 45
H W KP T 4 4
W A G9
N 4% H S 62 %
W A G- 1 HBSR AN D N. S
W A G- 2 H S
S O L E Y S
3 O
4 L E Y 3 5
# 1
S C L
U L # 1
RI V 23 1 7R I V2 3 39
RL
IP
VT
23 4
05
XF
SO
OL P T 4 4
62
6%
2S %
E L A N E Y
D
L Y N C H
RL
IP
VT
23 4
05
XF
SO
S %OL P T 4 4
62
6%
2
P U MP H R EY
S H I
P L Y 2 2
P U M P
H R E Y
U N I N
O
C ED A R C K
T O
W N S E N D
M .
R I V R 8 3
S C L
U L # 1
S U B P
RI V 23 1 7R I V2 3 39
G N
R 5
E 5
N
E 5 5 4
R E E
G
3
W S P T 2
5 1
W S P T
5 2 5
6
W S P T 5
2 4
D 7
F U L L R T 3
N
F U L L R T 3
N
F U L R T N 2
H D
Z E L W D G
A
6 O L D N T 4
H 8
O 1
N
E Y 5 8 2
A Z E L W
H
5
O
G
P L D N T H
O
3
P N E Y 5
E . P O I T
N 3
H I
L L C
E H
N E
T .
8 P
S
C K
H 2
E S . E
K
.
P 1
P O I N 1
T
N . E A T
S
E .
P O I N T 4
. R I V R 2 A
M
H I L E N 0 2
R A Y M N 1
G
R
H I L E N 0
Y N R 2
R D M N 5 0
E
5
G 4
R A M
Y N G
R R
4 A M
R D M N 5 6
E
0
S T L
S D 9 8 3
E 0
R 7
D M N 0
5 8
E R D M N
5
S T L S D 8
9 4
H I L L E N
T 7
S O L E Y 3
S 4
O L E Y 3 5
C H
U R C H
C H E S W L
O D
2
M I L L N G N
T
C H S
E T R T N
C U
H R C H
C H E S W L
O D
W A G E
N R
1 2
3 M V
W R
G E
N
F O O D
M I D C
T Y T I
N O
R T H S T
D A N N E R
4 M V
1
3
M R
W
0
M W
O N T
0
M R
V L E B A N
M R 1 M R 2
M C K E R U N
M R 3
A B
0 M V
W R
0
M D
R
W O V E R F
V
N M E R E D H
T
K E T
N
I B
C O R O
S N
HW K PT 45
H W KP T 4 4
W A G9N44% H S 62 %
W A G- 1 HBSR AN D N. S
W A G- 2 H S
2
M I L L N G N
T
C H S
E T R T N
W A G N
E R
1 2
3 M V
W R
M I D T
C Y T I
G N
E
F O O D
N R
O T H S T
D A N N E R
4 M V
1
3
M R
W
0
M W
N N T
0
M R
V L E B A O
M R 1 M R M
2 C K E R U N
M R 3
D O V E R F
A B
0 M V
W 0
R
M R
W
V
N M E R E D H
T
K E T
N
I B
C R
O N R
C O
R N R
F A
R Z I E R H R
O S P O N D
C O L L E
G R D
F A
R Z I E R H O
R S P O N D
C O L L E
G R D
D L E
M
T P
M I D L
D E # 1
D L E
M
T P
M I D L
D E # 1
M I D L
D E # 2
V A U G H N
B E S T G T 3
4
M I D D
L E # 2
V A U G H N
C H U R H
C H L
B E S T G T 3
4
C R N .
H W A Y
C H U R C
H H L
C R O N
W S 4 3
C R N H
. W A Y
C R O W
N S 4 3
C .
P K 4 3 T P
E N F I E L D
B
C P
. K 4 3 T P
W . C A
H P E L
E N F I E L D
B
W . C H
A P E L
C E D A R 4 3
W .65C0H%
A PE L
9%
60%
C E D A R 4 3
W6.505
CH
9%9
%%A PE L
W
W
C R O N
W S 4 4
E D A R
C
4
C H
A P E L
C R O N
W S 4 4
E D A R
C
5 4 7
R I
V A
5 4 8
4
C H
A P E L
B E T
S G . R D
B E S
T G . R D
R I
V A
B E S T T
G 4 4
R I
V A
5 4 7
R I
V A
5 4 8
B E S T T
G 4 4
M IL F OR D
H A R N G T N
E N T R E V
C
L
A R . H 5 7
M
4
M A .
R H 5 4 8
M IL F OR D
H A R N G T N
A R . H 5 7
M
4
M A .
R H 5 4 8
V E R N
O N
E N T R E L
C
V
V E R O
N N
P P
. B
R
. R
I I
D D
G G
6 5
P P
. .
R R
B
I I
D D
G G
6 5
S H A R G
N T N
S H A R G
N T N
V I ST A 1
I ST A 0
V I ST A 1
I ST A 0
M L F R D C
T Y
S T E V
N S V L
R A S O N L
G
V
M L F R D T
C Y
S T E V
N S V L
R A S O N L
G
V
H I L S B R O
W Y E M I L L
H I L S B R O
W Y E M I L L
S T EE L E
S T EE L E
K A
R T Z
L O N G
W O O D
K R
A T Z
G E
R E N W D
L O N G
W O O D
G R
E E N W D
6
2
M W
E A S T N
O T P
0 M V R
6
2
M W
E A S T N
O T P
O AK G V2 3 0
0 M V R
O AK G V2 3 0
E A
S T O N
E A S T N U
M 1
E A
S T O N
E A S T N U
M 1
B O M
Z A N
G V0 5 3 O A KG V 05 4
G R
A C E
S T
T R
A P P E T P
B O Z
M A N
G V0 5 3 O A KG V 05 4
G R
A C E
S T
T A
R P P E T P
T A
L B O T
T A
L B O T
L E E
W S
R O
Y A L O A K
B R I D E
G V L
R O
Y A L O A K
L E W
E S
L E W E S T P
F I V E
B R I D E
G V L
T
P S
R A P P E
T
L E W E S T P
F I V E
A N K
L
M I D W A Y
T S
P
R A P P E
T
A N K
L
M I D W A Y
A R B E S N
H
O
T A Y O
L R
A R B E S N
H
O
R O B I S
N O N
T A Y O
L R
R H
E O B O T H
6%
6%
66
S U S S
E X
S U S E
S X
F I
A R M O N T
T N
A Y A R D
06
6 0%
086
7
ALLBT
B 0082
68
F A
I R M O N T
T A
N Y A R D
008
6 66
BL B
70 %
ALT
B 0082
68
R E
H O B O T H
6%
6%
66
R O B I N
S O N
M L
I L S B R O
N S E A O
F R D
7 7
2 5
M W
M R
V 1 2
M W
0
M V R
M I
L L S B R O
N S E A O
F R D
7 7
2 5
P E P P E R
0
P E P P E R
M W
M V
R 1 2
M W
0
M V R
0
M V
W R
R 1
I
I R 1 0
I R 4
M V
W R
R 1
I
P I E
N
I R 1 0
S T
IndianRiver
I R 4
C D
E A R
P I N
E
S T
I R 2
IndianRiver
C E
D A R
I R 2
I R 3
0 M V
W R 0
B E H
T A N Y
W R
V
M
M U M F O R D
D U P - S F
R D
P R E S O
T N
N K
D U P - S R
F D
P R E S O
T N
N K
I R 3
0 M V
W R 0
B E T
H A N Y
W R
V
M
M U M F O R D
B A Y A R D
C HA L K2 3 0
T O
D D
H A LK 5 TS
B A Y A R D
C HA L K2 3 0
F R N
A K F R D
E N
M K T R E A
T O
D D
H A LK 5 TS
L A R
U E L
E
F R A
N K F R D
N E W K
M T
E N
M K T R E A
L A R
U E L
E
S H
O R T
1
N E W M
K T
J A C
K T O W N
S H R
A P T W N
1 3
8 T H
S H
O R T
J A C
K T O W N
1
C L V T C LF
S H A
R P T W N
1 3
8 T H
S T
V I N
E N A L C
C A M B R D
I G
C L V T C LF
E L S O N
N
W C A
M B R D G
V I E
N N A L C
A I E
R Y
E L S O N
N
A I R
E Y
V I EN N A
V N 8
0
0
S MR Y CE 7 2
S M RY C E7 4
S T
C A M B R I
D G
W C M
A B R D G
I S H O P
B
V N
1 0
M W 1 3 M W
M V R 2 M V R
O C E
A N B A Y
V I EN N A
V N 8
0
0
I S H O P
B
V 1
N 0
O C A
E N B A Y
E D G E O
W O D
M W 1 3 M W
M V R 9 M V R
A R D E L A
M
E D G E W
O O D
A R D E L A
M
S MR Y CE 7 2
M RY C E7 4
H E B R O N
82 9 M W 8 28 MW
31 4 M V R 3 14 MV R
H E B R O N
82 9 M W 8 28 MW
31 6 M V R 3 16 MV R
M A
R I D E L
W A S
L T N
S
E W
N
H O OE
P C E A N P N
I
N S A L S R
B Y
M A
R I D E L
W A L
S T N
S
E W
N
H O E
O
P C E A N P I
N
B E L
R I N
N S A L S B
R Y
O E
C A N C T Y
B E R L I N P
T
B E L
R I N
O C
E A N C T Y
B E R L I N P
T
S I N E P U
X N
W O
R C E S T R
H E W2 3 2 0W
HE W 23 3 0E
S I N E P X
U N
W R
O C E S T R
H E W2 3 2 0W
HE W 23 3 0E
C U L V R
E
M H E R M O N
C U L V E
R
M H E R M O N
F R U T
I L N D
P IN E Y G R
F R U I
T L N D
P IN E Y G R
L O R E T O
L O R E T O
M
P R I N
M
P R I N
L I V E 1
O
A N
L I V E 1
O
A N
E N N E Y
K
K I G
N S
C K
E N N E Y
K
K I N
G S
C K
W E S T V
O E R
W E S T V
O E R
I N G S T O N
K
C O T
S E N
I N G S T O N
K
C O S
T E N
C I
R S F L D
9 M W
0 M V R
C I
R S F L D
S T O K
C T O N
P O C O O
M K E
9 M W
0 M V R
S T O K
C T O N
P O C O O
M K E
C H I N C O G
T
C H I N C O T
G
W A T T V
S I L
O A
K
W A T T V
S I L
O A
K
H A L L
H A L L
W A L O
L P S T
W A L L P
O S
W A L O
L P S T
A L L W O D
H
O
W A L L O
P S
A L L W O O
H
D
P A
R K S L E Y
P A
R K S L E Y
T A S L E Y
2 0
5
M W
M R
V
T A S L E Y
2 0
5
M W
M V
R
B L
E L H A V N
B E
L L H A V N
K E L L M
A
K E L L M
A
Figure 7: Contours of the voltage variation in approximately 150 of the Delmarva Power and Light (DPL) 69 kV voltages
before (left) and after the outage of the 85 MW Indian River 2 Generator on July 6th 1999.
As to whether contour will ultimately prove to be
beneficial, we would certainly hypothesize that it is
dependent upon system size. With 30 buses it is possible
to display the bus voltage magnitude fields on a one-line
with sufficient size that it is relatively easy for the user to
get a good conceptual mapping of the problem voltage
areas. With a larger case this would probably be
substantially more difficult. For example Figure 7 shows a
one-line/contour of the DPL system showing several
hundred buses. Superimposed is a color contour showing
the hypothesized voltage levels immediately before and
after the Indian River 2 generator outage. From the
contours the area of low voltages after the outage seems to
be immediately clear. Trying to convey this same
information using one-line fields, or tabular displays would
appear to be much more difficult. Future experiments are
needed to help address these issues.
Even on the small 30 bus system tested here it is
possible that the contouring did affect performance in more
subtle ways that are not evident in the present data. For
example, contouring may have influenced participants’
response strategies, such as the order in which they
acknowledged violations or selected capacitors to close.
One hypothesis which we are currently investigating is that
in cases with multiple low voltages the contours
subconsciously helped the users focus initially on the areas
with the most pronounced contours, corresponding to the
areas of the system with the lowest voltage magnitudes.
More fine-grained testing and analyses are ongoing to
examine these issues.
4. Conclusion
At first glance each of the three methods appears to
offer some benefit. The Tabular group were quickest in
acknowledging the voltage violations, the Oneline group
were quickest in correcting the problems, and the Contour
group made the best capacitor choices. However, the
slight time savings the Tabular group had in
acknowledgement (a savings of less than a second) was
greatly offset by the overall longer time it took for them to
completely correct the voltage problems – almost three
times as long as the other groups – and the larger number
of capacitor switchings they required. Therefore at least
for systems with thirty buses using the one-line approach
with or without contours appears to be significantly better
than a purely tabular approach. Benefits of contouring
over a pure one-line approach were not evident with the 30
bus system. More experiments are needed to determine
whether the contour approach is more useful with larger
systems.
5. Acknowledgement
The authors would like to acknowledge the support of
NSF through its grant NSF EEC 96-15792 , PSERC (Power
System Engineering Research Center), and the U.S.
Department of Energy through the CERTS program. The
authors would also like to thank the University of Illinois
students who participated in this study.
6. References
[1] P.M. Mahadev, R.D. Christie, “Minimizing User
Interaction in Energy Management Systems: Task
Adaptive Visualization,” IEEE Transactions on Power
Systems, Vol. 11, No. 3, pp. 1607-1612, August 1996.
[2] L.D. Christie, “Toward a Higher Level of User
Interaction in the Energy Management Task,” Proceedings
of the IEEE International Conference on Systems, Man
and Cybernetics, San Antonio, TX, October 2-5, 1994.
[3] P.R. D'Amour, W.R. Block, "Modern User Interface
Revolutionizes Supervisory Systems," IEEE Computer
Applications in Power, January 1994, pp34-39.
[4] K. Ghoshal, L.D. Douglas, "GUI Display Guidelines
Driving Winning SCADA Projects," IEEE Computer
Applications in Power, April 1994, pp. 39-42.
[5] G.P. de Azevedo, C.S. de Souza, B. Feijo, “Enhancing
the Human-Computer Interface of Power System
Applications,” IEEE Transactions on Power Systems, Vol.
11, No. 2, pp. 646-653, May 1996.
[6] P.M. Mahadev, R.D. Christie, “Envisioning Power
System Data: Concepts and a Prototype System State
Representation,” IEEE Transactions on Power Systems,
Vol. 8, No. 3, pp. 1084-1090, August 1993.
[7] M.D. Anderson, H.J. Pottinger, C.M. Schroeder, R.
Adapa, “Advanced Graphics Zoom in on Operations,”
IEEE Computer Applications in Power, pp. 25-28, April
1993.
[8] T.J. Overbye, J.D. Weber, “Visualization of Power
System Data”, Proc. 33th Hawaii International Conference
on System Sciences, Maui, HI, January 2000.
[9] T.J. Overbye and J.D. Weber, “New Methods for the
Visualization of Electric Power System Information”,
accepted for presentation at InfoVis 2000, October, 2000,
Salt Lake City, UT.
[10] Interim Report of the U.S. Department of Energy’s
Power
Outage
Study
Team,
January
2000,
http://tis.eh.doe.gov/post/interim.pdf.
[11] http://www.powerworld.com
[12] J.D. Weber, T.J. Overbye, “Voltage Contours for
Power System Visualization,” IEEE Trans. on Power
Systems, Vol. 15, February, 2000, pp. 404-409
[13] C.D. Wickens, A.D. Andre, “Proximity compatibility
and information display: effects of color, space, and
objectness of information integration,” Human Factors,
vol. 32, 1990, pp. 61-77.
[14] C.M. Carswell, C.D. Wickens, “Information
integration and the object display: An interaction of task
demands and display superiority”, Ergonomics, vol. 30 (3),
1987, pp. 511.527.
[15] C.M. Carswell, “Graphical information processing:
The effects of the proximity compatibility,” Proceedings of
the 34th annual meeting of the Human Factors Society ,
Santa Monica, CA, 1990, pp. 1494-1498.
[16] R.W. Pew, “The speed-accuracy operating
characteristic,” Acta Psychologica, vol. 30, 1969, pp. 1626.