IEEE Power & Energy Magazine - November/December 2016 - 55

november/december 2016

1.8
1.4

60.06

1.2
1.0

60.04

0.8

60.02

0.6

1.6

60.08

0.4

60.00

0.2
0.0

:3
6
20 :49
:3
6
20 :58
:3
7
20 :06
:3
7
20 :15
:3
7
20 :24
:3
7
20 :32
:3
7
20 :41
:3
7
20 :49
:3
7
20 :58
:3
8:
07

59.98

Real Time EST
EI Frequency Hz

56-in TV LCD A

figure 3. The ac current for an individual TV receiver, with
EI system frequency, at the time of the Super Bowl XLVII
power outage. The horizontal axis is time of day (EST).

Halftime Show: Super Bowl XLVII
(3 February 2013)
notable frequency events also occurred during the halftime
show (performance during a long break between the first
and second half of the game), as shown in Figure 5.
A frequency drop started at 20:09:20, which was the time
that the show began with raised lights and flames shooting

60.06
PSFM1FREQ
PSFM3FREQ

:5
0:

Game

00
:4
8:

:4
6:

Game

:4
4:
18

00
2:
18
:4

59.94

Game

59.96

Commercial 1-1
Commercial 1-2

59.98

Commercial 1-3

60.00

60.02

Commercial 2-4

Commercial 2-1
Commercial 2-2
Commercial 2-3

60.04

Pronounced changes in the brightness of television images
could also explain other frequency events during the game,
which were found to correspond to commercial breaks. The
images of the football game itself were quite consistently
bright (at least while the power to the stadium was working),
but the image brightness during the commercials was highly
variable. Two commercial blocks within a few minutes of
each other during the first quarter of the game proved useful
for comparison. There was an important difference between
these two blocks: the first ad in the first block (started about
18:41:36) was generally "dark" while the first ad in the second block (started about 18:44:27) was generally "bright."
Correspondingly, there was a distinct frequency rise of more
than 20 mHz starting about 18:41:36, while from 18:44:15
to 18:44:55, there was comparatively little frequency change
(lower than ±5 mHz). See Figure 4.
The 30-mHz spike in frequency from 18:45:54 to 18:46:08
aligns with the black backgrounds at the ends of commercials
2-3 and 2-4; frequency dropped back 20 mHz over the following 10 s as the brighter images from the game telecast came
back onto the screen. note that these relatively fast load changes
arising from changes in the television image are superimposed
on other, generally slower, changes in load taking place over the
same time period. In addition, secondary frequency control, i.e.
automatic generation control is constantly acting to gradually
move frequency back toward 60 Hz.

2.0

60.10

Commercial Breaks: Super Bowl XLVII
(3 February 2013)

60.12

18
:4

This particular event provides a great deal of insight into
one of the primary reasons for the unusually high number of
frequency disturbances observed annually during the Super
Bowl. The brightness of the replay image being televised prior
to the power outage was quite consistent for a relatively long
time (more than 10 s) as was the brightness of the image of the
stadium lights following the outage. Furthermore, the difference in brightness between these two images was pronounced.
This particular image transition during a widely viewed program provided an ideal opportunity to observe the aggregate
impact of television image brightness on interconnection load.
With the frequency governing characteristic on the EI (Beta)
being approximately 25 MW/mHz, the 50-mHz rise indicates
that the total load on the EI dropped approximately 1,300 MW
over this time. Since the large frequency transition is so
closely synchronized with the distinct change in TV image
brightness, it is reasonable to conclude that the 1,300-MW
load loss was the aggregate reduction of load drawn by television sets all across the EI as the television image darkened.
The brightness of a television image significantly affected the
power consumed by many TV sets. An individual TV receiver
belonging to one of the authors was monitored by a DFR, and
the results during the stadium power outage are shown in Figure 3 along with the system frequency on the same time scale.
The drop in the TV power consumption aligns precisely with
the rise in EI system frequency.

figure 4. Recorded frequency during part of the first
quarter of Super Bowl XLVII. The horizontal axis is time of
day (EST).
ieee power & energy magazine

Table of Contents for the Digital Edition of IEEE Power & Energy Magazine - November/December 2016