sign in sign up
<<
Home , Super Bowl XLIV

The Impact of Synchronized Human Activities on Power System Frequency

Yin Lei1, Ye Zhang1, Jiahui Guo1, Dao Zhou1, Jerel Culliss1, Philip Irminger1,2, Yilu Liu1,2, Fellow, IEEE 1University of Tennessee 2Oak Ridge National Laboratory Knoxville, TN, USA Oak Ridge, TN [email protected], [email protected]

Abstract—It has been demonstrated in previous studies that centralized IMS via the Internet. The IMS then processes and nation-wide or largely synchronous societal activities have stores the incoming data while performing real-time impacts on the power grid frequency responses. This paper situational awareness analysis. The architecture of investigates NFL games as an example to evaluate FNET/GridEye is shown in Fig.1. the influence of synchronized human activities on the power system using data collected from a wide-area frequency monitoring network (FNET/GridEye). The statistics and plots of several frequency fluctuation phenomena and relevant analysis GPS are presented. Featured characteristics drawn from the End User frequency data detected during the Super Bowl games are FDR1 FDR 2 FDR N discussed. Ethernet Webpage PC WAN Index Terms—FNET/GridEye, FDR, Frequency Response, NFL Super Bowl, Power Grid Frequency Disturbance

I. INTRODUCTION Event Monitoring system The impact of large-scale societal events on power system Oscillation Monitoring system performance has drawn more attention in recent years. FNET Server Database Server Web Server Reference [1] presents a thorough observation of power system frequency variation during Super Bowl XLII, and Figure 1. Frequency Monitoring Network Architecture reference [2] illustrates the impact of the 2010 FIFA World Cup and Super Bowl XLIV on power grids. Reference [3] The FNET/GridEye system began collecting synchronized points out a possible reason for the unusual rapid frequency measurements in 2004. It currently has over 150 FDRs spread change may be the synchronized load changes of many throughout North America, and 50 units installed in Europe, television sets across the grid when the screens are changing Asia, and Africa. from bright to dark back and forth simultaneously. All of the above papers demonstrate that a large group of people This paper is presented in the following sequence: Section engaging in a synchronous, energy-intensive activity can II shows viewership of previous Super Bowl games indicating affect the power system as a whole. This paper summarizes how largely synchronized human activities are involved in the and analyzes the impact of Super Bowl games on the system study, Section III demonstrates several observations based on frequency from some other perspectives using FNET/GridEye FNET/GridEye frequency measurements, followed by some measurements. explanations of the possible reasons for the observed frequency dynamics. Finally, the conclusions and a discussion FNET/GridEye is a single phase GPS-synchronized of future research work are presented in Section IV. distribution-level phasor measurement system. Generally, it consists of two major components: the GPS-based II. SUPER BOWL INFORMATION synchronized sensor known as a Frequency Disturbance The Super Bowl is the annual championship game of the Recorder (FDR), and the Information Management System (NFL), the highest level of (IMS), which can also be thought of as a Phasor Data professional in the United States. The Concentrator, or PDC [4-5]. All the FDRs in the Super Bowl has frequently been the most watched American FNET/GridEye system measure the voltage signal from television broadcast of the year. Super Bowl XLVI, played in ordinary 120 V, 60Hz power outlets and send the 2012, became the most-watched American television program synchronized frequency and voltage phasor data back to the in history, drawing an average audience of 111.3 million This work was supported in part by the Engineering Research Center Program of the National Science Foundation and DOE under NSF Award Number EEC-1041877 and the CURENT Industry Partnership Program.

978-1-4799-6415-4/14/$31.00 ©2014 IEEE viewers and taking over the spot held by the previous year's Super Bowl. An estimated total audience of 166.8 million had watched that game, according to Nielsen [6], meaning that over half of the American population watched at least some of the initial broadcast. Table I lists the audience ratings of the last four Super Bowls, which are studied in this paper, as measured by Nielsen Company. The in column 5 represent an average percentage of all television-equipped households tuned in to Super Bowl games at any given moment.

TABLE I. SUPER BOWL AUDIENCE RATING

Super Stadium Participating Nielsen Date Figure 2. EI daily events histogram Bowl # Location Teams Ratings

Miami Saints, The numbers of triggered events with magnitudes higher Feb 7th 45.0 National XLIV Gardens, than 400 MW during previous four Super Bowl games are 2010 106.5 million FL demonstrated in Fig. 3 with their corresponding active power Feb 6th Alington, , 47.3 National changes in MW. The events with magnitudes lower than 400 XLV 2011 TX Green Bay Pachers 111 million MW are considered as minor events, and are not included in the study. Gaints, Feb 5th Indiananpol 47.0 National XLVI New England 2012 -is, IN 111.3 million Patriots New Feb 3rd Baltmore Ravens, 46.3 National XLVII Orleans, 2013 108.7 million LA III. FNET/GRIDEYE OBERSERVATIONS AND ANALYSIS OF POWER SYSTEM FREQUENCY The analysis work in this paper has paid particular attention to monitored power system frequency variations during the previous Super Bowl games. The FNET/GridEye database has stored information of power grid performance since 2004. In the FNET/GridEye system, the IMS is continuously analyzing the incoming frequency data. Once the rates of frequency change in several FDRs exceed an Figure 3. Super Bowl events histogram empirical threshold, a power system event is declared. Typical event types are frequency rise, frequency drop and oscillation. B. Observation 2: Frequency events tend to happen during Then, the frequency deviation of the disturbance is converted commercial breaks into the equivalent active power amount by using the It is elaborated in an online survey in reference [2] that empirical coefficient β [7]. This estimation has proven during regular commercial breaks, the viewers are more likely accurate after verifying with many confirmed Eastern to engage in other activities besides watching TV, such as Interconnection (EI) system events. using the restroom, getting food or beverages from the A. Observation 1: high density of frequency events during refrigerator, using a computer, stove, microwave, clothes dryer broadcasting period and washer, and other home appliances. Most of these activities involve the use of electricity. The increasing energy Large numbers of frequency disturbances were detected consumption explains why multiple grid frequency events during the game time. These events all involved hundreds of occurred during the commercials. Fig. 4 shows the distribution megawatts of power system variation. The FNET/GridEye of FNET/GridEye triggered frequency events for Super Bowl event trigger is set at 400 MW. Fig. 2 shows the histogram of XLVI and XLVII. daily triggered event numbers for the EI system. Typically, the EI has no more than 4 events per day. No four-hour period in the recorded events experiences as many events as there are during the Super Bowl. It is evidence that the high density of power system events during this time is tied to the behavior of the large number of people watching the game.

2012 2013 By comparison, we can clearly see that, the frequency experienced more fluctuation during Super Bowl halftime break time, a good quantity of sharp frequency rise and drop can be observed from the plot, which have strong relationship with human activities on that time period.

D. Observation 4: Sharp drop of frequency at the beginning of halftime break The frequency response is tightly associated with the Figure 4. Frequency events occurrence distribution for individual game progress and audience activities, so there is Super Bowl XLVI, 2012 and XLVII, 2013 no uniform pattern that can be observed from year to year. However, there are some characteristics common to all Super Consider that, on average, over 110 million people watch Bowl games. For example, the power system frequency drops these games. Assume the power of the home electrical dramatically during the halftime break and increases once the appliance is around 500 W. If only 1% of viewers use an third quarter begins. Fig. 7 to Fig. 10 shows the halftime break appliance, the load level would increase by 550 MW. This has frequency data from the last four Super Bowl games. There a measurable impact on the power system frequency. were sharp frequency drops at the beginning of each halftime break, and the slope of the frequency drop is marked with a C. Observation 3: large frequency fluctuation during black line. It also seems reasonable that the audience would halftime break participate in other activities right after the halftime break, The general impact on the system is relatively larger which would explain the drastic increases in electricity during halftime break due to the longer period of break time consumption and the resulting frequency drops. and the increase of viewership, according to Nielson Company 60.08 [5]. Fig. 5 shows a 30 minutes frequency plot during the halftime break of Super Bowl XLVII, while Fig. 6 plots the 60.06 same 30 minutes time window frequency for an ordinary 60.04 Sunday evening in late January. 60.02

60.08 60 Frequency, Hz

60.06 59.98

59.96 60.04

59.94 60.02

01:00:28 01:03:21 01:06:14 01:09:07 01:12:00 01:14:52 01:17:45 01:20:38 01:23:31 01:26:24 01:29:16 Time, UTC 60 Frequency, Hz Figure 7. Halftime 30 minutes plot for Super Bowl XLVII, 2013 59.98

60.08 59.96

60.06 59.94

01:00:28 01:03:21 01:06:14 01:09:07 01:12:00 01:14:52 01:17:45 01:20:38 01:23:31 01:26:24 01:29:16 60.04 Time, UTC Figure 5. 30 minutes frequency plot during halftime break on Super Bowl 60.02

XLVII 60 Frequency, Hz Frequency, 60.08 59.98

60.06 59.96

60.04 59.94

01:00:28 01:03:21 01:06:14 01:09:07 01:12:00 01:14:52 01:17:45 01:20:38 01:23:31 01:26:24 01:29:16 60.02 Time, UTC

60 Figure 8. Halftime 30 minutes plot for Super Bowl XLVI, 2012 Frequency, Hz

59.98

59.96

59.94

01:00:28 01:03:21 01:06:14 01:09:07 01:12:00 01:14:52 01:17:45 01:20:38 01:23:31 01:26:24 01:29:16 Time, UTC Figure 6. 30 minutes frequency plot on an ordinary Sunday Evening 60.08

60.06

60.04

60.02

60 Frequency, Hz

59.98

59.96

59.94

01:00:28 01:03:21 01:06:14 01:09:07 01:12:00 01:14:52 01:17:45 01:20:38 01:23:31 01:26:24 01:29:16 Time, UTC

Figure 9. Halftime 30 minutes plot for Super Bowl XLV, 2011 Figure 12. Different interconnections frequency swings on an ordinary Sunday evening

60.08 The frequency event can also be observed in the West 60.06 Electricity Coordinating Council (WECC) and Electric

60.04 Reliability Council of Texas (ERCOT) Interconnection at the same time. Fig.11 shows the 30 minute frequency 60.02 measurements of all three interconnections during halftime

60 break on Super Bowl XLVII 2013, red frames show frequency

Frequency, Hz Frequency, swings observed simultaneously. It can be seen that the 59.98 impacts of nationwide events are also reflected in interconnections other than EI. Fig. 12 shows the plots for an 59.96 ordinary Sunday evening frequency in all three 59.94 interconnections. We do not observe any consistency in these

01:00:28 01:03:21 01:06:14 01:09:07 01:12:00 01:14:52 01:17:45 01:20:38 01:23:31 01:26:24 01:29:16 plots. These observations are also true for shorter time Time, UTC window events. As is shown in Fig.13, an EI triggered event Figure 10. Halftime 30 minutes plot for Super Bowl XLV, 2011 can also be detected in WECC and ERCOT. These are very unique frequency signatures can only be observed during E. Observation 5: All US interconnections tend to oscillate Super Bowl game time, since EI, WECC and ERCOT are not together during large frequency disturbances synchronized interconnections. Fig. 14 is a similar amount event trigger by FNET/GridEye on non-Super Bowl time, we can see from this figure that the three interconnections do not always swing together during system events.

Figure 11. Different interconnections frequency swings during halftime 2013

Figure 13. NERC interconnections frequency swings during Super Bowl XLVII [Image Courtesy of NERC] [8] price is low. Such activities would have notable impact on the system frequency similar to the societal events discussed in this paper. It is evident that the societal effects could play a very significant role in future smart grid implementation. The observations in the paper will also provide valuable information in the study of responsive loads for power grid frequency regulation.

ACKNOWLEDGMENT This work was supported in part by the Engineering Research Center Program of the National Science Foundation and DOE under NSF Award Number EEC-1041877 and the CURENT Industry Partnership Program. Figure 14. NERC interconnections frequency swings during random event [Image Courtesy of NERC] [8] The authors would like to thank all of the FDR host universities, companies, high schools and individuals throughout the world. IV. CONCLUSION REFERENCES The statistical information of power system frequency [1] X. Tao, et al., "FNET observations on the impact of on behavior is investigated based on the real wide-area the power grid frequency," in Power & Energy Society General measurements taken in FNET/GridEye from Super Bowl Meeting, 2009. PES '09. IEEE, 2009, pp. 1-5. XLIV, 2010 to Super Bowl XLVII, 2013. It is apparent that [2] L. Chen, P. Markham, C. Chen, and Y. Liu, " Analysis of Societal large groups of people engaging in the same event at roughly Event Impacts on the Power System Frequency using FNET the same time can have significant impacts on the power grid Measurements," in 2011 IEEE Power Engineering Society General Meeting., pp. 1-8. frequency. One common characteristic drawn from the system [3] B. Starling, K. Thomas, R. Orndorff, J.Ingleson, E. Allen, “Effects of frequency recordings is the dramatic frequency drop during the Super Bowl on the Eastern Interconnection,” in Gerrgia Tech Fault the halftime breaks. The relatively longer commercial and AND Disturbance Analysis Conference, 2013 entertainment broadcasting time allows people to participate [4] Z. Zhong, C. Xu, B.J Li, Zhang, Tsai, S.-J.S., Conners, R.W.; Centeno, in other electric energy-related activities; hence, the V.A., Phadke, A.G., Y. Liu, "Power system frequency monitoring introduction of greater frequency variability during this time. network (FNET) implementation," IEEE Trans. Power Systems, vol. 20, pp. 1914-1921, Nov. 2005. Comparison of the Super Bowl data with non-game day [5] Y. Zhang, Markhan, P., T. Xia, L. Chen, Y. Ye, Z. Wu, Z. Yuan, L. data shows that there are far more power system events Wang, Bank, J., Burgett, J., Conners, R.W., Y. Liu, "Wide-Area Frequency Monitoring Network (FNET) Architecture and occurring in the former case. And most of the events occurred Applications," IEEE Trans. Smart Grid, vol. 1, pp. 159-167, Sept. during the commercial breaks. Clearly, the impacts of the 2010. Super Bowl on power grid frequency in this paper provide [6] "Super Bowl Ratings Record: Giants-Patriots Game Is Highest-Rated valuable information regarding the system dynamics of such TV Show In US History". Huffington Post. Retrieved February 7, 2012. popular events. The accumulation and statistical analysis of [Online]. Available: http://www.huffingtonpost.com/2012/02/06/super- bowl-ratings-record-tv-giants-patriots_n_1258107.html the FNET/GridEye frequency data present an incisive point of [7] R. Gardner, "A Wide-Area Perspective on Power System Operation and view on the power grid frequency behavior during such Dynamics," Ph.D. dissertation, Dept. Electrical Eng., Virginia Tech events. Univ. Blacksburg, 2008. [8] [Online].Available:http://www.nerc.com/AboutNERC/keyplayers/Docu Understanding the relationship between large-scale ments/NERC_Interconnections_Color_072512.jpg societal events and power frequency has important implications for power system. With the development of smart grid technology, similar large-scale, synchronous activities would be observed. Individual consumers may decide to switch on their home electrical appliances when the electricity