Demonstration of a Novel
Synchrophasor-based Situational
Awareness System: Wide Area Power
System Visualization, On-line Event
Replay and Early Warning of Grid
Problems
DOE-DE-OE0000128
Final Report, M arch 201 3
EPRI P roject M an a g e r
A. Del R osso
ELEC TR IC P O W E R R ESEA RC H IN STITUTE
3420 H illview Avenue, Palo Alto, C alifornia 94304-1338 ■ PO Box 10412, Palo Alto, C alifornia 94303-0813 ■ USA
800.313.3774 ■ 650.855.2121 ■ askepri@ epri.com ■ w w w .epri.com
DISCLAIM ER OF W A R RA NTIES AND LIMITATION OF LIABILITIES
DISCLAIMER: “THIS REPORT W AS PREPARED AS AN ACCOUNT OF W O R K SPONSORED BY AN
AGENCY OF THE UNITED STATES GOVERNMENT. NEITHER THE UNITED STATES
GOVERNM ENT NOR ANY AGENC Y THEREOF, NOR ANY OF THEIR EMPLOYEES, MAKES ANY
WARRANTY, EXPRESS OR IMPLIED, OR ASSUMES ANY LEGAL LIABILITY OR RESPONSIBILITY
FOR THE ACCURACY, COMPLETENESS, OR USEFULNESS OF ANY INFORMATION,
APPARATUS, PRODUCT, OR PROCESS DISCLOSED, OR REPRESENTS THAT ITS USE W OULD
NOT INFRINGE PRIVATELY OW NED RIGHTS. REFERENCE HEREIN TO ANY SPECIFIC
COMMERCIAL PRODUCT, PROCESS, OR SERVICE BY TRADE NAME, TRADEMARK,
MANUFACTURER, OR OTHERW ISE DOES NOT NECESSARILY CONSTITUTE OR IMPLY ITS
ENDORSEMENT, RECOMMENDATION, OR FAVORING BY THE UNITED STATES GOVERNMENT
OR ANY AGENCY THEREOF. THE VIEW S AND OPINIONS OF AUTHORS EXPRESSED HEREIN
DO NOT NECESSARILY STATE OR REFLECT THOSE OF THE UNITED STATES GOVERNMENT
OR ANY AGENCY THER EO F.”
THIS DOCUMENT W AS PREPARED BY THE ORGANIZATION(S) NAMED BELOW AS AN
ACCOUNT OF W O R K SPONSORED OR COSPONSORED BY THE ELECTRIC POWER RESEARCH
INSTITUTE, INC. (EPRI). NEITHER EPRI, ANY MEMBER OF EPRI, ANY COSPONSOR, THE
ORGANIZATION(S) BELOW, NOR ANY PERSON ACTING ON BEHALF OF ANY OF THEM:
(A) MAKES ANY W ARRANTY OR REPRESENTATION W HATSOEVER, EXPRESS OR IMPLIED, (I)
WITH RESPECT TO THE USE OF ANY INFORMATION, APPARATUS, METHOD, PROCESS, OR
SIMILAR ITEM DISCLOSED IN THIS DOCUMENT, INCLUDING MERCHANTABILITY AND FITNESS
FOR A PARTICULAR PURPOSE, OR (II) THAT SUCH USE DOES NOT INFRINGE ON OR
INTERFERE W ITH PRIVATELY OWNED RIGHTS, INCLUDING ANY PARTY'S INTELLECTUAL
PROPERTY, OR (III) THAT THIS DOCUMENT IS SUITABLE TO ANY PARTICULAR USER'S
CIRCUMSTANCE; OR
(B) ASSUMES RESPONSIBILITY FOR ANY DAMAGES OR OTHER LIABILITY W HATSOEVER
(INCLUDING ANY CONSEQUENTIAL DAMAGES, EVEN IF EPRI OR ANY EPRI REPRESENTATIVE
HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES) RESULTING FROM YOUR
SELECTION OR USE OF THIS DOCUMENT OR ANY INFORMATION, APPARATUS, METHOD,
PROCESS, OR SIMILAR ITEM DISCLOSED IN THIS DOCUMENT.
REFERENCE HEREIN TO ANY SPECIFIC COMMERCIAL PRODUCT, PROCESS, OR SERVICE BY ITS
TRADE NAME, TRADEMARK, MANUFACTURER, OR OTHERWISE, DOES NOT NECESSARILY
CONSTITUTE OR IMPLY ITS ENDORSEMENT, RECOMMENDATION, OR FAVORING BY EPRI.
THE FOLLOWING ORGANIZATION(S), UNDER CONTRACT TO EPRI, PREPARED THIS REPORT:
Electric Power Research Institute
NOTE
For further information about EPRI, call the EPRI Customer Assistance Center at 800.313.3774 or
e-mail askepri@ epri.com .
Electric Power Research Institute, EPRI, and TOGETHER...SHAPING THE FUTURE OF ELECTRICITY
are registered service marks of the Electric Power Research Institute, Inc.
ACKNOWLEDGMENTS
This material is based upon work supported by the Department of Energy under Award Number
DE-F G26-08NT01903.
The following organizations, under contract to the Electric Power Research Institute (EPRI),
prepared this report:
Electric Power Research Institute
3420 Hill view Ave.
Palo Alto, CA 94304
Principal Investigator
A. Del Rosso
University of Tennessee Knoxville
Principal Investigators
K. Sun, Y. Liu
Quanta Technology
Principal Investigator
G. Zhang
HTC
Principal Investigator
H. Chen
This report describes research sponsored by the U.S. Department of Energy. EPRI would like to
acknowledge the support and contributions of the following folks:
•
Phil Overholt (DOE Program Director)
•
Brian Mollohan (DOE Program Manager)
•
DeJim Lowe, Paul Trachian, and Josh Shultz (Tennessee Valley Authority)
•
Fred Elmendorf, Ritchie Carroll (Grid Protection Alliance)
•
Xiaochuan Luo, Qiang Zhang (ISO-New England)
•
Jade Wong (Con Edison)
This publication is a corporate document that should be cited in the literature in the following
manner:
Demonstration o f a Novel Synchrophasor-based Situational Awareness System: Wide Area
Power System Visualization, On-line Event Replay and Early Warning o f Grid Problems. EPRI,
Palo Alto, CA: 2013. DOE-DE-OE0000128.
m
EXECUTIVE SUMMARY
Introduction
Since the large North Eastern power system blackout on August 14, 2003, U.S. electric utilities
have spent lot of effort on preventing power system cascading outages. Two of the main causes
of the August 14, 2003 blackout were inadequate situational awareness and inadequate operator
training In addition to the enhancements of the infrastructure of the interconnected power
systems, more research and development of advanced power system applications are required for
improving the wide-area security monitoring, operation and planning in order to prevent largescale cascading outages of interconnected power systems. It is critically important for improving
the wide-area situation awareness of the operators or operational engineers and regional
reliability coordinators of large interconnected systems. With the installation of large number of
phasor measurement units (PMU) and the related communication infrastructure, it will be
possible to improve the operators’ situation awareness and to quickly identify the sequence of
events during a large system disturbance for the post-event analysis using the real-time or
historical synchrophasor data.
The purpose of this project was to develop and demonstrate a novel synchrophasor-based
comprehensive situational awareness system for control centers of power transmission systems.
The developed system named WASA (Wide Area Situation Awareness) is intended to improve
situational awareness at control centers of the power system operators and regional reliability
coordinators. It consists of following main software modules:
•
Wide-area visualizations of real-time frequency, voltage, and phase angle measurements and
their contour displays for security monitoring.
•
Online detection and location of a major event (location, time, size, and type, such as
generator or line outage).
•
Near-real-time event replay (in seconds) after a major event occurs.
•
Early warning of potential wide-area stability problems.
The system has been deployed and demonstrated at the Tennessee Valley Authority (TVA) and
ISO New England system using real-time synchrophasor data from openPDC.
Apart from the software product, the outcome of this project consists of a set of technical reports
and papers describing the mathematical foundations and computational approaches of different
tools and modules, implementation issues and considerations, lessons learned, and the results of
validation processes.
Project Developm ent
A multidisciplinary team from five institutions was formed for this project. The following table
describes the role and responsibility of each member institution:
Table ES-1
Role of Project Participants
Participant
Responsibilities
EPRI
P ro je c t m a n a g e m e n t a n d r e s e a r c h on e v e n t d e t e c tio n a n d trac king , inter­
a r e a oscillation m onitoring a n d e a rly w a rn in g of a n g u l a r stability p r o b le m s .
TVA
S o f tw a r e installation, in tegration a n d d e m o n s t r a t i o n ( s u p p o r t e d by Grid
P ro te c tio n Alliance).
University of Tennessee
D e v e l o p m e n t of t h e L ocation of D is tu r b a n c e (LOD) m o d u le .
HTC Tech
S o f tw a r e d e v e l o p m e n t a n d te stin g .
Quanta Technology
C o n s u ltin g for s o f tw a r e integration, te s tin g a n d training.
ISO New England
P ro v id e a d v i c e a n d c o s t s h a r e s th r o u g h E P R I s u p p l e m e n t a l pro jec ts.
This project was carried out in three phases.
• Phase 1 consisted in the analytical and theoretical studies to develop practical technologies for
early warning of potential stability problems and efficiently handling and transferring of large
volumes of real-time synchrophasor data. This phase commenced in October 2009 and
extended to approximately July 2010.
• Phase 2 was the prototype-development phase. The work in this phase focused on functional
specification of the proposed situational awareness system, software development, system
integration, and a pilot demonstration. EPRI designed the software architecture and proposed
synchrophasor-based stability analysis algorithms in the second half of year 2010. In October
2011, EPRI conducted testing studies with historical PMU data provided by TVA. Phase 2
was completed in December 2011.
• Phase 3 consisted of large-scale demonstration of the developed WASA software at the TVA
system using both historical and real-time synchrophasor data. The system interfaced with
TVA’s PDC (Phasor Data Concentrator). The final version of the software was delivered to
TVA and funders in December 2012.
The W A SA System
Application Modules
Figure ES-1 depicts the structure of the developed wide-area situational awareness (WASA)
system.
vi
Real-time PMU Data Stream in IEEE C37.118
Online Event
Detection
Location of
Disturbance
Near Real-Time
Event Replay
Early Warning
Power System
Visualization
Figure ES-1
Application Modules of the PMU-based WASA System
The WASA system provides five functional modules to enable real-time situational awareness
using wide-area synchrophasor measurement data:
Power System Visualization . This module is a graphical user interface (GUI) of the power
system visualization and early warning applications developed using Microsoft Windows
Presentation Foundation (WPF) and with the zooming and panning capability. It includes the
following main features: Voltage contour displays, phase angle contour displays, frequency
contour displays, angle differences between selected locations, trending charts, dashboards (users
can specify their own dashboard for reliability monitoring), and displays for early warning
functions.
Online Event Detection: The online event detection is used for detecting a new large system
event such as a large generator tripping or large load outages using real-time PMU
measurements. When a new large event is detected, the location of disturbance (LOD)
application is launched to identify the location, the event time, the magnitude in MW, and the
type of the new disturbance [16] [17], As soon as a new event is detected, the new event
information (time and event message) will be displayed on the visualization displays of each
user’s computer. When the LOD application completes the LOD, the results—including the type,
magnitude (MW), and the location of the event—will be shown in the visualization display.
Near Real-Time Event Tracking. It will be critically important for power system operators and
reliability coordinators to perform near-real-time event tracking with full resolution (up to 30
samples per second) when a large disturbance occurs to improve the operator situation
vii
awareness. The near-real-time event tracking using the real-time PMU measurements will help
the power system operators, managers, and engineers to quickly understand and analyze the
ongoing events and take appropriate corrective or preventive control actions if possible to
prevent large-scale cascading system outages.
Early Warning o f Potential Wide-Area Stability Problems. Predicting inter-area angular
instability and taking mitigation actions in real time for a large-scale power system poses a
complex problem. Some analyses on the system’s nature in topology and dynamics can be
conducted offline to simplify the problem. The approach applied in the WASA system predicts
potential generation out-of-step (OOS). This functional module provides the following
information to operators for early warning of the potential OOS: coherent generation groups,
modal information, vulnerable interfaces, risk of potential OOS, and decision support.
Location o f Disturbance. Power system events such as generator trips, load shedding, and
oscillations create perturbations in the voltage, frequency, and angle of the grid. These
disturbances propagate throughout the electrical network in time and space. By importing and
further analyzing time-synchronized event measurement data generated by the event-detection
module of EPRI visualization software, the LOD module determines event type and estimates
event size and location.
System Architecture
Figure ES-2 depicts WASA software architecture. As can be observed in this figure, it can be
used as either a local application or web application.
openPDC
Webserver
Data Transfer
I
PMU Data R eader
Application S ervices
"with Memory R esidence
D atab ases
Using IEEE C37.118 Protocols
Data Conditioning
Visualization
Web Service
Location of disturbance
Event Oriented
Application D atabase
Instability early Warning
WA PS Visualization Application
Via Internet or Intranet
Figure 1-1
WASA System Architecture
Application Server
Data Handling and Conditioning
The transfer of real-time PMU measurements is based on the need of the applications. Only the
PMU measurements (including voltages, phase angles, and frequencies) required for the
visualization and early warning applications are transferred to the users’ computers.
OpenPDC, which is an open source phasor data concentrator (PDC) developed by Grid
Protection Alliance (GPA), is used in this application. OpenPDC retrieves, processes, and
manages the real-time or historical PMU measurements from other PDCs or directly from PMUs.
PDC Connection Manager is a utility program for connecting the WASA system to the openPDC
and configuring the WASA database using the (1) real-time PMU data stream or from a PMU
data file in IEEE C37.118 format with configuration frame; (2) the PMU location data (longitude
and latitude); and (3) PDC connection data string and the port number used. The PDC connection
and the PMU information above are saved in the database.
The PMU Data Conditioning module provides comprehensive PMU data conditioning feature to
identify PMU data issues at PMU level or at PMU data point level.
Databases
Memory residence object-oriented database: The real-time PMU measurements are stored in the
memory residence object-oriented database in two first-in-first-out (FIFO) data object queues
with configurable parameters for the duration and the numbers of samples per second of the
PMU data. This feature significantly improves the performance.
Event-Oriented Database: The event-oriented database has unique features for efficiently
handling large amount of PMU measurements related to system events using binary large object
(BLOB) and data partitioning. The BLOB data in the database has a significant performance
advantage over the relational data during database access (write and read). The data partitioning
enables the event replay to start as soon as the first partition of data block is received in the client
and gives the user a quick response without waiting for downloading of the entire PMU data
related to the event.
Communication Technology
A new technology was develop to package and format the PMU data objects into an efficient
binary byte stream in the server and unpacked them back to data objects in the client with
minimal overhead. The same technique is applied in the data communication of real-time
monitoring and near-real-time event PMU data as well as database archiving of new or historical
events. The new approach implements the web service using Microsoft new technology of
Windows Communication Foundation (WCF) and hosting WCF in the same application service
as the in-memory database. This eliminates an extra communication overhead between web
service and the in-memory database if they are hosted separately.
Operating Modes
WASA can be operated in three different modes:
IX
Real-Time Monitoring: In this operation mode, the power system operators, engineers, or
managers can perform the reliability monitoring for the current operating condition using real­
time synchrophasor measurements. In this operation mode, the application tool will perform:
real-time reliability monitoring using visualization, trending charts and dashboards, online event
detection, and early warning of potential system angle stability problems.
Near Real-Time Event Tracking: In this operation mode, the power system operators, engineers,
or managers can perform online event tracking and replay in near real time (in a few seconds
after a new event is detected) or conduct offline post-event analysis for a selected system event.
The visualization in this operation mode will be able to use the full resolution of synchrophasor
measurements.
Post Event Analysis: In this operation mode, the power system operators, engineers, or managers
can select one of the historical events to perform post-event analysis.
Visualization Features
The visualization platform provides the following main features for a user to:
• Display real-time synchrophasor data on a geographic map at a 1-Hz refresh
rate.
• Provide the visualization displays for voltage magnitudes, phase angles, and frequencies.
•
Show defined angle difference links between PMUs.
• Pan and zoom the displays.
• Show trending charts for the selected data (voltage magnitude, phase angle).
• Show dashboards.
•
Show early warning related to visualization and information.
Test Results and Dem onstration Deploym ent
Different types of tests were conducted to evaluate the performance of the WASA system. The
main tests were as follows:
Off-line testing on a reduced system model. The different modules and algorithms were
demonstrated using a simplified model of the WECC system. The model consists of 29
generators, 179 buses, and 263 branches separated in five zones. In the analysis, it was assumed
that PMUs were installed at high-voltage substations. The PMU data—including voltage
magnitudes, phase angles, and frequencies of 30 samples per second—were created using a
stability program for different contingencies. The simulated PMU data were read and processed
as a PMU data stream in the near-real-time event replay mode and captured as a new event in the
database.
Offline demonstration using FNET data: The University of Tennessee maintains a wide-area
frequency measurement network called FNET based on distribution-level synchrophasors
(frequency disturbance recorders, or FDRs), which provide synchronized measurements at 10
samples per second. From April 25 to April 28, 2011, a tornado outbreak caused severe impacts
x
on power transmission systems in the southeast U.S. Oscillation events in TVA system were
captured by FNET during that tornado outbreak. There are about 70 FDRs monitoring the entire
Eastern Interconnection. Figure ES-3 shows the visualization on one oscillation event during that
tornado outbreak. The data are angle deviations (changes from the pre-disturbance values).
1 Synchrophasor-based Situational Awareness Tool J
2012-10-16162*06.579 jReal Ten* Mon.tor.ng
Fie
Control Panel
Tool
Ed*
Account ■
Angular StaMity Control Panel 2012-10-16102966.045
-
Display Option
A Separation Scenarios
S i Show Sut>i.ty-felated Information
Highest Risk Interface TVA - SOCO
Vi Show Separation Scenario
S3 Auto Tracking
7} Show Coherent Groups
$
/ Show Group Unk Data
Interface TVA-SOCO
Risk Angle dirt (deg.)
0608
Group 1
^
Gr0UP 2
0016
n
-14626
Phase drff (deg.) Freq (Hi) Damp(%)
148.441
LOO?
n
o
-2-255
* Visualization Option
UsTnKnoxsoter770
Datatype Voltage Angle *
M u t e r PMU 710 -IN1MPA_
g
V! Show Contour
583
•40 -50 -10 -M -20 -10 I
A historical Events
g -Kl
-SO -50 -40 - » -20 -10 0
Figure ES-3
Visualization of Historical FNET Data on Inter-Area Oscillations
Offline demonstration using simulated and historical TVA PMU data: PMUs data were created
by means of dynamic simulations. For this purpose, PMUs were assumed at 60 key TVA
substations. Also, the system was tested using historical TVA PMU data that captured an
oscillation event. The early warning algorithm correctly estimated the modal parameters of the
oscillations, as can be seen in Figure ES-4.
XI
a (D EE
Synchrophasor-based Situational Awareness Tool
2012-10-22 22:54:56.715
File
Control Panel
Tool
Edit
Account
E3
Help
10:54:56 PM
Mode
jl T|
=
0
RefGroupId
S3 I
|l Tj
-ls’o^lo'^E.'o -9.0 -70 -50 -30 -10 1.0 3.0 50 70 9.0 11.0 13.0 15.0 17.0 19.0 210
U-j
f
\ \
Freq(Hz) 1187
/
Angle Diff=11.6de
Ampl =0.43 deg
Phase Diff=20 deg
Freq=1.182 Hz
Damp=-0.35%
Angle Ditf=7.2 deg
Ampl =0.33 deg
Phase Diff=7 deg
Freq=1.184 Hz
Damp=-Q.15%
DataTvpe voltage Angle ▼
Refarece PMU 31 - 8WILSON 500
|V| ShowC
-listoncal Eventsl v
(
Display Option
»
[V] Shew Stability-related Information
[V] Show Separation Scenario
0 Show Coherent Groups
Replay Control
Separation Scenarios
Highest Risk: Interface 3-1+2
mAuto Tracking
[7| Show Group Link Data
Clean Risk Messages
20 - TVA_CUMBBUS1
20 - TVA_CUMBBUS1
Interface 3-1+2
0.042
p T m w i i m i i T P i 1*
T U 'W i'll i'l l LAMdl I Li 11
■ jm e u iH iu iiiii
E Group 3
-60 -50 -40 -30 -20 -10
3 -50 -40 -30 -20 -10 '
J -50 -40 -30 -20 -10
▼
Risk Angle diff (deg.)
20 - TVA.CUMBBUS1
□ Group 1-2
Ampl (deg.) Phase diff (deg.)
0262
-19.466
Freq (I-
-7.978
Damp(96)
1.142
Figure 1-2
Visualization of Historical TVA PMU Data
Field demonstration at TVA: The WASA system was successfully deployed at a server located
at the TVA control center, which is interfaced with openPDC at TVA to receive streaming PMU
data in IEEE C37.118 format in real time.
The openPDC at TVA retrieves and processes about 150 PMUs from TVA and other utilities in
the Eastern Interconnection. A snapshot is shown in the figure below.
W W H M J ffn iJ !
DataType Voltage Angle
0 Show Contour
Figure ES-5
Visualization of Real-Time TVA PMU Data
Field demonstration at ISO New England: This WASA system has also been deployed at a
server located at ISO New England to interface with ISO New England’s openPDC to receive
real-time PMU data. The openPDC at ISO New England retrieves and processes more than 60
PMUs from its transmission operators (TO). The example in the figure below shows a phase
angle contour display on ISO New England PMU data. The inter-area oscillations across an
interface (the dotted line) between the north and south regions were monitored. Real-time modal
information is shown on the interface and in the table. The two “phase clocks” indicate the mode
shapes on two detected oscillation modes at 0.2 Hz and 0.5 Hz. Trending charts on three selected
variables are shown in the figure.
xm
»
Synchrophasor-based Situational Awareness Tool
2012-10-13 1 6 * 0 03.233 Rm I Tune Momtonng
File
Control Panel
Tool
Edit
Voltage Angle
-MO 1 » 0 1300 -ISO -too -SO
__________________
Account
Help
00
100
SO
tSO
300
Angular StaMrty Control Panel 2012-10-13 16:-
3S0
' Display Option
•
V Show Stability-related Information
Visualization Contr...
5 j Show Separation Scenario
A Visualization Option
SBShow Coherent Groups
DataType Voltage Angle *
5 8 Show Group Unk Data
ReferecePMU CX.LUCXOW19S.2
Clean Risk Messages
V Show Contour
A Separation Scenarios
A Historical Events
Highest Risk. Interface O m ngton - South
Interface O m ngton - South
Risk Angle diff (deg.)
0.092
f ■
16.848
Dash Board
Group 1
GfOUp2
uSH
»
SB!
§
-27.632
0.421
3.033
-93.196
0.216
-2.614
4 :4002 PM a
Mode 1 «
ME.MAXCYS
td*9-> Ph*** * f f (deg.) Freq (Hz) Damp(%)
a i6 6
0-099
S
2
Ref Group Id 1 <
■ 4:4002 P Mode 2 »
a
0
S
Ref Group Id 1 J
«
|
—
*
: :: ! ■
iiiiiii
CX.DEERFIELD
9
1^.r.lVlflll
Freq(Hz) 0.421
Figure ES-6
Angle Contour Display Using ISO New England PMU Data
Conclusions and R ecom m endations for Further Investigation
An advanced synchrophasor-based wide-area situational awareness software system (WASA)
has been developed in this DOE R&D and demonstration project. The software is aimed to
provide wide-area power system visualization, near-real-time event tracking, and early warning
of potential stability problems for power system operators and reliability coordinators using real­
time or historical PMU measurements.
The application implements several advanced technologies, including the object-oriented
memory residence database, reduced PMU data sizing, and the enhanced event oriented database
using BLOB and data partitioning to efficiently handle large amount of real-time and historical
synchrophasor measurements and to support a large number of concurrent users. The WASA
configuration manager and the extensive data-conditioning features have greatly simplified the
software deployment, integration with the Phasor Data Concentrators, the software maintenance,
and upgrade.
WASA software has been extensively tested using the synchrophasor data of simulated 179
PMUs in the WECC system and the historical FNET data of large system events. It was
deployed and demonstrated at TVA and ISO New England using real-time synchrophasor
measurements from the openPDC Phasor Data Concentrators.
xiv
Future investigation to enhance the capability of the system should focus on wide-area
coordinating control, integrating the techniques applied by the WASA software. That system
would be based on highly reliable and secure communication infrastructures and highperformance computing resources for real-time optimization and validation of control actions.
xv
CONTENTS
1 INTRODUCTION.......................................................................................................................... 1-1
Project O bjective................................................................................................................................1-1
Motivation and Value of This P ro je ct..............................................................................................1-1
Report outline..................................................................................................................................... 1-2
2 PROJECT DEVELOPMENT........................................................................................................2-1
Project Main O utcom e...................................................................................................................... 2-1
Project T eam ...................................................................................................................................... 2-1
Project Activities and D evelopm ent................................................................................................2-2
3 DESCRIPTION OF THE WASA SYSTEM................................................................................. 3-1
Overview of the System .................................................................................................................... 3-1
Main Challenges.................................................................................................................................3-2
System Architecture.......................................................................................................................... 3-2
Functional Modules and Enabling Technologies.......................................................................... 3-3
Graphical User Interface............................................................................................................. 3-3
Real-Time Synchrophasor Data Handling................................................................................ 3-4
Application-Oriented Data H andling.................................................................................... 3-4
OpenPDC..................................................................................................................................3-4
PDC Connection M anag er.................................................................................................... 3-4
PMU Data Conditioning.......................................................................................................... 3-4
Online Event Detection and Location........................................................................................3-5
Near-Real-Time Event Tracking................................................................................................. 3-5
Post Event A nalysis..................................................................................................................... 3-5
Early Warning of Potential Wide-Area Stability Problems......................................................3-5
D atabases..................................................................................................................................... 3-6
Memory Residence Object-Oriented D atabase................................................................. 3-6
Event-Oriented Database.......................................................................................................3-6
xvii
Communication Technologies.................................................................................................... 3-7
Wide-Area Visualization................................................................................................................... 3-8
Visualization Features.................................................................................................................. 3-8
Wide-Area Visualization...............................................................................................................3-9
Trending C harts.......................................................................................................................... 3-12
Dashboard................................................................................................................................... 3-17
Early Warning of Wide-Area Angular Stability Problem s.......................................................... 3-18
Offline A nalysis........................................................................................................................... 3-18
Online Monitoring and Decision Support................................................................................ 3-20
Location of Disturbance.................................................................................................................. 3-22
Event Amount Estimation.......................................................................................................... 3-22
Event Location Triangulation.................................................................................................... 3-23
Time Difference of A rrival.................................................................................................... 3-23
Plant-Aided Method...............................................................................................................3-24
Least-Squares M ethod......................................................................................................... 3-25
Implementation of Inte rface......................................................................................................3-26
4 TEST RESULTS AND DEMONSTRATION DEPLOYMENT.................................................... 4-1
Performance Testing on the Early Warning A lgorithm ................................................................ 4-1
Performance Testing on PMU Data H andling............................................................................ 4-10
Offline Demonstration Using FNET D a ta .................................................................................... 4-12
Offline Demonstration Using Simulated andHistorical TVA PMU Data................................... 4-14
Results from Field Demonstration at T V A ................................................................................... 4-15
Results from Field Demonstration at ISO New England........................................................... 4-16
5 DISSEMINATION OF RESULTS/TECHNOLOGY TRANSFER...............................................5-1
Conference and Journal Papers..................................................................................................... 5-1
Patent Applications............................................................................................................................ 5-1
Participation in N A S P I...................................................................................................................... 5-1
Workshop at TVA facility in Chattanooga, T N ...............................................................................5-2
6 CONCLUSIONS AND RECOMMENDATIONS FOR FURTHER INVESTIGATION.............. 6-1
7 REFERENCES.................................................................................................................................1
xviii
LIST OF FIGURES
Figure 1-2 WASA System A rchitecture................................................................................................viii
Figure 1-4 Visualization of Historical TVA PMU D a ta .........................................................................xii
Figure 3-1 Application Modules of the PMU-Based W ASA S ystem ............................................... 3-1
Figure 3-2 WASA System A rchitecture...............................................................................................3-3
Figure 3-3 Snapshot of WASA System Running with Data from 179 Simulated WECC
P M U s............................................................................................................................................... 3-9
Figure 3-4 Phase Angle Display With One PMU Selected as the Common Reference
3-10
Figure 3-5 Phase Angle Display With a Different Common Reference....................................... 3-10
Figure 3-6 Voltage Magnitude Contour Display With Angle Differences....................................3-11
Figure 3-7 Frequency Contour Display With Angle Differences.................................................. 3-11
Figure 3-8 Voltage Magnitude Display W ithout C ontours..............................................................3-12
Figure 3-9 Trending Chart of an Angle Difference With Limit Violations Highlighted................ 3-13
Figure 3-10 PMU Measurement Trending Chart to Show One M easurement............................3-13
Figure 3-11 PMU Measurement Trending Chart to Show Two M easurem ents..........................3-14
Figure 3-12 Voltage Tending Chart for Multiple P M U s.................................................................. 3-14
Figure 3-13 Voltage Magnitude Trending Chart for Multiple PM Us.............................................. 3-15
Figure 3-14 Voltage Angle Trending Chart for Multiple P M U s......................................................3-15
Figure 3-15 Voltage Angle Trending Chart After Zooming in ........................................................ 3-16
Figure 3-16 Trending Chart on the Instability Risk Index and Angle Separation Between
Two A re a s..................................................................................................................................... 3-16
Figure 3-17 Trending Chart on the Oscillation Frequency and Damping Ratio of a
Selected M o d e ............................................................................................................................. 3-17
Figure 3-18 Dashboard Integrating Trending Charts and Wide-Area Visualization..................3-17
Figure 3-19 Proposed A pproach........................................................................................................3-18
Figure 3-20 Generator Clustering in the Eastern Interconnection................................................ 3-19
Figure 3-21 Phase Clock on One Inter-Area Mode......................................................................... 3-21
Figure 3-22 Event Type Determination..............................................................................................3-23
Figure 3-23 Illustration of Frequency Propagation in a Generation Trip Event....................................3-24
Figure 3-24 Illustration of Plant-Aided M ethod................................................................................... 3-25
Figure 3-25 Illustration of Least-Squares M ethod........................................................................... 3-26
Figure 4-1 179-bus power system ........................................................................................................4-1
Figure 4-2 Angles Monitored by 12 Synchrophasors........................................................................4-3
xix
Figure 4-3 FFT Results Over 40-80s, 80-120s, 120-160s, and 160-200s.....................................4-4
Figure 4-4 Angle Distances................................................................................................................... 4-4
Figure 4-5 Phase Differences................................................................................................................4-5
Figure 4-6 Risk Indices Without Remedial A ctio ns........................................................................... 4-5
Figure 4-7 Angle Distances With Remedial A c tio n s......................................................................... 4-6
Figure 4-8 Phase Differences With Remedial A ctio ns......................................................................4-6
Figure 4-9 Risk Indices With Remedial A ctions................................................................................. 4-7
Figure 4-10 Phase Clock on an Inter-Area M o d e ..............................................................................4-7
Figure 4-11 Visualization Before the Final C ontingency.................................................................. 4-8
Figure 4-12 Visualization After Load Shedding.................................................................................. 4-9
Figure 4-13 Frequencies and Angles After the Grid
IsSeparated on Interface 1-234............. 4-10
Figure 4-14 Visualization of Historical FNET Data on Angle D eviations.....................................4-13
Figure 4-15 Visualization of Historical FNET Data on Inter-Area O scillations............................. 4-13
Figure 4-16 Visualization of Simulated TVA PMU D ata................................................................. 4-14
Figure 4-17 Visualization of Historical TVA PMU D a ta .................................................................. 4-15
Figure 4-18 Visualization of Real-Time TVA PMU D a ta ................................................................ 4-16
Figure 4-19 Voltage Contour Display Using ISO New England PMU D a ta ................................ 4-17
Figure 4-20 Angle Contour Display Using ISO New England PMU D a ta ....................................4-18
xx
LIST OF TABLES
Table 2-1 Role of Project Participants.................................................................................................2-2
Table 2-2 Project Phases and M ilestones.......................................................................................... 2-3
Table 3-1 Input and Output of LOD Module.....................................................................................3-27
Table 4-1 Generator Clustering........................................................................................................... 4-2
Table 4-2 Benchmark Testing Result on Simulated 179PMUs (on a Laptop Com puter)
4-11
Table 4-3 Benchmark Testing Result Using aLaptop Connected To an Application
Server via Secured VP N ............................................................................................................. 4-12
xxi
LIST OF ACRONYMS
BLOB: Binary Large Object
GPA. Grid Protection Alliance
GUI: Graphical User Interface
FIFO: In First Out
ISO: Independent System Operator
RTO: Regional System Operator
NERC: North American Reliability Corporation
NYSERDA
NASPI: North American SynchroPhasor Initiative
OOS: out-of-step
PDC: Phasor Data Concentrator
PMU: Phasor Measurement Units
TVA: Tennessee Valley Authority
WASA: Wide Area Situation Awareness
WECC: Western Electricity Coordinating Council
WCF: Windows Communication Foundation
1
INTRODUCTION
Project Objective
The purpose of this project was to develop and demonstrate a novel synchrophasor-based
comprehensive situational awareness system for control centers of power transmission systems.
The situational awareness system consists of four main software modules:
•
Wide-area visualizations of real-time frequency, voltage, and phase angle measurements and
their contour displays for security monitoring.
•
Online detection and location of a major event (location, time, size, and type, such as a
generator or line outage).
•
Near real-time event replay (in seconds) after a major event occurs.
•
Early warning of potential wide-area stability problems.
The developed system is intended to improve situational awareness at control centers of the
power system operators and regional reliability coordinators. It would provide effective widearea power system monitoring and visualization utilizing real-time synchrophasor data and
enable system operators to assess the stability levels of their power grids in real time, become
better aware of on-going disturbance events, and receive early warning of potential wide-area
stability problems in order to take appropriate control actions to mitigate the risk.
Motivation and Value of This Project
Since the large North Eastern power system blackout on August 14, 2003, U.S. electric utilities
have spent lot of effort on preventing power system cascading outages. Two of the main causes
of the August 14, 2003 blackout were inadequate situational awareness and inadequate operator
training [1], In addition to the enhancements of the infrastructure of the interconnected power
systems, more research and development of advanced power system applications are required for
improving the wide-area security monitoring, operation, and planning in order to prevent largescale cascading outages of interconnected power systems. It is critically important for improving
the wide-area situation awareness of the operators or operational engineers and regional
reliability coordinators of large interconnected systems. With the installation of a large number
of phasor measurement units (PMU) and the related communication infrastructure, it will be
possible to improve the operators’ situation awareness and to quickly identify the sequence of
events during a large system disturbance for the post-event analysis using the real-time or
historical synchrophasor data.
With the funding available by the American Recovery and Reinvestment Act of 2009 and the
large investments of the electricity utilities, a large number of PMUs as well as the required
1-1
Introduction
communication infrastructure and the related online applications are being implemented in the
Eastern Interconnection (El), WECC, and ERCOT. The North American SynchroPhasor
Initiative (NASPI) is collaborating with utilities, ISOs/RTOs, NERC, transmission companies,
researchers, and vendors to create a robust and secure synchronized measurement infrastructure
and the related monitoring, operation, and analysis tools for better power system operations and
planning, and improved system reliability [2] [3][4] [5],
There are significant and tangible benefits from the practical uses of the real-time GPS
synchrophasor measurements. Transmission grid operators may have greater situational
awareness on the potential risk of cascading outages and may have useful diagnosis and analyses
on how to respond to the potential risks. However, advanced computational, visualization, and
data-management tools are needed to materialize the benefits from synchronized measurement
systems. In the last few years, a lot of research and development effort has been undertaken to
develop applications utilizing synchrophasor measurements for real-time monitoring, state
estimation, stability control, and post-event analysis of interconnected power systems. EPRI has
performed several R&D projects for wide-area power system visualization using PMU
measurements since 2006 [7][9][11][19] [28],Virginia Tech and EPRI have developed advanced
applications to perform on-line triggering and to identify the location of disturbances (LOD) of
power systems using the synchronized frequency measurements of the synchrophasor
measurements [22] [23], EPRI has been working with NYSERDA, TVA, and Virginia Tech to
develop a wide-area real-time security monitoring and visualization application using
synchrophasor measurements [9][11] [12] [28],
In this project, EPRI, along with the project team participants, developed an advanced wide-area
situation awareness (WASA) system using real-time synchrophasor measurements. The system
has been deployed and demonstrated at Tennessee Valley Authority (TVA) and ISO New
England systems using real-time synchrophasor data from openPDC. The WASA configuration
manager and the extensive data conditioning features have greatly simplified software
deployment, maintenance, and update. The applications have implemented several advanced
technologies, including the object-oriented memory residence database, reduced PMU data
sizing, and the enhanced event oriented database using BLOB and data partitioning to efficiently
handle large amount of real-time and historical synchrophasor measurements and to support
large number of users.
Transmission grid operators, ISO, RTO, reliability coordinators, operating engineers,
transmission planners, and organizations with investment in the installation of PMUs can
certainly benefit from this new system and the results of this project. Indeed, the novel WASA
application can help improve wide-area power system monitoring and visualization at control
centers. It provides system operators the ability to assess the stress on their power grids, become
more aware of ongoing events, obtain early warning signals of potential grid problems, and then
take timely corrective or preventive control actions.
Report outline
This report is structured as follows:
Section 2 describes the project development, including project main outcomes, project team,
project activities, and phases.
1-2
Introduction
Section 3 presents a detailed description of the WASA software. First, an overview of the system
architecture is provided. Each of the modules that comprise the software is described in detailed,
including the mathematical formulation and the computation algorithm. This section also
describes the main challenges and difficulties encountered during the project development.
Section 4 deals with software test results and demonstration deployment. The topics addressed in
this section include: Performance testing of the early warning algorithm, performance testing of
the PMU data-handling techniques, offline demonstration using TVA’s simulated and historical
PMU data, results from the field demonstration at TVA system, and results of the field
demonstration at ISO New England.
Section 5 discusses the main activities conducted by the project team to disseminate the project
value and results and transfer the developed technology.
Finally, Section 6 provides the main conclusions and suggests actions for further development
and future research.
1-3
2
PROJECT DEVELOPMENT
This section describes the project main outcome, the project team, and project activities for the
entire period of funding, describing how the work evolved from the initial hypotheses, to the
methodology formulation, and to preliminary tests towards the final implementation of the
developed software.
Project Main Outcome
The main deliverable of this project is the wide-area situational awareness (WASA) software.
The intent of WASA is to furnish power system operators and reliability coordinators with a
superior tool for improving real-time system operating security. The WASA system provides
five functional modules to enable real-time situational awareness using wide-area synchrophasor
measurement data:
•
Real-time security monitoring using real-time PMU data .
•
Near-real-time event tracking (in few seconds after a major event, such as loss of a large
generator or load, occurs).
•
On-line event detection and location using wide-area frequency measurements.
•
Post-event replay and analysis.
•
Early warning of potential wide-area stability problems.
WASA software has been deployed and demonstrated at TVA and ISO New England systems.
Apart from the software product, the outcome of this project consists of a set of technical reports
and papers describing the mathematical foundations and computational approaches of different
tools and modules, implementation issues and considerations, lessons learned, and the results of
validation processes.
Project Team
EPRI assembled a highly qualified, experienced, and multidisciplinary team from five
prestigious institutions to carry out this project. The following table describes the role and
responsibility of each member institution:
2-1
Project Development
Table 2-1
Role of Project Participants
Participant
R esponsibilities
EPRI
Project management and research on event detection and tracking,
inter-area oscillation monitoring, and early warning of angular
stability problems.
TVA
Software installation, integration, and demonstration (supported by
Grid Protection Alliance).
University of
Tennessee
Development of the Location of Disturbance (LOD) module.
HTC Tech
Software development and testing.
Quanta Technology
Consulting for software integration, testing, and training.
ISO New England
Provide advice and cost shares through EPRI supplemental projects.
Project Activities and Development
This project was carried out in three phases. The initial phase, Phase 1, consisted of the
analytical and theoretical studies to develop practical technologies for early warning of potential
stability problems and efficiently handling and transferring of large volumes of real-time
synchrophasor data. This phase commenced in October 2009 and extended to approximately July
2010 .
After the basic foundations and initial tests were performed, Phase 2—the prototype
development phase— started -. This phase focused on prototype studies aimed at functional
specification of the proposed situational awareness system, software development, system
integration, and a pilot demonstration. EPRI designed the software architecture and proposed
synchrophasor-based stability analysis algorithms in the second half of year 2010. In October
2011, EPRI conducted testing studies with historical PMU data provided by TVA. Phase 2 was
completed in December 2011.
Phase 3 consisted of large-scale demonstration of the developed WASA software at TVA’s
system using both historical and real-time synchrophasor data. The system interfaced with
TVA’s PDC (Phasor Data Concentrator). The main steps and milestones in the implementation
process were as follows:
2-2
•
EPRI delivered the WASA beta release to TVA in February 2012.
•
EPRI worked with TVA, GPA, and Quanta Technology to install WASA onto a TVA
computer in May 2012.
•
WASA was deployed and demonstrated at ISO New England using real-time
synchrophasor data in August 2012.
Project Development
•
TVA hooked up the server to real-time PMU data in October 2012.
•
A real-time demonstration was conducted during the open workshop held at the TVA
office in Chattanooga on October 23rd, 2012.
•
After that date, the software was tested by TVA’s reliability coordinator and system
operator.
•
EPRI and the project team provided technical support.
•
The final version of the software was delivered to TVA and funders in December 2012.
Table 2-2 summarizes the activities (tasks) at each of the project phases.
Table 2-2
Project Phases and Tasks
Phase
Phase 1:
Analytical Study
Tasks
1. Research on early warning of stability
problems.
Periods
From 10/01/2009
to
2. Development of efficient methods on data
processing and transfer for visualization.
07/31/2010
Phase 2:
3. Functional specification.
From 07/01/2010
Pilot Study
4. Software development.
to
5. System integration and offline testing using
historical or simulated data.
12/31/2011
6. Installation and demonstration at TVA and
ISO New England systems.
From 01/01/2012
(Research)
(Development)
Phase 3:
Large-Scale
Demonstration
7. Technical training and documentation.
to
12/31/2012
8. Technical support for 6 months.
2-3
3
DESCRIPTION OF THE WASA SYSTEM
This section provides a detailed description of the WASA software. First, an overview of the
system architecture is provided. The main challenges and difficulties encountered during the
software formulation and development are then described. The section continues with the
description of each of the modules that comprise the software package.
Overview of the System
Figure 3-1 depicts the structure of the developed wide-area situational awareness (WASA)
system.
Real-time PMU Data Stream in IEEE C37.118
I
I
)
Online Event
Detection
Location of
Near Real-Time
Disturbance
Event Replay
Early Warning
Power System
Visualization
Figure 3-1
Application Modules of the PMU-Based WASA System
As shown in Figure 3-1, the WASA system provides five functional modules to enable real-time
situational awareness using wide-area synchrophasor measurement data:
3-1
Description o f the WASA system
•
Real-time security monitoring using real-time PMU data.
•
Near-real-time event tracking (in a few seconds after a major event, such as loss of a
large generator or load, occurs).
•
On-line event detection and location using wide-area frequency measurements.
•
Post-event replay and analysis.
•
Early warning of potential wide-area stability problems.
Using the situational awareness system, system operation engineers at control centers would be
able to online assess the stress on their power grids, become more aware of ongoing disturbance
events, and receive early warning signals of potential stability problems such that early control
actions could be taken.
Main Challenges
The main challenges and difficulties faced in developing this situational awareness system are
described as follows:
•
Transferring a large volume of PMU measurements from openPDC to the application
server.
•
Inserting the event-related PMU measurements into and a relational database when a
large event occurs.
•
Transferring a large volume of PMU measurements at a high sampling rate (such as 30
sample per second) related to an event from the application server to users’ computers/
laptops via secured Internet or intranet or a dedicated communication network
particularly for tracking the ongoing events.
•
Presenting useful and high-fidelity visualization displays, including voltage contour
displays, frequency contour displays, phase angle contour displays, angle difference
links, trending charts, and dashboards using real-time or historical PMU data.
•
Analyzing real-time PMU measurements to provide early warnings of potential stability
problems for operators to take appropriate corrective or preventive control actions if
necessary.
System Architecture
WASAT can be used as either a local application or web application. For the latter, Figure 3-2
gives a feasible software architecture, which consists of the following servers, which could be
undertaken by the same computer server:
3-2
Description o f the WASA system
•
Application server: This application server has key software components and two
databases installed. One database is an object-oriented memory residence database, and
the other is an event-oriented application database.
•
Web Server: Provides web services by software, such as Microsoft IIS, to client
computers to remotely access the databases and visualize synchrophasor data, events, and
analysis results.
o penPDC
W e b se rver
Data Transfer
PMU D ata R e a d e r
A pplication S e rv ic e s
with M em ory R e sid en ce
D a ta b a s e s
Using IEEE C37.118 Protocols
Data Conditioning
I
V isualization
W eb Service
On-line Event Detector
Location o f d istu rba nce
E v en t O riented
A pplication D a ta b a s e
In sta b ility early W arning
WA PS Visualization Application
Via Internet or Intranet
A p p lic a tio n Server
Figure 3-2
WASA System Architecture
Key elements of this system are described in the following subsections:
Functional Modules and Enabling Technologies
Graphical User Interface
The graphical user interface (GUI) of the power system visualization and early warning
applications was developed using Microsoft Windows Presentation Foundation (WPF) and with
the zooming and panning capability. It includes the following main features:
•
Voltage contour displays.
3-3
Description o f the WASA system
•
Phase angle contour displays.
•
Frequency contour displays.
•
Angle differences between selected locations.
•
Trending charts.
•
Dashboards (users can specify their own dashboard for reliability monitoring).
•
Displays for early warning functions.
Real-Time Synchrophasor Data Handling
Application-Oriented Data Handling
The transfer of real-time PMU measurements is based on the need of the applications. Only the
PMU measurements (including voltages, phase angles, and frequencies) required for the
visualization and early warning applications are transferred to the users’ computers. When there
is no large system event, it may be sufficient to transfer one sample of the PMU measurements
for real-time monitoring. For the near-real-time event replay mode and the post event analysis
mode, the full resolution of the real-time PMU measurements are transferred from the
application server to the user’s computer via secured Internet or intranet.
OpenPDC
The openPDC is an open source phasor data concentrator (PDC) developed by Grid Protection
Alliance (GPA). The openPDC retrieves, processes, and manages the real-time or historical
PMU measurements from other PDCs or directly from PMUs.
PDC Connection Manager
This is a utility program for connecting the WASA system to the openPDC and configuring the
WASA database using the (1) real-time PMU data stream or from a PMU data file in IEEE
C37.118 format with configuration frame; (2) the PMU location data (longitude and latitude);
and (3) PDC connection data string and the port number used. The PDC connection and the
PMU information above are saved in the database.
PMU Data Conditioning
PMU measurements may contain bad measurements, some of the PMU measurements may be
missing, or a PMU may not be functioning. This module provides a comprehensive PMU data
conditioning feature to identify PMU data issues at the PMU level or at the PMU data point
level. A quality flag will be set to invalid for a malfunctioning PMU or a bad PMU data point.
When a PMU measurement is temporarily missing, it will be replaced, if possible, by its previous
value.
3-4
Description o f the WASA system
Online Event Detection and Location
The online event detection is used for detecting a new large system event such as a large
generator tripping or large load outages using real-time PMU measurements. When a new large
event is detected, the location of disturbance (LOD) application is launched to identify the
location, the event time, the magnitude in MW, and the type of the new disturbance [16] [17], As
soon as a new event is detected, the new event information (time and event message) will be
displayed on the visualization displays of each user’s computer. When the location of
disturbance (LOD) application completes the LOD, the results—including the type, magnitude
(MW), and the location of the event—will be shown in the visualization display.
Near-Real-Time Event Tracking
It will be critically important for power system operators and reliability coordinators to perform
near-real-time event tracking with full resolution (up to 30 samples per second) when a large
disturbance occurs to improve the operator situation awareness. The near-real-time event
tracking using the real-time PMU measurements will help the power system operators, managers,
and engineers to quickly understand and analyze the ongoing events and take appropriate
corrective or preventive control actions, if possible, to prevent large scale cascading system
outages.
Post Event Analysis
The event related PMU data required for visualization and early warning applications are stored
in the database in special binary format and in multiple partitions. Each user can select one of
these events to perform the post-event analysis by transferring the related event data from the
event database to the users’ computers in compact binary format and on a partition basis so that
the user can start the event replay as soon as the first partition of the event data is available.
Early Warning of Potential Wide-Area Stability Problems
Predicting inter-area angular instability and taking mitigation actions in real time for a largescale power system pose a complex problem. Some analyses on the system’s nature in topology
and dynamics can be conducted offline to simplify the problem. The approach applied in the
WASA system predicts potential generation out-of-step (OOS). The detailed approach is
introduced later in this report. This functional module provides the following information to
operators for early warning of the potential OOS:
•
Coherent generation groups: Coherent generation groups identified offline using EPRI
DYNRED program are visualized to show their locations and territories. Each group can
be monitored by one PMU or a group of PMUs.
•
Modal information: Modal analysis is performed online on PMU data over a 30- to 60second time window about angle differences between coherent generator groups to
identify dominant inter-area modes and estimate modal parameters, such as oscillation
frequencies, damping, and mode shapes.
3-5
Description o f the WASA system
•
Vulnerable interfaces: Mode shapes tell dynamic clustering of all PMUs, indicating
how generation groups swing together or against each other. Thus, network interfaces
vulnerable to OOS are identified.
•
Risk of potential OOS: Across each vulnerable network interface, the angle difference
will be compared to a threshold in real time to estimate a risk index of OOS.
•
Decision support: Once the risk of OOS on any interface becomes high, a mitigation
action is suggested from a strategy table (developed in the offline stage for representative
scenarios of instability). The strategy may reconfigure power flows by load/generation
shedding to relieve the stress on the interface or even controlled separation of the system
on the interface.
Databases
Memory Residence Object-Oriented Database
In order to significantly improve the performance of the visualization application, the real-time
PMU measurements are stored in the memory residence object-oriented database in two first-infirst-out (FIFO) data object queues with configurable parameters for the duration and the
numbers of samples per second of the PMU data. One of the object queues is used for the real­
time security monitoring and the second one is used for the near-real-time event tracking. A
reduced resolution (such as one sample per second) is used for the first object queue to improve
the performance and reduce the system resource and communication requirements. The second
object queue has a variable length to cover the entire duration of a system event with reduced or
full-resolution data objects (configurable) and is activated only when a new system event is
detected.
Event-Oriented Database
The event-oriented database, typically deployed at the application server, is a relational database
developed using Microsoft SQL Server 2005 or 2008. The event-oriented database has unique
features for efficiently handling large amounts of PMU measurements related to system events
using binary large object (BLOB) and data partitioning. The BLOB data in the data base has a
significant performance advantage over the relational data during database access (write and
read). The data partitioning enables the event replay to start as soon as the first partition of data
block is received in the client and gives the user a quick response without waiting for
downloading of the entire PMU data related to the event [20], When a new event is detected, the
PMU measurements are converted into BLOB in byte-arrays format with multiple partitions.
Each partition consists of a configurable number of PMU data objects. The event-oriented
database stores only the PMU measurements, including voltages, phasor angles, and frequencies
required for the visualization and early warning applications.
3-6
Description o f the WASA system
Communication Technologies
During the extensive testing, it was found that the main performance bottleneck was related to
the transfer of the large size of the PMU data from the application server to the user’s computer
in the real-time event replay mode or post-event replay modes. Each PMU data in a data frame is
represented by an instance of a data object in order to be processed by object-orientated
programming. However, the object serialization process using conventional method is not
efficient, and the resulting buffer is not compact.
In order to improve the communication performance for large PMU data, a new technology was
develop to package and format the PMU data objects into an efficient binary byte stream in the
server and unpacked them back to data objects in the client with minimal overhead. The same
technique is applied in the data communication of real-time monitoring and near real-time event
PMU data as well as database archiving of new or historical events.
The test results showed that the new technique significantly reduces the PMU data buffer size by
about 80% compared to the conventional object serialization approach presented in [11], This
new technique, plus the implementation of new communication infrastructure, dramatically
reduces the times required for managing, transferring, and processing the PMU data and
improves the overall system performance.
The clients access the application data through a web service request. Traditionally and in an
earlier implementation [11], the web service is hosted in a web server, which in turn makes
another inter-process communication request to the memory resident database that is hosted by a
separated application service. In order to reduce the communication overhead, a new approach
was developed to implement the web service using the Microsoft new technology of Windows
Communication Foundation (WCF) and host WCF in the same application service as the in­
memory database. This eliminates an extra communication overhead between web service and
the in-memory database if they are hosted separately.
Operation modes
This software system can be operated at three modes.
Real-Time Monitoring
In this operation mode, the power system operators, engineers, or managers can perform the
reliability monitoring for the current operating condition using real-time synchrophasor
measurements. In this operation mode, the application tool will perform the followings:
•
Real-time reliability monitoring using visualization, trending charts, and dashboards.
•
On-line event detection.
•
Early warning of potential system angle stability problems.
In this mode, real-time synchrophasor data will be transferred from the PDC to the application
server.
3-7
Description o f the WASA system
Near-Real-Time Event Tracking
In this operation mode, the power system operators, engineers, or managers can perform online
event tracking and replay in near real time (in few seconds after a new event is detected) or
conduct offline post-event analysis for a selected system event. The visualization in this
operation mode will be able to use the full resolution of synchrophasor measurements.
When a new large event is detected, the synchrophasor data related to the event will be
transferred from the PDC to the event-oriented data base for the near-real-time event replay. For
post-event analysis, a user can select one of the existing events stored in the database to perform
post-event analysis using wide area power system visualization.
Post-Event Analysis
In this operation mode, the power system operators, engineers, managers can select one of the
historical events to perform post-event analysis. The WASA software can be interfaced to a
Phasor Data Concentrator (PDC), such as OpenPDC, to obtain real-time synchrophasor data.
Wide-Area Visualization
A wide-area visualization platform integrating the GUI, databases, and other technologies for
efficient synchrophasor data handling and communication was developed and field-demonstrated
under this project. Main features of this platform are introduced as follows.
Visualization Features
The visualization platform provides the following main features for a user to:
•
Display real-time synchrophasor data on a geographic map at a 1-Hz refresh rate.
•
Provide the visualization displays for voltage magnitudes, phase angles, and frequencies.
•
Show defined angle difference links between PMUs.
•
Pan and zoom the displays.
•
Show trending charts for the selected data (voltage magnitude, phase angle).
•
Show dashboards.
•
Show early warning related visualization and information.
Figure 3-3 gives a sample display of the visualization platform, which is running with data from
179 simulated WECC PMUs.
3-8
Description o f the II'. IN. 1system
2012-10-22 225036,441 I
1= a - i
■ Angular Stability Control Panel 2012-10-22 22:30.36.403
( A) Display Option
[Vj Show Stability-related Information
■ 1050:36 PM
«■! W
.
RefGroupId
DataType | Voltage Angle t
®
Separation Scenarios
Highest Risk: Interface 14-23
E l Show Separatio
|E] Show Coherent
iV] Show Group Li
Ib I Autjo Tracking
0 Show Con
Interface 1-234
-e
Risk Angle diff (deg.)
0311 453.7
( 3 Group
Ampl (deg.) Phase diff (deg.) Freq (Hz)
7.501
122.634
0.228
Damp(%)
5.609
-
■■
= SSSSSS'i'T'pS
5 niSSiiilillBiii
s
■
Angle Diff=12.6 deg
Ampl=3.82 deg
Phase Diff=126 deg
Freq=0.230 Hz
Damp=1.21%
aiding Chardnlerface 1-234
lyl Auto Treed Refresh Legend I Docking Insid
Trending for Interface 1-23410122/201222:30:34.408 Eastern Daylight Tim
Angle Diff=50.5 deg
Ampl=0.00 deg
Phase Diff=1 deg
Freq=0.000 Hz
Damp=0.00%
™ ssi» 2 ie 3
-528.08^1
|7.9
-3 *
4
-3 * 6 5
-24.940.7
-36
-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
-10
-5
0
Time (Second)
Figure 3-3
Snapshot of WASA System Running with Data from 179 Simulated WECC PMUs
Wide-Area Visualization
The visualization platform may display one from the wide-area visualizations (with or without
contours) of bus voltage magnitudes, bus phase angles, and frequencies. For a phase angle
display, the user needs to select the phase angle of a PMU as common reference for the
visualization displays. The angle visualization display will be updated in real time based on the
newly selected common phase angle reference. The following figures illustrate visualizations
with different common angle references for the same area. The data are from 60 simulated TVA
PMUs. More sample visualizations are shown in Figure 3-6 to Figure 3-8.
3-9
Description o f the II'. IN. 1system
!■
S y n c h ro p h aso r-b asc d S itu atio n al A w aren ess Tool
I 2 0 1 1 -1 1 -2 8 15:54:02.698
File
C o n tro l Panel
V o lta g e Angle,
20.0 -15.0 ” -10.0
E vent Replay [
Tool
-5.0
Edit
0.0
Help
5.0
10.0
15.0
20.0
25.0
300
35.0
A n g le D ifference Ratio
inna
ms
nvt
0.00
”
0.15
0.30
0.45
Figure 3-4
Phase Angle Display With One PMU Selected as the Common Reference
5 7 S y n c h ro p h a so r-b a se d Situatio n al A w aren ess Tool
2011-11-28 15:53:37.030 Event Rep 1.1 [
File
C o n tro l Panel
V o lta g e A n g le
-200
-ISO
-100
Tool
-SO
Edit
0,0
H elp
SO
LOO
ISO
20L0
25.0
30.0
3S.0
A n g le D ifference Ratio
0.00
0.15
0.30
Figure 3-5
Phase Angle Display With a Different Common Reference
3-10
Description o f the WASA system
0
Sy n ch ro p h aso r-b ase d S ituational A w areness T o o l^
•
•
•
File
C o n tro l Panel
Tool
V o ltag e M a g n itud e
0.94
036
097
Edit
0.99
'-------II | ~
.
20 1 1 -1 2-1 1 01:55:59.163 | Real Time M onitoring
I
-1
Help
LOO
1.02
L03
A ngle D ifference Ratio
0.00
0.15
030
L06
0.45
0.60
075
090
A
■
1.04
102
1.04
/ f 0 3 10
1.01
1.01
1827<;
m
/
Figure 3-6
Voltage Magnitude Contour Display With Angle Differences
r
Synchrophasor b ased Situational A w areness Tooj^
2011-12-11 01:54:18.661
File
Control Panel
Real Tim e M onitoring
Tool
Edit
H
Help
Frequence
59.750 59.800 59.850 59.900 59.950 60000 60050 60.100 60150 60.200 60250 60.300
Angle D ifference Ratio
000
015
0.30
0.45
0.60
075
090
LOS
1.20
59 997
59.998 5 9 .9 ^ 998
50.996
59996
59 997
59.997
59 997
59.997
59.997
59 997
59.997
59.997
59.998
S^.9^76 1
59.998
“ ” ^97
5® W 9 97
* » 7
S&90?
59.997
5 50S97
59 9®
«
58#986P-998
Figure 3-7
Frequency Contour Display With Angle Differences
3-11
Description o f the WASA system
I
S y n c h r o p h a s o r - b a s e d S itu a tio n a l A w a r e n e s s T ool
2 0 1 1 - 1 2 -1 1 0 3 :1 5 :4 1 .2 5 9
File
C o n tro l P a n e l
R eal T im e M o n ito rin g
Tool
V o lta g e M a g n itu d e
0.94
0.96
0.97
E d it
0.99
H elp
1.00
1.
28.41
Figure 3-8
Voltage Magnitude Display Without Contours
Trending Charts
The visualization platform provides the following types of trending charts for monitoring
dynamic changes of one or multiple specific quantities over time:
•
PMU measurement trending charts.
•
Angle difference trending charts.
•
Stability index trending charts.
Sample displays are shown in the figures below.
3-12
Description o f the WASA system
_) Phase A n g le D ifference Chart
[•/ A u to Trend | R e fre s h jp ra g to D a sh b o a rd )
P h a s e A ngle diffe re n ce 1 1/28 /2 0 1 1 15:53:59.298 P acific S tan d a rd Tim e
32
p
3 0 .6
r
2 9 .2
I
2 7 .8
L
2 6 .4
■
________________________________ _____
25
M
M
i i i i i i B
■
1
-3 0
-2 8
-2 6
-2 4
-2 2
-2 0
-1 8
-1 6
-1 4
-1 2
-1 0
-8
-2
Tim e (Second)
Figure 3-9
Trending Chart of an Angle Difference With Limit Violations Highlighted
Q I rending
31 - 8WILSON 500 - VOLT
^_ ^ _ L l = _ * * l
v
Secondary None
-
0 Auto Trend Refresh Legend [Docking Inside ▼ Drag to Dashboard
Trending far 31 - 8WILSON 500 11/28/2011 15:53:58.698 Pacific Standard Time
Time (Second)
Figure 3-10
PMU Measurement Trending Chart to Show One Measurement
3-13
Description o f the WASA system
Primary: 31 - 8WILSON 500 -VOLT
,
Secondary: p i - 8WILSON 500 -ANGL
-
1171 Auto Trend Refresh Legend ( Docking Inside v Drag to Dashboard
Trending for 31 - 8WILSON 500 11/28/2011 15:53:53.364 Pacific Standard Time
1.1 3 -
1.016H
0.902
-
0.8
0.788
-5 .2
0.674
-9.6
1
0.56
-30
18
-28
-16
-14
-12
-10
-14
-6
-4
Time (Second)
Figure 3-11
PMU Measurement Trending Chart to Show Two Measurements
E
Trending
B
I S
E
Data Type: Voltage M agnitude
I °
T
[V] Auto Tre nd Refresh L sgend | Docking Inside
I ^
------ '
Drag to Dashboard
Trending for Voltage Ma gnitude 11/28/21 11 15:53:58,69 Pacific Standard Time Trending
oj
1 0 44
m
§
CTr
U.OdZ
>
a>
0 .4 2 6
o
-2
8
-2 6
-2>4
-2 2
-2 0
-1 8
-1 6
-1 4
Time (Second)
-----------------Figure 3-12
Voltage Tending Chart for Multiple PMUs
3-14
-1 2
-1 0
-8
6
4
2
C
»
Description o f the WASA system
U
m m m tm
Trending
^
Z
Data Type: V o lta g e M a g n itu d e
.
3
A u to Trend
Refresh
'1
N
Legend [ D ockin g Inside
Drag to D ashboard
Trending fo r Voltage Magnitude 11/28/2011 15:53:58.698 Pacific Standard Time Trending
1 .2 5
§
1 .0 4 4
0 .8 3 8
0 .6 3 2 -
>
0 .4 2 6 -
0.22
-3 0
-2 8
-2 6
-2 4
-2 2
-2 0
-1 8
-1 6
-1 4
-1 2
-1 0
-8
-6
-4
-2
0
Time (Second]
Figure 3-13
Voltage Magnitude Trending Chart for Multiple PMUs
tni
|l_ l Trending
i
Data Type: V o lta g e A n g le
[Vi A u to Trend
Refresh Legend I D ockin g Inside
Drag to D ashboard
Trending for V oltage A ngle 1 1 /2 8/2011 15:54:01.364 P acific S ta n d a rd T im e Trending
28
22.2
1 6 .4
10.e
*2
-1
-3 0
-2 8
-2 6
-2 4
-2 2
-2 0
-1 8
-1 6
-1 4
-1 2
-1 0
-8
-6
-4
-2
0
Tim e (S econd)
Figure 3-14
Voltage Angle Trending Chart for Multiple PMUs
3-15
Description o f the WASA system
|
b
T re n d in g
"
V o lta g e A n g le
[V] A u to Trend
Refresh
Legend
D o c k in g In side
T
D rag to D ash b o ard
Trending for Voltage Angle 11/28/2011 15:53:57,364 P acific Standard Tim e Trending
16.16
11.16
J2
>
6. 16-
1.16
-19.6625
-18.6625
-17.6625
-16.6625
-15.6625
-14.6625
-13.6625
-12.6625
-11.6625
-10.6625
Time (Second)
Figure 3-15
Voltage Angle Trending Chart After Zooming in
ME!
I T r e n d in g C h a r tl n l e r f a c e NE-5W
Secondary:!
PrimarY:^k
Auto Trend Refresh Legend | Docking Inside
Diff
Trending for nterface N E -SW 11/02/2011 14:08:10.237 Pacific Daylight Tim e
0 66
0 .614-
0 .568-
0 .522-
0 .476-
0.43
-30
-28
-26
-24
-22
-20
-18
-16
-14
-12
-10
6
-
4
-
2
0
Time i.Second i
Figure 3-16
Trending Chart on the Instability Risk Index and Angle Separation Between Two Areas
3-16
Description o f the WASA system
n srn
I T re n d in g C h a rt
Primary:
| ^ j Secondary:|"
Damping Ratio (% )
3 17
Auto Trend Refresh Legend | Docking Inside
Frequency (Hz)
Trending for Interface N&-SW11102/2011 14:03:42.233 Pacific DaylightTi[re
0.37
E
38.2
0.28
25.4
0.19
14.6
0.01
-0.08
Time (Second)
Figure 3-17
Trending Chart on the Oscillation Frequency and Damping Ratio of a Selected Mode
Dashboard
The dashboard functions are provided for users to create their own dashboards for power system
monitoring. Each dashboard may integrate selected parts of the wide-area visualization (for
specific regions) and trending charts. Sample displays are provided below.
FI DashBoardWindow
.Jn x
File
■21
S PAR ADISES DD
- A n g le D iff:
3
1.14
c<
S - 8 R U TH ER F 0 5 0 0
23 - S D A V ID S O N 5 D ro p In H e re
it
d n a .it r%r nr,inert
60.14
" .J '2
6 0 .0 8 2
60.024
.
:: :
-- -- - ■-
ss.sss
55.™
55.35
-30
J
Figure 3-18
Dashboard Integrating Trending Charts and Wide-Area Visualization
3-17
Description o f the II'. IN. 1system
Early Warning of Wide-Area Angular Stability Problems
Predicting inter-area angular instability and taking mitigation actions in real time for a largescale power system pose a complex problem. Some analyses on the system’s nature in topology
and dynamics can be conducted offline to simplify the problem. Thus, an approach for online
prediction of OOS is proposed under this project to conduct tasks of two stages, as shown in
Figure 3-19. Techniques applied in each stage are introduced below.
O fflin e
A n a ly s is
■Identify generator clusters to be monitored by PMUs
•Build a mitigation strategy table for different OOS
scenarios, which separate clusters into OOS groups
•M onitordynam ics of identified clusters using PMUs
O n lin e
M o n ito rin g
& D e cisio n
S u p p o rt
•Estimate the risks of OOS scenarios by weakness in
dynamic connections among clusters
•Suggest a mitigation action from the strategy table
when the risk exceeds a threshold
Figure 3-19
Proposed Approach
Offline Analysis
The objective of this stage is to identify potential inter-area angular instability scenarios and
optimize their mitigation actions. Two steps are taken in this stage.
Step 1 is to identify a number (say K) of generator clusters by geographic control regions or
generator coherency. Those clusters may oscillate against each other under disturbances and
might become OOS. The proposed approach focuses on potential OOS (for short, POOS) among
those clusters. The average angle 5k (k=\~K) of each cluster needs to be monitored by one or a
group of PMUs.
Generator clusters defined may match those actual control regions in the grid that own local
generation and have relatively weak inter-connections. Another approach is generator coherency
analysis. EPRI’s DYNRED software offers two methods for generator coherency identification
(see [14] for more details):
•
The Tolerance Based Method, which deems generators as coherent if the phase of their right
eigenvector entries, relative to a common reference entry, is higher than a given threshold
(values 1 and 0 respectively correspond to perfect and no coherency);
•
The Weak Links Based Method, which simplifies all synchronous machines as classical
machines and forms a coherency matrix-based on the angular response to current injections
at those machines.
3-18
Description o f the II'. IN. 1 system
C o h eren c y
Threshold
a
a»080
*«Q WQ
43Q
K K
«Q
*0
X
“Q
H
i \
a =085
iQ
i« Q
n Q
'\ Y
4»Q
soQ
3y
\
erQ
II
N X
4ft\
23
a=G90
A
so Q
4Q ,9Q ”Q *£) 20
x
xjx:
K.
42
48Q ,e£) «Q
xx \
x < v \x
X V X
n
a = 094
3
o
rQ
"Q
nQ
\ tX
o
\
”Q
2v
y
’O
20
60
*o
* 6
4,0
43Q
”Q
” 0
“6
\/\
X x ft
>< \
M \ A
a = 0 96
" 0
,60
,e<b
ii
ied
X
’O
,20
kO
Figure 3-20
Generator Clustering in the Eastern Interconnection
For example, Figure 3-20 shows a clustering tree with the coherency study results on a simplified
Eastern Interconnection system model. The Tolerance Based Method is used with the coherency
threshold decreasing from 0.96 to 0.80. Each step aggregates the generators more to form bigger
but fewer clusters. The system planner may conduct such a coherency study to decide the
optimal number and sizes of generator clusters and place one or more PMUs to monitor each of
them. Later, in the case study on the WECC system, the Weak Links Based Method is used.
The above Step 1 ignores OOS within one cluster, which has either a lower probability or a
smaller impact to the system. Once a cluster itself becomes OOS, it can be mitigated by tripping
a few leading unstable generators to stabilize the others.
Step 2 is to build a table of mitigation strategies, each corresponding to a representative operating
condition and a POOS scenario describing how clusters aggregate into coherent groups under an
OOS condition. The clusters that keep synchronization aggregate to a bigger group, which is
called a POOS group in this paper. At the beginning of OOS, it is usually for one or a few
leading clusters to separate from the rest, which can respectively be regarded as two groups.
3-19
Description o f the WASA system
Thus, this work focuses on a typical type of POOS scenario that aggregates all clusters into two
POOS groups. For if clusters, they have at most 2K'l-\ different scenarios of forming two POOS
groups. For each scenario, the network interface (called POOS interface) separating two POOS
groups can be studied offline, which correspond to relatively weak connections (such as tie-lines
between control regions). Then, its mitigation strategies should be studied under representative
operating conditions. Each strategy may be:
• Reconfiguring power flows by load shedding in the load-rich side of the interface or/and
generation rejection on the other side to relieve the stress on the interface.
• Separating the grid on the POOS interface into two electric islands associated with load
shedding in the load-rich island and/or generation rejection in the generation-rich island if
coordinative separation devices are placed on the interface.
System planners may conduct cost-effect analysis and simulation studies to determine the
optimal strategy for each POOS scenario under a given operating condition. A strategy table can
be maintained on a monthly or daily basis.
Online Monitoring and Decision Support
This stage assumes that each POOS scenario is linked to an inter-area oscillation mode, and the
POOS interface can be indicated by its mode shape. Three steps in this stage are taken every
second or every few cycles on PMU data over the latest time window of T (s). T should be longer
than two periods of targeted modes but should not be too long. Otherwise, the latest changes in
modal parameters may be buried by old data in the window. Because inter-area oscillation
frequencies are mainly in 0.1 to 0.7 Hz [15], 7=20 to 60(s) is recommended for this approach.
Step 1 is to identify a number (say AT) of representative inter-area oscillation modes in PMU data
of 8\~8k over the latest time window. Apply fast Fourier transform (FFT) to each 8k over the
time window to identify the center frequencies of apparent oscillation modes indicated by peaks
of magnitudes in its frequency spectrum. For S\~Sk, a number of oscillation frequencies are
identified as candidates for the system. Represent the modes with very close center frequencies
by one mode that averages their frequencies and magnitudes. Finally, a few representative modes
with distinct center frequencies are identified. Apply the modal analysis technique in [16] to
judge whether each mode is an inter-area or local mode; utilize squared-coherency function yki2
as defined in (1) about PMU data at two locations (8k and Si). In (1), F(-) denotes FFT, <a is the
angular frequency of a representative mode, Ski(a>) is the cross-spectral density (USD), and Suf to)
and Su((d) are power spectral densities (PSD).
2
, .
K H ’
E f'iSp-F iS,)}
USD and PSD can be estimated using PMU data of 8k and Si over the latest time window. If
yk?(co) is close to 1 (such as >0.7) for some k and /, the mode is considered an inter-area mode
between Sk and Si rather than a local mode; otherwise, that mode is filtered out. Order the
3-20
Description o f the II'. IN. 1system
identified inter-area modes by their magnitudes, and select the top-M (such as 2 to 5) as
representative inter-area modes with @,-6^.
Step 2 is to estimate the modal parameters of each mode. As indicated by (2), 5k can be
approximated by summing up a steady-state angle 5m and M oscillatory signals, respectively,
about identified modes. The modal parameters to be estimated for one mode (mode z) include
oscillatory frequency con, amplitude Ak-i, damping coefficient f t, and phase (ph. The first three
parameters can be obtained from Prony analysis [17] or analytical wavelet transform (AWT)
[18], Phases can be estimated by mode-shape analysis [5], The phase angle of Ski{co) in (1) gives
the phase difference between 5k and Si.
M
+
-Ck,
I
„
.
2 cok i f
’
c o s(a > „ r +
)
(2 )
7=1
As illustrated in Figure 3-21, for mode i over a time window, draw K unit phasors with phases
<jhi~<j)Ki in a 0-360° “phase clock.” Identify the biggest two-phase differences (such as \(jhi-(jhi\ and
\(pii-(j>2 i\ in Figure 3-21) between any two adjacent phasors to partition K phasors into two groups,
whose interface in the network indicates a POOS interface linked to mode z. There may be
multiple modes linked to the same POOS interface if their mode shapes indicate the same twogroup partition in the “phase clock.”
</>1 , (reference)
<fa>
Figure 3-21
Phase Clock on One Inter-Area Mode
Step 3 is to estimate the OOS risk at each time step 1=t {s) on each POOS interface. First, the
average angles (denoted by S and <5ii) of the 5k s from two sides of the interface are respectively
calculated. The angle distance across the interface is |<5i-<5n|. Second, the modal parameters about
M inter-area modes are re-estimated for the angle distance over the latest time window of
which can help predict <5nax; that is the maximum value \S-Si\ may reach in next T. Smw can be
approximated by the sum of its steady-state value and amplitudes of M modes (see [12] for
3-21
Description o f the II'. IN. 1system
details). The security limit A m for \Si-Su\ can either take a fix value, such as 180°,or a stability
limit estimated dynamically. For example, the method proposed in [13] can estimate the stability
boundary of a monitored angle distance based on its phase-plane trajectory over a time window.
For mode i linked to this interface, the phases (denoted by (j)n and <pm) of S and Si can be
estimated, and their phase difference |<j>n-(j)iii\ is calculated. Then, the maximum phase difference
(denoted by (zWx) of all modes linked to this interface can be known. The security limit Om for
|<j>n-(j)iii\ can be a fixed value, such 180°.
max | <f>n ( t ) - (f>,n ( t )
max | 8j (t) - 5 n (t) \
Ri s k(z ) =
t=T~T+T
x
AM
M
0 ) ■x
AM
/ ,
z ,\r
(3 )
M
Finally, an OOS risk index is defined and calculated at each time step by (3), where y is used to
adjust the balance between the weights of the angle distance and phase difference in the index.
When y=0, the influence from the phase difference across the interface is ignored, and the risk is
purely based on the angle distance.
Location of Disturbance
Power system events such as generator trips, load shedding, and oscillations create perturbations
in the voltage, frequency, and angle of the grid. These disturbances propagate throughout the
electrical network in time and space. By importing and further analyzing time-synchronized
event measurement data generated by event-detection module of EPRI visualization software, the
LOD module determines event type and estimates event size and location.
Event Amount Estimation
Once a frequency event is detected, a rough power mismatch during events can be estimated. If
the average frequency deviation before and after the event is represented as Af, the frequency
change Af is related to the system active power mismatch during the event, which is generally
referred to as.5 (frequency response characteristic) [24] then:
A P = jffA /
(4)
The /? term is usually an empirical coefficient that can be obtained from statistics analysis of
known event cases. Some value of P for El and WECC can be found in [24],
Event type can be determined by the sign of Af, as shown in Figure 3-22.
3-22
Description o f the WASA system
Vi T.Tw i Tk KO
-V
LVWcr
l.VWM’10^
—utiiflv^n)
fiR**S3$5$551SS88388r?S;8RX «RKfl|f!
xxxxxRxxxxxxx$th$tstith:tSth;t&$ttt&{t&&
Rf i &RKRRK&Kf i &f i KKR k R R KRR R KR KKR R K
ftne<VTC)
Generation Trip
MVlBlKUM|rtt
888$:::;«r.M*R6<X8;5l533iX)t$888
tm m m sism m tm m m
♦ s : : ; rxkr
888888888
88888888888888888888888888888 888888888
Load Shedding
Figure 3-22
Event Type Determination
Event Location Triangulation
The event location triangulation method is based on geometrical triangulation algorithm, which
utilizes the Time Difference of Arrival (TDOA) [25] of frequency disturbance propagation and
geographical location information of each measurement device. At the first step, the time
difference of arrival will be calculated for each measurement device to form a TDOA profile of
this event. For the second step, the algorithm will pick the first several devices that see the event
and apply the plant-aided method to pinpoint the specific tripped plant or hydroelectric storage
unit based on the TDOA profile. If the plant-aided method fails, the least-squares method will be
used instead to obtain a general event location (latitude, longitude).
Time Difference of Arrival
Thorp et. al. in [26] discussed the frequency disturbance propagation in the power system as a
traveling wave. The propagation velocity of the disturbance is decided by the electromechanical
inertia of the electrical path. Therefore, the disturbance detection times at different measurement
locations differ.
3-23
Description o f the II'. IN. 1system
Figure 3-23 provides a close-up view of a generation trip event case such that the arrival time
differences at the frequency threshold marked by the black line can be easily seen. The
hypocenter of the disturbance can be back traced with the information of frequency delay of
arrival and FDR geographical information. There is a very common pattern that the closer the
units to the event location, the more proportional their TDOAs are to their distances to the
disturbance location, which is justified empirically by FNET historical event cases [27], So in the
practical event location algorithm design, only the information of the first several units will be
utilized.
FDR7-M6S
FDR11 -CALV
60.025
D isturbance
C enter
0008Hz
60.015
60.01
I
FDR18-NY
■ FDR7-MISS
I
FDR11-CALV
■ FDR13-MidWest
5 9 98 5
FDR26-Louisville
12:12:57
■ FDR28-METC
Figure 3-23
Illustration of Frequency Propagation in a Generation Trip Event
Plant-Aided Method
According to previous observations, the first FDR unit that detects the disturbance is usually the
closest one to the actual event location. Therefore, we can have the assumption that the possible
event location would be within such an area that any spot in it is sufficiently close to the first
FDR unit. We can approximately get such area by drawing a suitable circle (with radius r )
centering the first responding FDR unit. An empirical value of 200 miles is set as the radius r.
Take the generation trip disturbance as an example, at the first step; a database containing the
generation plant information is created, which includes plant capacity, plant name, plant location,
and so on. Then this plant-aided approach works in the following steps, as illustrated in Figure
3-24.
•
Plant Candidate Selection
When the disturbance is determined as a generator trip and the generation drop amount is
estimated, the possible tripped generation plants can only be located within the circular area
centering the first responding FDR, with its capacity larger than or close to the estimated trip
amount. By filtering out power plants with capacity under trip amount and those out of
3-24
Description o f the II'. IN. 1system
geographical range, the candidate pool will be narrowed down to only a few power plants and
therefore make the next validation step more efficient.
•
Plant Filtering with Proportionality Validation
After the candidate pool is chosen, the proportionality of TDOAs of devices and their physical
distances to these generation plants are checked to further select the most possible tripped plant.
The residues for linear fitting can be used as the indicator of the proportionality of the points
concerned. In this approach, each possible power plant is assumed to be the hypothesized event
location, the linear fitting is performed, and residues are obtained. The event location can then be
finalized by choosing the power plant that has the minimal residue.
1st device that sees the
disturbance
(with smallest TDOA)
■2S0
Possible plants within a certain
X
range from 1st device
100
Plant actually caused
disturbance
T 1 m (M C )
Distance: the distance o f device to plant;
Time: Wave front arrival tim e of device;
1. Plant Candidate Selection
2. Plant Filtering
Figure 3-24
Illustration of Plant-Aided Method
Least-Squares Method
As shown in Figure 3-25, the basic equations used in event location tri angulation can be drawn
as:
L , = V ( t ( - t fl)
(5)
Zf = ( xi ~ x e')2 + ( y i - y By
(6)
3-25
Description o f the II'. IN. 1system
Where L, is the distance between device, and the event hypocenter, and x t and y : are the latitude
and longitude of devicet location respectively, t- is the time of arrival determined by information
from device; and tBis actual event time. V is the wave propagation speed. For simplification, we
assume the wave propagation speed, V, is unique in different directions. By solving these
equations, the latitude x e and longitude ye of the disturbance center can be obtained.
FDTC2
L3
LA.
F-DR4
Figure 3-25
Illustration of Least-Squares Method
Implementation of Interface
The LOD algorithm is encapsulated as dynamic-link library (DLL) file to interface with EPRI
visualization software to ease the integration. Detailed information about the input and output
parameters of this DLL file is presented in Table 3-1.
3-26
Description o f the II'. IN. 1system
Table 3-1
Input and Output of LOD Module
Function
Input
Output
Parameters
Comments
Device Configuration File
Includes geographical information of devices in
XML format.
Event Frequency Data
Files
The data files include time-synchronized
frequency data of all the devices during the event
time interval in CSV format. They are generated
by event detection module of EPRI software.
Event Margin per
Interconnection
Only if the tripping amount of an event in a certain
interconnection is larger than this number, the
triangulation algorithm (the part that gets event
location) will be executed.
g_cs Location
Event location. Possible value: El, WECC, TX
g_csPlants, g_csRegion,
g_csCity,g_csState,
g_csZip,
Information of the power plant, to which the event
is close. These pieces of information are only
available for El events.
m_Year, m_Mon, m_Day;
m_Hour, m_Min, m_Sec;
Event time.
gJTripAm ount
Amount of generation trip or load shedding (in
terms of MW).
g_fLatitude, g_fLongitude
Estimated event location information. If the value
is zero, it means there is not sufficient data for
computation of event location.
bGenTrip
Bool type value. If the value is TRUE, the event
type is generation trip. If the value is FALSE, the
event type is load shedding.
3-27
4
TEST RESULTS AND DEMONSTRATION
DEPLOYMENT
Performance Testing on the Early Warning Algorithm
The proposed early warning algorithm is demonstrated using a 179-bus test system as shown in
Figure 4-1, which is a simplified WECC AC transmission system with 29 generators, 179 buses,
and 263 branches in five zones [12],
Zone I'-A
V>
Zone 1-B
6
5
-
■ rss
;
Zone 2-A
j
tu
.
/ Zone 2-B
//
"1
Zone 1-C
3
-»4
- - .
•*. o4-f
...
-- ' .
_
Jr.**.,
Figure 4-1
179-bus power system
4-1
TEST RESULTS and demonstration deployment
In the Offline Analysis stage, a study on generator coherency using DYNRED, indicates four
coherent clusters given in Table 4-1. The clusters with neighboring load buses are connected by
four groups of tie-lines:
•
Tie-lines 83-168, 83-170 and 83-172 (named TLG 1-2)
•
Tie-line 81-99 (named TLG 1-4)
•
Tie-lines 142-153 and 153-154 (named TLG 2-3)
•
Tie-line 28-29 (named TLG 3-4)
Table 4-1
Generator Clustering
Clusters
Generator Buses and PMU
Locations
Territory
1
30, 35, 65, 70, 77, 79
Zone 1-A
2
103, 112, 116, 118
Zone 2-A
3
13, 15, 40, 43, 47, 138, 140, 144,
148, 149
Zone 2-B
4
4, 6, 9, 11, 18, 36, 45, 159, 162
Zones 1-B&C
Outage of any of those lines may increase the OOS risk of the system. Obviously, the system has
six POOS scenarios: 1-234, 2-134, 3-124, 4-123, 12-34, and 14-23, which each correspond to
one POOS interface that comprises two of the above TLGs and partitions the system into two
POOS groups. For example, the POOS interface 1-234 contains TLG 1-2 and TLG 1-4 and
separates four clusters into two groups, which respectively include cluster 1 and clusters 2
through 4. A mitigation-strategy table can be built to address each POOS scenario. To avoid
OOS on a POOS interface, each strategy may be power flow reconfiguration or controlled grid
separation as mentioned above. The optimal strategy can be selected by simulating and
comparing different options.
In the Online Monitoring and Decision Support stage, the inter-area oscillations among the four
generator clusters should be monitored. Here, assume each cluster is monitored by three PMUs
located at large generation units as cycled in Figure 4-2. The rotor angles measured by three
PMUs are averaged to represent the angle of that cluster. To test the algorithms in this stage, a
sequence of contingencies is simulated by TSAT software of PowerTech Labs. The sequence
includes six successive three-phase faults at Os, 40s, 80s, 120s, 160s, and 200s near bus 83
(Malin substation) on lines 83#-172, 83#-170, 114#-124, 115#-130, 83#-94, and 83#-98 (
indicates the fault bus). The contingencies are also indicated by numbers 1 through 6 in Figure
4-1. The three tie-lines at bus 83 in the 179-bus system correspond to the famous COI
(Califomia-Oregon Intertie) of the WECC system, which was involved in the U.S. western
blackout events in 1996. Each fault is cleared after six cycles by tripping the fault line. Those
contingencies never break the connection at COI but keep increasing the system’s vulnerability
4-2
TEST RESU LTS and demonstration deployment
until the OOS indicated the simulation result in Figure 4-2 (12 angles monitored by PMUs are
highlighted).
Every 1 second, conduct modal analysis on the latest 40s (7) PMU data. Figure 4-3 gives the
FFT results over the 40s time windows ended at t=80s, t=120s, t=160s, and t=200s. The results
from four time windows are differentiated by increasing line thickness. Obvious oscillation
modes are found around 0.2 Hz, 0.5 Hz, and 0.75 Hz. The mode around 0.2 Hz is the dominant
mode. As an example, letM =l and predict OOS related to the 0.2-Hz mode.
Generator relative angle d eg ): Reference Generator = 35
20
0
9
•
11
e
15
□
30
45
■
A
A
A
V
19
T
140
149
V
80
103
116
118
-
100
120
220
Time (sec)
Figure 4-2
Angles Monitored by 12 Synchrophasors
4-3
FFT
of A n g u la r D iffe re n c e
(D e g )
TEST RESULTS and demonstration deployment
5,0,8
0,8
0,6
0,6
0.4
fr 0,4
0.2
0,2
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0,2
0.3
0.4
0.5
0,6
Frequency (Hz)
(a)
0,7
0,8
F req uency (Hz)
S\2
(b)
Si4
0.4
0.35
0.25
0.25
of A n g u l a r
D iffe re n c e
<D eg)
0.35
FFT
? 0.15
0.05
0.05
Frequency(Hz)
Frequency (Hz)
(C ) & 4
(d ) &3
Figure 4-3
FFT Results Over 40-80s, 80-120s, 120-160s, and 160-200s
------1-234
------14-23
-------------------A/v^n^-—- - - 40
Figure 4-4
Angle Distances
4-4
80
fU
120
Time (s)
----------- :
160
200 210
TEST RESULTS and demonstration deployment
180
O)
1-234
14-23
120
Time (s
160
200 210
Figure 4-5
Phase Differences
0.8
1-234
14-23
0.6
0.4
0.2 L
40
80
120
160
200 210
Time (s)
Figure 4-6
Risk Indices Without Remedial Actions
As shown in Figure 4-4, POOS interfaces 14-23 and 1-234 have the largest angle distances, and
they are both linked to the 0.2-Hz mode. Interface 14-23 actually corresponds to the NE/SE
Separation Scheme of the WECC system. To estimate OOS risks on those two interfaces, the
phase differences are calculated at every second and are shown in Figure 4-5. The phase
difference on interface 1-234 is larger than 150° for most of the time, which indicates the two
sides being oscillating against each other. In equation (3), let y=l, Am=180° and Om=180° The
risk indices on two interfaces are calculated and shown in Figure 4-6. After the contingency at
t=160s, the system has a high OOS risk around 0.8. At t=202~203s, both risk indices exceed 1.
Early warning should be sent in t=160~200s or even earlier to the grid operator to take mitigation
4-5
TEST RESULTS and demonstration deployment
actions on one of the interfaces (usually, actions on one interface can also mitigate the risk on the
other, which can be verified by simulations, as the described in the following study).
In order to test the performance of mitigation actions triggered by the proposed OOS risk index,
this load-shedding strategy is tested: shedding 5% system load in the area around cluster 3 (south
side of the two interfaces) once the OOS risk index on the interface 14-23 reaches 0.8. The
results are given in Figure 4-7 through Figure 4-9.
Figure 4-7 gives the angle distances on two interfaces with that mitigation strategy being
performed. After the control, the angle distances on both interfaces have an obvious drop.
100
------1-234
------14-23 _
O)
H. 80
k b -—
y V ---------------
<
u
o
$ £ ----------- :
—-——1
\Z\An------------
b
1
k r ~ :
^0 5 40
C
<
20
-
40
80
120
I
200
160
_
240
Time (s)
Figure 4-7
Angle Distances With Remedial Actions
From Figure 4-8, the oscillations with a high phase difference still exist on at least interface 1234, which operators should be aware of.
180
6)
CD
2,
1-234
CD
O
|
it
□
14-23
90
CD
V)
CD
.C
0.
0
40
80
120
160
200
240
Time (s)
Figure 4-8
Phase Differences With Remedial Actions
Figure 4-9 shows the effects of the control in terms of the risk indices, which drop to around 0.3.
4-6
TEST RESULTS and demonstration deployment
Time (S)
Figure 4-9
Risk Indices With Remedial Actions
For illustration purposes, simulation results at 12 PMUs in the above case study on the 179-bus
system are regarded as real-time PMU data and continuously fed into the software.
S ynchrophasor-based Situational A w are n ess Tool
2011-10-31 13:07:30.297
Real Time Monitoring
Rle
Edit
Control Panel
Tool
BMM!8HE3lJiBS
I needed on Interface NE-SW (Risk>0.8)
Help
i V d ta^e An^le^^ ^n n ^c n dnnccn ynn ocn innni1cn
I 10.0 25.0 40.055.0 70.0 85.0 100.0115.0
|
I, Angleprfference^ R a tio ^
1.00
0.20
0.40
0.60
0.80
1.00
1.20
- F Show Stability-related Information! Id [ 4^
F Show Separation Boundaries - C S f m 7 C S ^ s iE .5 m u to
F Show Group Link Data
Reference Group
t
Clean Risk M essages |
Highest Risk: | Interface NE-SW . j - ^
Separation Scenarios | Interface NE-SW
I
Freq(Hz) 10.231
Group Id 1-4
r ~ Group Id 2-3
'
Trac|dng
Risk | Angle diff CdegQ |
Ampl (deg.)
Phase diff (deg.)
0
0
Freq (Hz) | Damp(%) |
0.234
10.27
0
0
1 - North
(*
PMU
C
Power Flow
DM . T ,K | Voltage A
Angle Diff=78.1 deg
Ampl=2.90 deg
P h a se Off=132 deg
Freq=0.234 Hz
D arn p = 1 (1 6 6 % _ _
Angle Dm=20.3 deg
Amp! =3.49 deg
P hase Diff=130deg
Freq=0.234 Hz
Damp=11.17%
Show Contour
n ^
A « f« c « P M u |U 8 - T B /A T R 2 M .O
T ren d in g C h a r tln te r f a c e NE-SW
Angle Diff=57.5 deg
Ampl=0.20 deg
P h a se Diff=-3 deg
Freq=0.21S Hz
D arn p ^2 1 % ^ _ ^_
Figure 4-10
Phase Clock on an Inter-Area Mode
4-7
TEST RESULTS and demonstration deployment
Figure 4-10 shows a snapshot around t=160s. The risk index on interface 1-234 jumps to 0.8, and
a red early warning message is shown on the top of the visualization. A matched mitigation
strategy could be suggested as well in the message as decision support.
Figure 4-11 illustrates angle contour before t=200s, where the phase clock indicates that the
phase difference on the interface 1-234 has been close to 180°. The software can display the real­
time (updated every second) OOS risk on any selected POOS interface. Interface 1-234 is
visualized on the screen and colored (in red for this case), depending the color codes of risk
levels.
80$
82.9
56.7
46.l7xx
NT
\ /
/
A
3
39.6
-10.5
' ' ' ' i t
-17.6
Figure 4-11
Visualization Before the Final Contingency
Figure 4-12 shows the result of the load shedding strategy mentioned above, which corresponds
to Figure 4-7 through Figure 4-9.
4-8
TEST RESULTS and demonstration deployment
i1mmmrnfiimmwwm
Pimary Poi nt:| Rjsk
Secondary Point:| None
Trending for Interface 1 - 2+3+4 09/08/2011 14:43:49.604 Pacific Daylight Time
0.91
0.754
0.598
0.442
0 286
-30
-25
-20
-15
-10
Time [Seconds)
I Q 2011-09-08 14:43:50.604
Visualization
Angular Stability
-W Show Stability-related Inform ation------------------------------------------------------------------Show Separation Scenario • p - Show Coherent Groups ■
Reference Gro up:
Highest Risk:
1
| 7 show Me de Shape Clock
Interface 1 + 4 - 2+3
I- Auto Tracking
Separation Scenarios
H I 1 Risk
| 0.309
| Angle diff {deg.) 11
41.241
V Group 1
Amplitude | Phase diff (deg.)
Freq (Hz)
Damp(%) |
V Group 2-3-4
17.802
0
0.308
0
19.812
0
148.174
0
Figure 4-12
Visualization after load shedding
Figure 4-13 demonstrates the result of an alternative mitigation strategy—performing controlled
separation on interface 1-234 when its risk reaches 1 and immediately shedding 5% system load
around cluster 3. From the figure, angles of two resulting islands separate, but frequencies of
both islands are stabilized around 60 Flz.
4-9
TEST RESULTS and demonstration deployment
Figure 4-13
Frequencies and Angles After the Grid Is Separated on Interface 1-234
Figure 3-3 shows the visualization using 179 simulated WECC PMUs. Cascading events are
added to the Califomia-Oregon Intertie to create inter-area oscillations leading to generation out
of step in the system. Four generation groups are monitored respectively by four PMU groups.
Real-time modal information is shown for each pair of PMU groups. The phase clock clearly
tells that the group No. 1 (in the north) is oscillating against the other three, which indicates a
vulnerable interface, as visualized on the map. The trending chart in the figure is about the real­
time risk of OOS and angle difference across that interface.
Performance Testing on PMU Data Handling
Extensive performance tests were performed using the simulated PMU data of 179 PMUs in the
WECC region. It is assumed that PMUs are installed at 179 high-voltage substations of the
WECC system. The PMU data, including voltage magnitudes, phase angles, and frequencies of
30 samples per second, were created using a stability program for five successive contingencies
over 200 seconds. The simulated PMU data were read and processed as PMU data stream in the
near real-time event replay mode and captured as a new event in the database. The new event
data were inserted into and stored in the event oriented database using the binary large object
(BLOB) and data partitions (data blocks). Each data partition (block) contains 200 PMU data
objects (these parameters are configurable). Depending on the length of the event time period,
each event may consist of many data partitions in the database. For the 179 PMU test cases, there
were about 25 to 30 PMU data partitions.
4-10
TEST RESULTS and demonstration deployment
The performance benchmark testing results are shown in Table 4-2 by comparing the data sizing
of the event data and the required time for inserting the event data into the event database. In this
testing, all the database servers, application server, and web server were installed on a small Dell
laptop. The database was implemented using Microsoft SQL database. The test results showed
that the size of each partition was reduced from 5.6 MB to 1.0 MB, and the size of the entire
event data was reduced from 168 MB to 30 MB by using the new formats of the PMU data
presented in the paper. On a laptop computer, the time required for inserting the entire PMU
event data into event database was reduced from 54 seconds to 8 seconds. The time required for
reading the entire event data from the event database to read the entire PMU event data into the
event database was reduced from 4 seconds to 1.39 seconds.
Table 4-2
Benchmark Testing Result on Simulated 179 PMUs (on a Laptop Computer)
Using BLOC and Partitions Without
Special Byte-Array Data Protocol
Using BLOC and Partitions
With Special Byte-Array
Data Protocol
Data Size (MB) per Block
5.6 MB
1 MB
Data Size (MB) for Entire
Event (30 Blocks)
168 MB
30 MB
Time to Insert Entire Event
Data Into Event Database
54 Seconds
8 Seconds
Time to Read the Entire
Event Data From the Event
Database
4.02 Seconds
1.39 Seconds
For post-event analysis, it is necessary to read the PMU event data from the database and transfer
the data from the application server to each user’s computer. Using the technical approach and
the special PMU data format, the performance has been significantly improved using the new
technical approach. In this testing, all the database servers, application server, and web server
were installed on a HP Proliant 385 server. The performance tests were performed using a
Lenovo Laptop to access the applications via a secured VPN. The benchmark testing results are
shown in Table 4-3. It took 0.90 second to read and transfer the PMU data for each data block
from the database to the user computer. It took 27 seconds to read and transfer the entire PMU
event data from the database. It took about 1.5 to 2 seconds for the first visualization display to
show on the user’s computer after the event replay request. It took about 2 seconds for the first
visualization display to show up at user computer after a new event is detected in the near real­
time event replay mode.
4-11
TEST RESULTS and demonstration deployment
Table 4-3
Benchmark Testing Result Using a Laptop Connected To an Application Server via Secured VPN
Average Time (Seconds)
Read and Transfer PMU Data for Each PMU Data Block From
Database to User Computer
0.90
Read and Transfer PMU Data for Entire PMU Event Data From
Database to User Computer
27
First Visualization Display Shows up at User Computer After
Requesting Replay of an Existing Event in Event Database
1.5
First Visualization Display Shows up at User Computer After a New
Event Is Detected in the Near Real-Time Event Replay Mode
2
Offline Demonstration Using FNET Data
The University of Tennessee maintains a wide-area frequency measurement network called
FNET, based on distribution-level synchrophasors (Frequency Disturbance Recorders or FDRs),
which provide synchronized measurements at 10 samples per second.
Between April 25 and April 28, 2011, a tornado outbreak caused severe impacts on power
transmission systems in the southeast U.S. Oscillation events in the TVA system were captured
by FNET during that tornado outbreak. There are about 70 FDRs monitoring the entire Eastern
Interconnection.
Figure 4-14 and Figure 4-15 give the visualization on one oscillation event during that tornado
outbreak. The data are angle deviations (changes from the pre-disturbance values). The LOD
application accurately estimates the location and size of the event, as visualized in the two
figures below.
4-12
TEST RESU LTS and demonstration deployment
2012-10-161547:50.999 Real Tim# Monitoring
File
Control Panel
Tod
Edit
Account
_________
Help
-?e -16 -16 -16 16 16 16 76 16 116 11
o li e lie
Angle Deference Ratio
300
2012-10-16 1S37:29.96S
Display Option
□ Show Stability-related Information
A Separation Scenarios
□ Show Coherent Groups
Clean Risk Messages
:,
.
Amp! (deg.) Phase diff (deg.) Freq (Hz) Damp(%)
0.013
179.083
L01S
-2.126
A Visuakzation Option
A
OataType Voltage Angle
UtTnKnox$otir770
1976
Referee# PMU 710-INIMP4
E •»
-60 -50 -40 X -20 -10 !
Figure 4-14
Visualization of Historical FNET Data on Angle Deviations
■ Synthrophasor-based Situational Awareness Tool ]
2012-10-16 10:2908 579 Real T«n* Monrtonng
File
- «
Control Panel
«
6
Tool
Edit
Account
Help
Angle Difference Ratio
,6 .16 -16 -16 16 16 16 76 .6
•
4k
Angular Stability Control Panel 2012-10-1610:2908.04$
A Display Option
A Separation Scenarios
y Show Stability-related Information
Hghest Risk interface TVA - SOCO
Vj Show Separation Scenario
83 Auto Tracking
i- ] Show Coherent Groups
V Show Group link Data
Interface TVA-SOCO
Risk Angle diff (deg.)
0.008 -14.826
Aropl (deg) Phase diff (deg) Freq (Hz) Damp(%)
0016
.
102»08_Mode j t *
148441
1.003
-2.2SS
S _
Ref Group Id 1 <
* Visuakzation Option
*
U$TnKno«dar770
OataType Voltage Angle *
Referee# PMU 710 -INIMPA.
if
@ Show Contour
-60 -SO -40 -30 -20 -10 0
• Historical Events
Freq(Hz) 0.197
'3
't
2 • GA*FI.North
I
-60 -50 -40 - » -20 -10 0
4 - ISO-NE.N
Figure 4-15
Visualization of Historical FNET Data on Inter-Area Oscillations
4-13
TEST RESU LTS and demonstration deployment
Offline Demonstration Using Simulated and Historical TVA PMU Data
As shown in Figure 4-16, the WASA system was tested using simulated TVA PMU data (PMUs
were assumed at 60 key TVA substations) about an oscillation event. On a laptop computer, the
software can meet real-time performance requirements.
LlS.O-S.O-lLO-B.O -7.0 -5j0 -3.0 -L0 LO 3.0 5.0 7.0 9.0 11.0 13.0 15.0 17.0
Mode 1 _
Ref G
O K>
Referees PMU 31 - 8WILSON 500
[j/1 ShowC
Angle Diff=0.7 deg
Ampl=0.12 deg
Phase Diff=33 deg
Freq =0458 Hz
Damp=1.27%
Angle Uitt=-9.8 deg
Ampl =0.23 deg
Phase Diff=-94 deg
Freq=1.119Hz
Damp=4.95%
Angle Uitt=b 4 deg
Ampl=0.07 deg
Phase Diff=12 deg
Freq =0.459 Hz
Damp=-0.61 %
Angle Diff=7.1 deg
Ampl=0.13deg
Phase Diff=-21 der
Freq =0.452 Hz
Damp=0.07%
Angle Diff=16.2 deg
AmpdM3.19 deg
Phase Diff=135 deg
Freq=1.259 Hz
Damp=3.60%
Display Opt
Show Stability-related information
y l Show Separation Scenario
O Show Coherent Groups
51 Show Group Link D
□ Group 5
Angle Diff=4.8 deg
Ampl=0.00 deg
Phase Diff=-22 deg
Freq=0.000 Hz
Damp^J.00%
Ampl [deg.) Phase diff [deg.) Freq (Hz) Damp[%)
0.229
-45.849
L I 13
5.04
i iimiii ■■
Figure 4-16
Visualization of Simulated TVA PMU Data
As shown in Figure 4-17, the system was tested using historical TVA PMU data that captured an
oscillation event. The early warning algorithm correctly estimated the modal parameters of the
oscillations.
4-14
TEST RESULTS and demonstration deployment
o
An-gle Diff=7.2 deg
Ampl =0.33 deg
P h a se Diff=7 deg
Freq=1.184 Hz
D a r n p = A )3 5 % _ _
■ I Angular Stability Control Panel 2012-10-22 22:54:56.681
Figure 4-17
Visualization of Historical TVA PMU Data
Results from Field Demonstration at TVA
This WASA system has been successfully deployed at a server located at the TVA control
center, which is interfaced with openPDC at TVA to receive streaming PMU data in IEEE
C37.118 format in real time. The openPDC at TVA retrieves and processes about 150 PMUs
from TVA and other utilities in the Eastern Interconnection. With some initial training, the EPRI
technical project team worked with the technical team project at TVA to perform the
configuration and customization of the WASA software, including the common reference angle,
color maps for visualization displays, filtering of invalid PMUs, angle difference links, and
dashboards. A snapshot is shown in the figure below.
4-15
TEST RESULTS and demonstration deployment
is u tilization C ontrol P ee
A Visualization Option
D a taT y p e
V o lta g e A ngle v
R e fe re e s PMU
MONTGOMERY
0 5 h o w C ontour
Figure 4-18
Visualization of Real-Time TVA PMU Data
Results from Field Demonstration at ISO New England
This WASA system has also been deployed at a server located at ISO New England to interface
with ISO New England’s openPDC to receive real-time PMU data. The openPDC at ISO New
England retrieves and processes more than 60 PMUs from its transmission operators (TO). The
data-conditioning module played an important role for the identification of some PMUs with
data problems or invalid or missing data during the initial testing and data validation. Some of
the PMUs with data problems were marked as invalid and were excluded from the visualization
and early warning applications. Some of the missing PMU data were replaced by the latest data
available. The WASA software also provides the capability to allow users to manually exclude a
malfunctioning PMU from the calculations of the WASA applications via the visualization
displays. The quality of the visualization contour displays and the results of the early warning
4-16
TEST RESU LTS and demonstration deployment
application were dramatically improved using the above handling of the bad or missing PMU
data.
Figure 4-19 and Figure 4-20 show the voltage magnitude contour display and phase angle
contour display of ISO New England PMU data. The inter-area oscillation across an interface
(the dotted line) between the north and south regions were monitored. Real-time modal
information is shown on the interface and in the table. The two “phase clocks” indicate the mode
shapes on two detected oscillation modes at 0.2 Hz and 0.5 Hz. Trending charts on three selected
variables are shown in the figures.
■3 Synch ro p haso r-b ased Situational A w areness Tool
2012-10-13 16:37:23.266
File
Control Panel
Real Time M onitoring
Tool
Edit
A ccount
Help
V oltage M agnitude
Angle Difference Ratio
i 1 A ngular Stability Control Panel 2012-10-13 16:..
visualization C ontr
Bi 23
A Display O ption
B1 23
0 Show Stability-related Information
10 Show Separation Scenario
A Visualization O ption
O Show C oherent Groups
DataType | Voltage M agnitude
10 Show G roup Link Data
Referece PMU CX_LUDLOW19S_2
Clean Risk M essages
10 Show Contour
A Historical Events
A Separation Scenarios
Highest Risk: Interface O rrington - South
0 A uto Tracking
Interface O rrington - South
Angle DifM7.3d
Ampl=0 22 deg
Phase Difff-20 d<
Freq*0 443 Hz
Damo- 5.54%
Risk Angle diff (deg.)
0.093
17.21
* 1 Dash Board
0
Ampl (deg.) Phase diff (deg.) Freq (Hz) Damp(%)
G roup 1
0.178
Angle Diff:
-25.239
0.423
2.081
0 G roup 2
*
a
4:37:22 PM
M ode j l ▼|
s
Ref G roup Id 1
Ref G roup Id | l
1.0 3-03
Freq(Hz) 0.000
Freq(Hz) 0.423
Figure 4-19
Voltage Contour Display Using ISO New England PMU Data
4-17
TEST RESU LTS and demonstration deployment
■ Synchrophasor -based Situational Awareness Tool
2012-10-13 1 & 4003J33
File
Control Panel
Voltage Angle
-h o 3 s o -reo
-:to
Real Tan# Monitoring
Tool
-m o
Edit
Account
Help
-to
oo
mo
to
:to
Angle Difference Ratio
zoo
•
Angular Stabibty Control Panel 2012-10-13 16:n Display Option
[<fl Show StabiWty-related Information
! /i Show Separation Scenario
A Visualization Option
S ] Show Coherent Groups
DataType Voltage Angle *
V , Show Group Link Data
Referee, PMU CX.IUCXOW19S.2
Clean Rok Messages
V Show Contour
A Separation Scenanot
*- Historical Events
Highest Risk Interface O m ngton - South
6 8 Auto Tracking
/
79
Interface O m ngton • South
82
Risk Angle diff (deg.)
0.092
•
16.848
Dash Board
Ampi (deg-) Phase diff (deg.) Freq (Hz) Damp(%)
j
-27.632
0.421
3.033
0099
-93.196
0.216
-2.614
Li___ 1_____ °__ °__
1
■
4 :4002 PM <=
Mode 1 ,
ME.MAXCYS
i
a i6 6
®
2
Ref Group Id 1 -
■ 4:4002 P_
M o d e .2 ,
«=> . ®
_
HBBills
105^2
CX.DCERF1ELD
■■■■■■
L
£
so ora
8
58
FreqfHz) 0.421
rVrvn In Moca
Figure 4-20
Angle Contour Display Using ISO New England PMU Data
4-18
S
Ref Group Id T 1
5
DISSEMINATION OF RESULTS/TECHNOLOGY
TRANSFER
The project team actively sought throughout the project ways to disseminate and effectively
share approaches and results publicly. Dissemination of results and technology transfer were
carried out in various forms, including publication in conferences, seminars, webinars, peerreviewed journal articles, patents, and dedicated workshop. This section provides a description of
such activity.
Conference and Journal Papers
1. G. Zhang, K. Sun, et al, “Application of Synchrophasor Measurements for Improving
Operator Situational Awareness,” IEEE PES General Meeting, Detroit, MI, 24-29 July 2011.
2. K. Sun, K. Hur, P. Zhang, “A New Unified Scheme for Controlled Power System Separation
Using Synchronized Phasor Measurements,” IEEE Trans Power Systems, Aug 2011.
3. K. Sun, S. Lee, P. Zhang, “An Adaptive Power System Equivalent for Real-time Estimation
of Stability Margin using Phase-Plane Trajectories,” IEEE Trans. Power Systems, vol. 26, pp.
915-923, May 2011.
4. K. Sun, X. Luo, J. Wong, “Early Warning of Wide-Area Angular Stability Problems Using
Synchrophasors,” IEEE PES General Meeting, 23-26 July 2012, San Diego, 2012.
Patent Applications
1. K. Sun, K. Hur, P. Zhang, Application of Phasor Measurement Units (PMU) for Controlled
System Separation, U.S. Patent Application No.12/948,188 (Pub. No. US 2012/0123602
A l), Nov 17, 2010.
2. K. Sun, Q. Zhou, et al, “Application of Phase-Locked Loop (PEL) in Oscillation
Monitoring for Interconnected Power Systems,” Pending U.S. patent application filed in
2011 .
Participation in NASPI
Project team members, especially former project managers Gourui Zhang and Kai Sun, have
actively participated in the North American SynchroPhasor Initiative (NASPI), giving
presentations describing the objective, value, approach, and results of this project.
NASPI is a collaborative effort between the U.S. Department of Energy, the North American
Electric Reliability Corporation, and North American electric utilities, vendors, consultants,
5-1
Dissemination o f results/technology transfer
federal and private researchers, and academics. NASPI activities are funded by DOE and NERC,
as well as by the voluntary efforts of many industry members and experts.
The mission of the NASPI is to improve power system reliability and visibility through widearea measurement and control technologies. The NASPI community works to advance the
deployment and use of networked phasor measurement devices, phasor data-sharing, applications
development and use, and research and analysis.
Hence, NASPI is the most appropriate forum to expose the value and outcome of this project to
the electric industry community. The presentations given by the project members at the different
NASPI workshops1 can be found in the dedicated website https://www.naspi.org/.
Workshop at TVA facility in Chattanooga, TN
The EPRI project team organized a dedicated software demonstration workshop, which was held
at the TVA office in Chattanooga, Tennessee, on October 23rd 2012. Near twenty professionals
from different electric utilities, ISOs, and research and government institutions participated in the
event, including DOE, EPRI, TVA, UTK, University of Tennessee Chattanooga, Quanta Tech,
Oakridge National Laboratory, GPA, ISO New England, and New York Power Authority.
At the workshop, TVA presented its vision on situational awareness, and DOE presented an
overview of SGIG projects and synchrophasor projects. EPRI, along with UTK and Quanta
Technology, presented a detailed description of this project, including: project organization and
motivation, approach, project development, mathematical foundations, software development
challenges and achievements, and project results and outcomes. The project team also conducted
the following demonstration of WASA operation and performance:
• Offline demonstration with simulated WECC 179 PMUs.
• Offline demonstration with simulated TVA 60 PMUs.
• Offline demonstration with historical ISO-NE PMU data (about 70 PMUs).
• Offline demonstration with historical TVA PMUs on TVACumberland oscillation event
in 2006.
•
Offline demonstration with historical FNET data (about 70 FDRs) on TVA tornado event
in 2011.
•
Online demonstration with real-time PMU data stream at TVA.
Participants of the workshop provided important preliminary feedback regarding the value and
performance of the software, as well as ideas and suggestions for further development and future
research.
Overall, the participants agreed that a more comprehensive testing is required to fully evaluate
the capability, features, and value of the software, as well as to identify “bugs” and
improvements. In line with that conclusion, one of the main action items derived from the
workshop was to organize a pilot demonstration at the TVA’s Reliability Coordinator.
1NASPI meeting in Orlando, FL Oct. 2012, and NASPI meeting in Atlanta, GA Oct. 18, 2012.
5-2
Dissemination o f results/technology transfer
Further, one of the main conclusions regarding the usability and value of a wide-area
synchrophasor system is that it extends the capability to not only monitor and assess system
security but also to provide automatic control actions to mitigate fast instability conditions and
widespread outages.
5-3
6
CONCLUSIONS AND RECOMMENDATIONS FOR
FURTHER INVESTIGATION
The advanced synchrophasor-based wide-area situational awareness software system has been
developed to provide wide-area power system visualization, near-real-time event tracking and
early warning of potential stability problems for power system operators and reliability
coordinators using real-time or historical PMU measurements. The applications have been
successfully deployed and demonstrated at TVA and ISO New England using real-time
synchrophasor measurements from the openPDC Phasor Data Concentrators. The WASA
configuration manager and the extensive data-conditioning features have greatly simplified the
software deployment, integration with the Phasor Data Concentrators, and the software
maintenance and upgrade.
The WASA applications developed in this DOE R&D and demonstration project have also been
extensively tested using the synchrophasor data of simulated 179 PMUs in the WECC system
and the historical FNET data of large system events. The applications have implemented several
advanced technologies, including the object-oriented memory residence database, reduced PMU
data sizing, and the enhanced event oriented database using BLOB and data partitioning to
efficiently handle large amounts of real-time and historical synchrophasor measurements and to
support large numbers of concurrent users. The performance testing results show that it is
possible to perform the near-real-time event tracking a few seconds after a new event is detected
to track the ongoing event with high resolution (up to 30 samples per second) to improve the
operator wide area situation awareness. The integrated post-event analysis of this system allows
users to perform high fidelity post-event analysis using historical PMU measurements in the
event-oriented historical database. The early warning application conducts measurement-based
oscillation mode analysis and angle stability analysis in real time on PMU data over a specified
length of time window. The information provided includes the frequency, damping and mode
shape of the dominant oscillation mode, clusters of PMUs indicating generation coherency, and
vulnerable grid interfaces with instability risks, which indicate potential system separation. The
test results also show that the early warning application has good potentials in providing useful
information to system operators about potential wide-area angle stability problems.
The WASA software is ready for deployment, testing, and demonstration in the control centers at
RTOs, ISOs, and utilities for improving the system operator situation awareness. For real-time
situational awareness, the key functional modules integrated by this software can provide system
operators with comprehensive information, such as visualization of wide-area PMU
measurements with potential reliability concerns to be highlighted by effective use of colors and
event information including times, locations, types, sizes, recorded high-resolution event data,
system oscillation information, and early warning information based on stability analysis results.
The software has demonstrated the state-of-the-art technologies on PMU-based wide-area
6-1
conclusions and recommendations fo r further investigation
situational awareness. The next step will be to investigate a wide-area coordinating control
system integrating the techniques applied by the WASA software. That system would be based
on highly reliable and secure communication infrastructures and high-performance computing
resources for real-time optimization and validation of control actions.
6-2
7
REFERENCES
[1], U.S.-Canada Power System Outage Task Force, “Final Report on the August 14, 2003
Blackout in the United States and Canada: Causes and Recommendations”, April, 2004.
[2], “NASPI - North American SynchroPhasor Initiative”, http://www.naspi.org.
[3], Tao Xia, Hengxu Zhang, Robert Gardner, Jason Bank, Jingyuan Dong, Jian Zuo, Yilu Liu,
Lisa Beard, Peter Hirsch, Guorui Zhang, and Rick Dong, "Wide Area Frequency based
Event Location Estimation”, Presented at 2007 IEEE PES General Meeting.
[4], B. Qiu, L. Chen, V.A. Centeno, X. Dong, Y. Liu, “Internet Based Frequency Monitoring
Network (FNET)”, IEEE Power Engineering Society Winter Meeting, 28 Jan.-l Feb. 2001,
Vol. 3, pp 1166-1171.
[5], C. Maryinez, M. Parashar, J. Dyer, J. Coroas, “Phasor Data Requirements for Real Time
Wide-Area Monitoring, Control and Protection Applications”, EIPP White Paper, January
26, 2005.
[6], Manu Parashar, Jim Dyer and Terry Bike, “EIPP Real-Time Dynamics Monitoring
System”, http://certs.lbl.gov/certs-rt-pubs.htmk February 2006
[7], G. Zhang, P. Hirsch and S. Lee, “Wide Area Power System Visualization Using Smart
Client Technology”, Presented at 2007 IEEE PES General Meeting, Tampa, Florida, 2007.
[8], Manu Parashar, Jianzhong Mo, "Real Time Dynamics Monitoring System (RTDMS):
Phasor Applications for the Control Room," 42nd Hawaii International Conference on
System Sciences, 2009
[9], G. Zhang, L. Beard, R. Carroll, R. Zuo and Y. Liu, “WAVA-PMU and Near Real-Time
Event Replay Using SynchroPhasor Measurements”, Presentation at NASPI meeting at
Chattanooga, TN, October, 2009.
[10], Ritchie Carroll, “openPDC Specifications”, NASPI Working Group Meeting, Austin, TX,
February 25, 2010
[11], Zhang, S. Lee, R. Carroll, J. Zuo, L. Beard and Y. Liu, “Wide Area Power System
Visualization using Real-Time SynchroPhasor Measurements”, Presented at 2010 PES
General Meeting at Minneapolis, Minnesota, July, 2010.
1
Error! No text o f specified style in document.
[12],
K. Sun, K. Hur, P. Zhang, A New Unified Scheme for Controlled Power System
Separation Using Synchronized Phasor Measurements, submitted to IEEE Transactions
on Power Systems.
[13],
K. Sun, Application of Phasor Measurement Units for Controlled System Separation,
EPRI Report #1017800, 2009
[14],
EPRI DYNRED Software Manual, EPRI, Palo Alto, CA: 2010. Software Product ID #:
1020268
[15],
P.Kundur, Power System Stability and Control. New York: McGraw-Hill, 1994.
[16],
D. I. Trudnowski, “Estimating Electromechanical Mode Shape From Synchrophasor
Measurements,” IEEE Trans. Power Syst., vol. 23, no. 3, pp. 1188-1195, August 2008
[17],
I. F. Hauer, “Application of prony analysis to the determination of modal content and
equivalent models for measured power system response,” IEEE Trans. Power Syst., vol.
6, no. 3, pp. 1062-1068, Aug. 1991.
[18],
K. Hur and S. Santoso, “Estimation of system damping parameters using analytic wavelet
transforms,” IEEE Trans. Power Del., vol. 24, no. 3, pp. 1302-1309, Jul. 2009.
[19],
G. Zhang, K. Sun, H. Chen, R. Carroll and Y. Liu, “Applications of Synchrophasor
Measurements for Improving Operator Situational Awareness”, IEEE PES General
Meeting, Detroit, Michigan, 2011.
[20],
K. Sun, X. Luo, J. Wong, “Early Warning of Wide-Area Angular Stability Problems
Using Synchrophasors”, IEEE PES General Meeting, 23-26 July 2012, San Diego, 2012
[21],
FNET, http://fnetpublic.utk.edu/index.html
[22],
T. Xia, H. Zhang, et al, “Wide-area Frequency Based Event Location Estimation”, 2007
IEEE PES General Meeting
[23],
R.M. Gardner, et al, "Non-Parametric Power System Event Location Using Wide-Area
Measurements," 2006 IEEE PES PSCE
[24],
J.W. Ingleson, E. Allen, "Tracking the Eastern Interconnection frequency governing
characteristic," in IEEE Power and Energy Society General Meeting, 2010, pp. 1-6, 25-29
July 2010
[25],
Parashar, M., J.S. Thorp, and C.E. Seyler, Continuum modeling of electromechanical
dynamics in large-scale power systems. [Circuits and Systems I: Fundamental Theory
and Applications, IEEE Transactions on], 2004. 51(9): p.1848.
2
Error! No text o f specified style in document.
[26], J.S. Thorp, C.E. Seyler, A.G. Phadke, “Electromechanical wave propagation in large
electric power systems,” IEEE Transactions on Circuits and Systems I: Fundamental
Theory and Applications, vol. 45, issue 6, pp. 614 - 622, June, 1998.
[27], T. Xia, Frequency Monitoring Network (FNET) Algorithm Improvements and Application
Development, 2009, Virginia Tech: Blacksburg
[28], EPRI NYSERDA R&D Project Final Report,” Real-Time Applications of Phasor
Measurement Units (PMU) for Visualization, Reactive Power Monitoring and Voltage
Stability Protection”, Final Report 10-33, November, 2010
3