Upload
others
View
9
Download
1
Embed Size (px)
Citation preview
S.C. Srivastava, ProfessorDepartmentof Electrical Engineering
Indian Institute of Technology KanpurEmail: [email protected]
SC Srivastava / IITK WAMCS QIP course 10 May 2019 1
General Introduction The size and complexity of the system has increased.
A.C. transmission systems at 1100/1200 kV.
Generating unit sizes have gone upto 800/1000 MW.
HVDC links (+/- 500/800 kV) are in operation for back to back and long distance power transfer.
Concerns to its operators: Maintaining security, reliability, quality and stability of power system
Ensuring economic operation. ( Monopoly-Market based operation)
Leading to use of large number of sensors, DSP, PE devices, communication networks and IT enabled services ( Large CPS)
Computer aided monitoring & dispatching SCADA (Supervisory Control and Data Acquisition) based Energy
Management System (EMS)and Distribution Automation System (DMS).
Wide Area Monitoring and Control Systems (WAMCS) using synchrophasors.
2SC Srivastava / IITK WAMCS QIP course 10 May 2019
Power System Stability
Frequency
Stability
Small-Signal
Stability
Transient
Stability
Short
Term
Long
Term
Large-
Disturbance
Voltage Stability
Small-
Disturbance
Voltage Stability
Voltage
Stability
Rotor Angle
Stability
Consideration
for
Classification
Physical
Nature/ Main
System
Parameter
Size of
Disturbance
Time
Span
Short Term
Short
Term
Long
Term
P. Kundur, J. Paserba, V. Ajjarapu, G. Andersson, A. Bose, C. Canizares, N. Hatziargyriou, D. Hill, A. Stankovic, C. Taylor, T. V. Cutsem, and V. Vittal, "Definition and classification of power system stability IEEE/CIGRE joint task force on stability terms and definitions," IEEE Transactions on Power Systems, vol. 19, no. 3, pp. 1387-1401, Aug. 2004.
3
RTU RTU RTU
SUB LDC SUB LDC SUB LDC
SLDC SLDC SLDC
ERLDC WRLDC NRLDC SRLDC NERLDC
NLDC
SCADA EMS System+
(Typical Architecture in India)
+ Wide area monitoring system since 1960s
SC Srivastava / IITK WAMCS QIP course 10 May 2019 4
Components of SCADA EMS
• Remote Terminal Units – Intelligent units to collect data from field,convert these to suitable form ( through A/D or D/A converters,Transducers etc.), communicate the data to Control Center.
• Communication Networks or Data Transmission Networks –PLCC, Microwave Telemetry Link, Dialup Network, Fiber Optics etc.
• Computer Systems & Man Machine Interface- at Control Center
• Software:
1. Operating system software for Data base management & MMI
2. Communication Software
3. Application Software (to perform advance functions of EMS)
Problems with SCADA based WAMS:• Data time skewed. Data scan rate upto 10 sec.
• Only magnitude measurements and phasors through stateestimation-time extensive.
5SC Srivastava / IITK WAMCS QIP course 10 May 2019
Advance Functions of Energy Management System
• Supervisory Control & Data Acquisition (SCADA) functions
• System Monitoring and Alarm Functions
• State Estimation
• On line Load Flow
• Economic Load Dispatch
• Optimal Power Flow ( including Optimal Reactive Power Dispatch)
• Security Monitoring and Control
• Automatic Generation Control
• Unit Commitment
• Load Forecasting
• Log Report Generation ( Periodic & Event logs), etc.
A program scheduler may invoke various Application programs at fixed intervals. 6
PHASOR DATA CONCENTRATOR
APPLICATIONSOFTWARE
SYSTEM CONTROL CENTER
MONITORINGCONTROLDATABASE
PMU
GPS
PMU
PMU PMU
PDC
PDC SUPER PDC
A Typical Synchrophasor based WAMCS Architecture
7
Recommended PMU reporting rates (frames/sec): as per IEEE std. 37.118
System frequency 50 Hz : 10 25 50System frequency 60 Hz : 10 12 15 20 30 60
(New standard also encourages fps of 100, 120 or less than 10, if required for a specific application)
Event Millions of
people
affected
Location Date
July 2012 India blackout 620 India 30–31 July 2012
2014 Bangladesh blackout 150 Bangladesh 1 Nov 2014
2005 Java-Bali blackout 100 Indonesia 18 Aug 2005
1999 Southern Brazil blackout 97 Brazil 11 March 1999
2009 Brazil and Paraguay
blackout
87 Brazil, Paraguay 10–11 Nov 2009
Northeast blackout of 2003 55 United States,
Canada
14–15 Aug 2003
2003 Italy blackout 55 Italy, Switzerland,
Austria, Slovenia,
Croatia
28 Sep 2003
Northeast blackout of 1965 30 United States,
Canada
9 Nov 1965
Few Major Black-out Events across the World(Ref. http://en.wikipedia.org/wiki/List_of_power_outages)
SC Srivastava / IITK WAMCS QIP course 10 May 2019 8
SC Srivastava / IITK WAMCS QIP course 10 May 2019 9
RTO/ISO in USA
August 14, 2003 North-East Blackout in US, Canada
10
Reliability Coordinators in Midwest
SC Srivastava / IITK WAMCS QIP course 10 May 2019
2003 US Blackout: Few Eventshttps://en.wikipedia.org/wiki/Northeast_blackout_of_2003
12:15 p.m. Incorrect telemetry data renders inoperative the state estimator, operated by the Indiana-based MISO. An operator corrects the telemetry problem but forgets to restart the monitoring tool.
2:02 p.m. The first of several 345 kV overhead transmission lines in northeast Ohio fails due to contact with a tree in Walton Hills, Ohio
2:14 p.m. An alarm system fails at FirstEnergy's control room, not repaired.
3:05 p.m. A 345 kV transmission line known as the Chamberlin-Harding line sags into a tree and trips in Parma, south of Cleveland.
3:32 p.m. Power shifted onto another 345 kV line, the Hanna-Juniper interconnection, causes it to sag into a tree, bringing it offline as well. While MISO and FirstEnergy controllers concentrate on understanding the failures, they fail to inform system controllers in nearby states.
3:39 p.m. A FirstEnergy 138 kV line trips in northern Ohio.
4:05:57 p.m. The Sammis-Star 345 kV line trips due to undervoltage and overcurrent interpreted as a short circuit. Later analysis suggests that the blackout could have been averted prior to this failure by cutting 1.5 GW of load in the Cleveland–Akron area.
SC Srivastava / IITK WAMCS QIP course 10 May 2019 11
4:09:02 p.m. Voltage sags deeply as Ohio draws 2 GW of power from Michigan.
4:10:34 p.m. Many transmission lines trip out, first in Michigan and then in Ohio, blocking the eastward flow of power around the south shore through Erie, Pennsylvania and into the Buffalo, New York area.
4:10:38 p.m. Cleveland separates from the Pennsylvania grid after a series of line and generator trip.
4:10:40 p.m. Flow flips to 2 GW eastward from Michigan through Ontario (a net reversal of 5.7 GW of power), then reverses back westward again within a half second.
4:10:43 p.m. International connections between the United States and Canada start to fail.
4:10:45 p.m. Northwestern Ontario separates from the east The first Ontario power plants go offline
4:10:46 p.m. New York separates from the New England grid.
4:10:50 p.m. Ontario separates from the western New York grid, cascaded separation and tripping happens.
4:13 p.m. End of cascading failure. 256 power plants are off-line, 85% of which went offline after the grid separations occurred, most due to the action of automatic protective controls. 12
13
Blackout Root Cause : FE Situational Awareness & Vegetation Management
FE did not ensure a reliable system after contingencies occurred because it did not have an effective contingency analysis capability
FE did not have effective procedures to ensure operators were aware of the status of critical monitoring tools
FE did not have effective procedures to test monitoring tools after repairs
FE did not have additional high level monitoring tools after alarm system failed
FE did not adequately manage tree growth in its transmission rights of way
SC Srivastava / IITK WAMCS QIP course 10 May 2019
14
Blackout Cause: Reliability Coordinator Diagnostics
MISO’s state estimator failed due to a data error.
MISO’s flowgate monitoring tool didn’t have real-time line information to detect growing overloads
MISO operators couldn’t easily link breaker status to line status to understand changing conditions.
PJM and MISO ineffective procedures and wide grid visibility to coordinate problems affecting their common boundaries
SC Srivastava / IITK WAMCS QIP course 10 May 2019
Grid Disturbances in Indiaon 30th & 31st July 2012
Enquiry Committee report posted at
http://www.cercind.gov.in/2012/orders/
Final_Report_Grid_Disturbance.pdf
Loss of load: 36000 and 48000 MW
SC Srivastava / IITK WAMCS QIP course 10 May 2019 15
NRNER
ER
SR
WR
Schd : 2220 MW
Actual : 2044 MW
Schd : 832 MW
Actual : 975 MW
Schd : 1192 MW
Actual : 3123 MW
Schd : 835 MW
Actual : 962 MW
Schd : -33 MW
Actual : 95 MW
Schd: 278 MW
Actual : 2654 MW
Antecedent Conditions on 30th July 2012
Time 02:30 Hrs
592 MW
Mh’garh
Mundra
FrequencyNEW Grid- 49.68 Hz
Over-drawal : 1755 MW
Under-drawal: 2343 MW
Over-drawal : 446 MW
SCADA
Snapshot
16
Loss of load: 36000 MW and 48000 MW
SC Srivastava / IITK WAMCS QIP course 10 May 2019
NR NER
ER
Inter-regional lines under outage on 30th July 2012Z
erd
aB
hin
ma
l
Ka
nkro
li
Gw
alio
r
Au
raiy
aM
ah
alg
ao
nM
ala
np
ur
Sipat Ranch
iRourkela
Sterlite
Raigarh
RaigarhBudhipadar
Korba
Sala
kat
iB
irpara
Bin
aguri
Bongoig
aonGorakhpur
BaliaFa
teh
pu
r
Gaya
Agra
Ko
taB
adod
Mo
rak
To SRHVDC
BhadrawatiHVDC
Talcher Kolar
HVDC
Gazuwaka
HV
DC
Vin
dh
.
WR
17SC Srivastava / IITK WAMCS QIP course 10 May 2019
Center of angular Separation on 30th July 2012
18SC Srivastava / IITK WAMCS QIP course 10 May 2019
PMU plots on 30th July 2012. The trippings at ~2:33:15 due to large angular separation – correlates with DRs and WAFMS
19
Balia-Biharsharif
Balia-Patna
Grkpr-Muzzpr.
SC Srivastava / IITK WAMCS QIP course 10 May 2019
NRNER
ER
SR
WR
Schd : 740 MW
Actual : 1978 MW
Schd : 830 MW
Actual : 830 MW
Schd : 850 MW
Actual : 2044 MW
Schd : 894 MW
Actual : 900 MW
Schedule: -21 MW
Actual : 150 MW
Schd : 73 MW
Actual : 1822 MW
Antecedent Conditions on 31st July 2012
Frequency NEW Grid-49.84 Hz
Time 12:57 Hrs
432 MW
Mh’garh
Mundra
Over-drawal : 2432 MW
Under-drawal: 2987 MW
Over-drawal : 720 MW
SCADA
Snapshot
20SC Srivastava / IITK WAMCS QIP course 10 May 2019
NR NER
ER
Inter-regional lines under outage on 31st July 2012
Ze
rda
Bh
inm
al
Ka
nkro
li
Gw
alio
r
Au
raiy
aM
ah
alg
ao
n
Ma
lan
pu
r
Sipat Ranch
iRourkela
Sterlite
Raigarh
RaigarhBudhipadar
Korba
Sala
kat
iB
irpara
Bin
aguri
Bongoig
aonGorakhpur
BaliaFa
teh
pu
r
Gaya
Agra
Ko
taB
adod
Mo
rak
To SRHVDC
Bhadrawati
HVDC
Talcher Kolar
HVDC
Gazuwaka
HV
DC
Vin
dh
.
WR
21SC Srivastava / IITK WAMCS QIP course 10 May 2019
Separation on 31st July 2012
22SC Srivastava / IITK WAMCS QIP course 10 May 2019
PMU Plots
23
Jabalpur, Bhadrawati, Raipur
Gwal-Bina
SC Srivastava / IITK WAMCS QIP course 10 May 2019
Important Observations
Few Major Causes of the Grid Disturbances:
• Lack of Situational Awareness and real time monitoring tools
• Early security assessment/warning system.
• Unintended operation of Protection/ Improper coordination of Control Actions
• Lack of enough reactive compensation
• Human error & Grid Indiscipline
Few Remedial Measures:
• Wide area monitoring & control. PGCIL-URTDMS Project (http://www.cea.nic.in/reports/powersystems/sppa/scm/allindia/agenda_note/1st.pdf )
• Ensuring activation of all emergency controls and protection
• Smart grid - Proper automation, information flow and data management.
• Regulatory changes to strengthen system security and operation.
24SC Srivastava / IITK WAMCS QIP course 10 May 2019
➢Pilot projects in all regions already in place.
➢ Integrate the pilot project PMUs with the National plan.
➢Unified Real Time Dynamic Measurement System(URTDMS) being implemented.
➢About 1700 PMUs are planned, to be placed at all the400 kV and above voltage buses and major generatingplant buses. Analytics being developed by IIT Bombay.
➢Few analytics to be implemented in the first phase
include; Line parameter estimation, Oscillation
monitoring, Distance relays’ vulnerability analysis,
Linear/dynamic state estimation, CT/CVT calibration,
Supervision of zone-3 distance protection.
25SC Srivastava / IITK WAMCS QIP course 10 May 2019
Roadmap of WAMCPS Deployment in India
Some Research Work on Synchrophasor based WAMCPS carried out at IIT Kanpur
• Phasor Assisted State Estimation
• Dynamic Phasor Estimation
• Machine Rotor Angle Estimation for Transient Stability
• Critical Mode Identification for Small Signal Stability
• Synchrophasor based Voltage Stability Assessment
• Wide Area Measurement based Adaptive Distance Protection
• Network based Wide Area Damping Control
• Optimal Frequency and Voltage stability based load shedding
• Load, Transmission Line and Generator Model Parameter
Estimation.
• On Line Tuning of Power system Stabilizers
• Testing in WAMS Lab using RTDS
26SC Srivastava / IITK WAMCS QIP course 10 May 2019
*P. Tripathy, S. C. Srivastava, and S. N. Singh, “A Modified TLS-ESPRIT-Based Method for Low-Frequency Mode Identification in Power Systems Utilizing
Synchrophasor Measurements", IEEE Transactions on Power Systems, vol. 26, no. 2, May 2011, pp. 719-727.
Modified TLS-ESPRIT+ based Low Frequency Mode Identification*
27SC Srivastava / IITK WAMCS QIP course 10 May 2019
+ Total Least Squares Estimation of
Signal Parameters via Rotational
Invariance Technique (TLS-ESPIRIT)
Test signals corresponding to inter-Area modes
The test signal, considered for simulations had a small signal
oscillation* of 0.4 Hz, having attenuation factor -0.07, amplitude
set to unity with an initial phase of 60 degree.
Two area test system assuming the availability of power signals
from PMUs (Ref: Kundur’s book)
Probing test data of the Western Electricity Coordination Council
(WECC) system (obtained from BPA site)
PMU data obtained on Northern Grid system (source: POSOCO)
Test Cases
28SC Srivastava / IITK WAMCS QIP course 10 May 2019
Test Signal corresponding to Inter- Area Oscillations with SNR 30 dB
29
Test Signal corresponding to Inter- Area Oscillations with SNR 30 dB
30SC Srivastava / IITK WAMCS QIP course 10 May 2019
Improved Voltage Instability Monitoring Index (IVIMI) :
VDR – Voltage Deviation from its Reference
CVD – Consecutive Voltage Deviation (Rate of
Voltage Deviation)
• is system wide voltage instability index at instant-i.
• Based on voltage deviation and rate of voltage change.
• Rate of change of voltage reflects the dynamic variation.
max
2
max
1CVD
)(VDR
)(IVIMI ii
i
CVDiw
VDRiw +=
setbusesloadtheallpIVIMIISVIMI p
ii= )max(
Improved Voltage Instability Monitoring Index*
* Sodhi, R., Srivastava, S.C., Singh, S.N., "A Simple Scheme for Wide Area Detection of Impending Voltage Instability," IEEE
Trans. on Smart Grid, vol.3, no.2, pp.818-827, June 2012.
*Ch. V. V. S. Bhaskara Reddy, S. C. Srivastava and Saikat Chakrabarti, “An improved Static Voltage Stability Index using
Synchrophasor Measurements for Early Detection of Impending Voltage Instability,” National Power Systems
Conference (NPSC), IIT (BHU), Varanasi, India, December 12‐14, 2012.
iISVIMI
Voltage Stability Risk Index (VSRI)#
1. Calculate the Moving Average
2. Calculate the % diversity of the ith measurement
3. Calculate the Risk Index
where, Vi= voltage measurement at ith interval
Vj=moving average of jth window,
N=length of the moving window,
dj=diversity of th ith measurement for the jth window
Zj = VSRI of the jth window.
=
=N
iij
vN
v1
1
100−
=j
ji
iv
vvd
+=
=
−
2
)(11
)1(
N
iii
j
tdd
Nz
# Kim, et al., “System an method for calculating real-time voltage stability risk index in power system using
time series data,” Patent No. US007236898B2, Dt. Jun. 26,2007.
SC Srivastava / IITK WAMCS QIP course 10 May 2019
0 50 100 150 200 250 300
0.7
0.8
0.9
1
1.1
1.2
TIME in sec.
VO
LT
AG
E in P
U
156
158
174
Fig: Critical bus voltages for load increase.
i. Slow increase of load
Simulation Results
NRPG 246 bus Indian test system
SC Srivastava / IITK WAMCS QIP course 10 May 2019
0 50 100 150 200 250 300-1
-0.5
0
0.5
VS
RI
Time in Sec.
156
158
174
Fig.: VSRI plot for
load increase.
0 50 100 150 200 250 300-3
-2
-1
0
1
2
3
4
TIME in sec.
VIM
I
156
158
174
Fig: IVIMI plot for
load increase.
(195 s early prediction)
SC Srivastava / IITK WAMCS QIP course 10 May 2019
0 50 100 150 200 250 300
0.65
0.7
0.75
0.8
0.85
0.9
0.95
1
TIME in sec.
VO
LT
AG
E in P
U
156
158
174
ii. A line outage and simultaneous load increase
Fig: Critical bus voltages
for line outage
SC Srivastava / IITK WAMCS QIP course 10 May 2019
0 50 100 150 200 250 3000
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
TIME in sec.
VIM
I
156
158
174
Fig: IVIMI plot for
load increase.
(170 s early prediction)
➢ Power system events that cause the impedance
trajectories to enter into tripping zones, under no-
fault situations, cause system security concerns.
Resistance
Re
acta
nc
e
36
Distance Relay Protection
SC Srivastava / IITK WAMCS QIP course 10 May 2019
Unintended Operation of Distance Relays➢ Due to impedance trajectory encroachment into
relay tripping zone.
➢ System events which, sometimes, lead to distance
relay operation
➢ Dynamic
• Power Swings - Blinders and Timer blocking
scheme
• Voltage Instability
➢ Static
• Load encroachment
➢ Operation under Zone-3 has been observed to result in
cascading failures of power systems.37SC Srivastava / IITK WAMCS QIP course 10 May 2019
A Method* using SVM to avoid Unintended
Operation of Distance Relays
➢ Relay Ranking Index
▪Ratio of normalized apparent impedance of the relay to
branch loss sensitivity used to rank the relays in terms of its
vulnerability to Power Swings and Voltage Instability.
➢Fault and Disturbance Classifiers
▪Two support vector machines used, SVM-1 for
distinguishing between fault and no fault condition, and
SVM-2 for classifying the no fault disturbance in ‘Voltage
Instability’ or ‘Power Swing.
38
*Seethalekshmi K., S.N. Singh and S.C. Srivastava, “A Classification Approach Using Support Vector
Machines to Prevent Distance Relay Mal-operation under Power Swing and Voltage Instability”, IEEE
Transactions on Power Delivery, Vol. 27, No. 3, July 2012, pp. 1124-1133.
SC Srivastava / IITK WAMCS QIP course 10 May 2019
Critical Relays’ Identification➢ Normalized Apparent Impedance for a relay, Rij in line i-j
where, and are the sending and receiving end voltages of the line i-j and is the impedance corresponding to third zone setting of the relay. is the impedance of the line.
➢ Branch loss sensitivity:
where,
➢ Relay Ranking Index
( )3
ij
iliner
i j
r ij
VZ Z
V V
Z
= −
3r ijZiV jV
lineZ
39
= + −loss loss
h P jQ in line i j
2 22 2
j ji iVV
ij ij ij ij ijS S S S S
= + + +
ij
ij
ij
ZrRRI
S=
,
=
iV
ij
i
hS
V,
=
i
ij
i
hS and
SC Srivastava / IITK WAMCS QIP course 10 May 2019
Proposed Classification StrategyRelay impedance measurement unit
1= Trip
0= Block
1= Relay pick-up
0= ResetV
1= Enable SVM-2
0= Disable SVM-2
1= Power swing
-1= Voltage instability
Digital distance relay block
for sampling & DFT based
phasor calculation
Feature vectorSVM-1
(Fault
classifier)
SVM-2
(Disturbance classifier)Feature vector
Inputs , , ,s s s lossV I P Q
s sInputs , , , ,s s sV I P Q
i
➢SVM-1 classifier
Label ‘0’ for fault condition
Label ‘1’ for other situations
SC Srivastava / IITK WAMCS QIP course 10 May 2019 40
Bus 7Bus 2 Bus 8 Bus 9 Bus 3
Bus 5 Bus 6
Bus 4
Bus 1
Simulation StudiesWSCC 9-bus System
41SC Srivastava / IITK WAMCS QIP course 10 May 2019
Relay No. Normalized Apparent
Impedance
Branch Loss
Sensitivity
Relay Ranking
Index
Relay Ranking
R5-7 0.0994 2.5031 0.0339 1
R7-5 0.1025 2.4308 0.0360 2
R6-9 0.1201 1.6859 0.0608 3
R9-6 0.1226 1.6534 0.0632 4
R8-7 0.5456 2.1443 0.2171 5
R7-8 0.5513 2.1232 0.2215 6
R5-4 0.5643 1.4612 0.3294 7
R4-5 0.5804 1.4376 0.3444 8
R8-9 0.7929 0.7515 0.9001 9
R9-8 0.8056 0.7440 0.9237 10
R6-4 0.9881 0.8614 0.9786 11
R4-6 1.0000 0.8531 1.0000 12
Step-1: Identification of Critical Relays
42SC Srivastava / IITK WAMCS QIP course 10 May 2019
Rf=0.01
Bus 8 Bus 9Bus 7Bus 2 Bus 3
Bus 5 Bus 6
Bus 4
Bus 1
tclg=0.3 sec
Simulation of Power Swing Scenario
SC Srivastava / IITK WAMCS QIP course 10 May 2019 43
5 6 7-200
-100
0
100
200
300
Time (s)
Vo
lta
ge
(p.u
.)
4.5 5 5.5 6 6.5-200
-100
0
100
200
Time (s)
Vo
lta
ge (
p.u
.)
4.5 5 5.5 6 6.5 7
0
0.2
0.4
0.6
0.8
1
Time (s)
Pred
icte
d l
ab
el
(SV
M-1
)
Classification by SVM-1
Fast Swing Slow Swing
4.5 5 5.5 6 6.5
0
0.2
0.4
0.6
0.8
1
Time (s)
Pred
icte
d l
ab
el
(SV
M-1
)
Classification label by SVM-1 Classification label by SVM-1
44SC Srivastava / IITK WAMCS QIP course 10 May 2019
Performance of SVM-1 (Fault Classifier)
2 4 6 8 10 12 14 16 1840
60
80
100
120
140
Time (s)
Volt
age a
t th
e r
ela
y l
ocati
on (
bus-
7)
2 4 6 8 10 12 14 16 18
0
0.2
0.4
0.6
0.8
1
Time (s)
Pre
dic
ted
lab
el
(SV
M-1
)
Voltage unstable case Classifier output
SVM-1 Comparison with ANFIS* Scheme
Scheme Classification accuracy (%) Testing accuracy (%)
ANFIS based classifier 80.02 75.83
SVM based proposed classifier 98.85 96.18
* H. K. Zadeh and Z. Li, “A novel power swing blocking scheme using adaptive neuro-fuzzy inference system, “Electric
Power Systems Research, vol. 78, pp 1138-1146, 2008.
Accuracy of SVM-2 classifier in WSCC 9-bus system
Cases Training accuracy
(%)
Testing accuracy (%)
32 100 96.71
45
Proposed Methodology:
• Logarithm of the singular values of the Hankel matrix formed from
the Measurement data used for model order estimation.
• Second order Taylor series extended TLS-ESPRIT used for accurate
dynamic phasor estimation*.
• Taylor’s second order approximation used to estimate first and
second derivatives of the phasor amplitude and phase derivatives.
• Deviation in the derivative values used to distinguish between fault
and no fault conditions**.
* P. Banerjee and S. C. Srivastava, “An Effective Dynamic Current Phasor Estimator for SynchrophasorMeasurements,” IEEE Transactions on Instrumentation and Measurements, vol.64, no.3, Mar. 2015, pp. 625-637.
** Paramarshi Banerjee, Improved Estimation of Dynamic Phasors, and their Applications in Distance Protection &Stability Assessment, Ph.D. Thesis, IIT Kanpur, January 2015.
Avoiding Relay Unintended Operation
Using Phasor Derivatives
46SC Srivastava / IITK WAMCS QIP course 10 May 2019
Proposed Approach for Relay Blocking
• The predicted and the accurate phasor at the present time
• The vector deviation of the predicted phasor as a percentage
of the amplitude at the fifth previous data window is given by
• Corresponding values for the voltage and the current phasors
are denoted as 𝐸_𝑉 and 𝐸_𝐼, respectively.
• The Relay Blocking (RB) signal is high and releases the relay
for operation when 𝐸_𝑉 and 𝐸_𝐼 are simultaneously greater
than 15% (experimentally established).
2 2
(0), 5
ˆ ˆ(Re( Re( ) (Im( Im( )%
) ) ) )100k k k k
b k
P PE
P P
X −
− + −=
(0) (0), ,
ˆ(0) (0), ,
ˆ ˆ , b k b kj jk kb k b kP X e P X e = =
SC Srivastava / IITK WAMCS QIP course 10 May 2019 47
▪ The proposed algorithm for generating RB signal should remain low for power swings and voltage instability but high for faults .
▪ The test cases considered in WSCC 9 bus system are
• A:Unstable swing;
• B:Fault during unstable swing;
• C:Fault during stable swing;
• D:Voltage Instability;
• E:Fault during voltage instability;
▪ The test cases considered in NE-39 bus system are
• F:Unstable swing;
• G:2 cases of fault during unstable swing;
• H:Stable swing;
• I:Fault during stable swing;
• J:Voltage Instability;
• K:Fault during voltage instability;
Performance Evaluation
SC Srivastava / IITK WAMCS QIP course 10 May 2019 48
Case B: Fault during unstable swing in WSCC 9 bus system
Voltage and current at bus 8 in WSCC 9 bus system for fault during unstable power swing.
RB at bus 8 in WSCC 9 bus system for fault during unstable power swing.
SC Srivastava / IITK WAMCS QIP course 10 May 2019 49
Case D: Voltage Instability in WSCC 9 bus system
Voltage at bus 7 in WSCC 9 bus system for fault during voltage instability.
RB at bus 7 in WSCC 9 bus system during voltage instability
SC Srivastava / IITK WAMCS QIP course 10 May 2019 50
Local Stabilizing Control: Power System Stabilizers
Turbine
Vt
Gen
Governor
Exciter
Voltage
Regulator
PSS
Transmission Line
(Power System Network)
Output
KPSS 1
w
w
sT
sT+
T(s)
Input
PSS
Gain Washout Lead-Lag Compensator
Limiter
51
▪ It provides damping to the
oscillations in the range of
0.2Hz-3 Hz, by modulating
generator field excitation.
▪ The input signal can be rotor
speed or frequency or line
power flow power deviation.
Major Problems
▪ Robustness towards multiple
operating conditions and
topology-needs retuning.
▪ Lack of global observation.
▪ Coordination of multiple
controllers
SC Srivastava / IITK WAMCS QIP course 10 May 2019
52
➢ A two-loop framework of a wide-area stability control system.
➢ The Wide-Area Damping Controller (WADC) in the higher level receives remote
signals, from few pre-selected PMUs located in a wide-
geographical area in the network & provides additional supplementary damping
signals, to some of the pre-decided generators/FACTS devices
POWER SYSTEM
G-1
G-2
G-N
PMU-1
PMU-2
PMU-K
AVR-1
AVR-2
AVR-N
CPSS-1
CPSS-2
CPSS-N
2Δω
NΔω
+
+
+
+
+
+
delay
delay
delay
N_caτ
2_caτ
1_caτ
WADC
delay
delay
delay
1_scτ
2_scτ
K_scτ
1VS
2VS
NVS
KY
2Y
1Y
1Δω
Exciter
Exciter
Exciter
1 2 KY Y Y
1 2 NVS VS VS
SC Srivastava / IITK WAMCS QIP course 10 May 2019
Centralized Wide Area Damping Controller
Some of the major issues:
❑ The choice of the control I/O signals
❑ Control law or algorithm
❑ Latency in communication network
❑ Packet Loss
❑ Packet Disorder
❑ Channel Bandwidth
❑ Communication Failure
53
Issues with Wide-Area Damping Controller
SC Srivastava / IITK WAMCS QIP course 10 May 2019
Wide-Area TS Fuzzy Output Feedback Controller with
Input/Output Signal Selection
54
➢ Coherency based approach* for input/output signal selection
• Data transformed into orthogonal space to make correlated variables
uncorrelated by applying Principal Component Analysis (PCA).
• Self-Organizing Map (SOM) for final data clustering. For clustering data,
few critical line contingencies were considered.
➢ Producing better damping effect to the critical inter-area modes of
oscillations. TS-Fuzzy Controller (WATSF) was applied**. Results
compared with existing H2/H∞ Controller (called as WARDC)+.
* B.P. Padhy, S.C. Srivastava, and N.K. Verma, "A Coherency-Based Approach for Signal Selection for Wide Area Stabilizing
Control in Power Systems", IEEE Syst. Journal, vol.7, no.4, pp.807-816, Dec. 2013.
**B.P. Padhy, S. C. Srivastava, and N. K. Verma, “Robust Wide-Area TS Fuzzy Output Feedback Controller for Enhancement of
Stability in Multimachine Power System”, IEEE Systems Journal, vol.6, no.3, pp.426,435, Sept. 2012.
+ Y. Zhang and A. Bose, "Design of Wide-Area Damping Controllers for Inter area Oscillations," IEEE Transactions on Power
Systems, vol. 23, no. 3, pp. 1136-1143, Nov. 2008.
SC Srivastava / IITK WAMCS QIP course 10 May 2019
Implementation of the proposed method
Line Outage Contingency
Index Value
1.2463
1.2073
1.1953
0.3666
0.1280
55
Line Outage Contingency
Index Value
0.096
0.064
0.058
0.039
0.0374
28 29L
−
1 39L−
1 2L−
2 25L−
16 21L
−
1 2L−
8 9L−
2 25L−
21 22L
−
16 21L
−
Dynamic Contingency Index (39-bus system)
Dynamic Contingency Index (68-bus system)
Dynamic contingency
index is defined
ξ -ξI Jsys sysDCI =
J ξIsys
2600 2800 3000 3200
0.9985
0.999
0.9995
1
1.0005
1.001
1.0015
1.002
Samples
Gen
-Spe
ed in
P.U
.
G1
G2
G3
G4
G5
G6
G7
G8
G9
G10
G11
G12
G13
G14
G15
G16
Gen Speed of 68-bus system for line L1-2 outage –Data used
as input to the signal selection algorithm
56SC Srivastava / IITK WAMCS QIP course 10 May 2019
39
-bu
s N
ew E
ngl
and
Sy
stem
Type. Pole vector approach
Geometric approach
Proposed approach
Generators G1, G5, G7, G9 G3, G7, G8, G9 G3, G5, G9, G10
Power Flow
In Lines
L(8-9), L(1-2), L(2-3), L(14-15),
L(3-18)
L(9-39), L(16-17), L(2-3), L(2-25),
L(4-5)
L(17-27), L(16-19), L(9-39),
L(2-25), L(8-9)
68
-bu
s N
ew
Engl
and
New
Yo
rk S
yste
m
Generators G11, G12, G13, G14, G15, G16
G3, G9, G10, G13, G14, G16
G1, G3, G5, G9, G13, G15
Power Flow
In Lines
L(36-37), L(50-52), L(40-41),
L(34-36), L(1-30), L(9-36)
L(1-2), L(42-52), L(1-47), L(33-38), L(26-27), L(9-30)
L(36-37), L(1-2), L(9-30),
L(50-52), L(9-36), L(4-5)
Signal Selection for 39-Bus and 68-Bus Systems
57SC Srivastava / IITK WAMCS QIP course 10 May 2019
GG15
67
42
52
49
G
G14
G
G16
41
66
68
46
38
33
51
50
45
39
44
35
34
G
G12
64
36
43
37
G
G13
65
40
31
G
G10
62
48 47
1
30
32
G
G11
63
9
8
G
G2
7 5
6
54
4
3
2
G
G1
53
25
G
G8
G
G9
292826
18 17
27 24
21
16
61
G
G3
10
55
11
12
13
14
15 19
G
G5
G
G4
20
57
56
G
G6
58
G
G7
59
23
22
60
WADC
New England and New York Interconnected 68-bus system showing
selected input and output signals with the proposed method58
Comparative Results (68-bus system)
59
30 32 34 36 38 40 42
-2
0
2
x 10-3
Time (sec)
Sp
ee
d D
iv W
7 (
pu
)
3-phase fault
With only PSS
With poledir
With geometric
With proposed WARDC
With proposed WATSF
50 52 54 56 58 60 62 64 66 68 70-100
0
100
200
Time (sec)
P
9_39(M
W)
Load outage
With only PSS
With poledir
With geometric
With proposed WARDC
With proposed WATSF
Speed deviation of Gen 7 and power deviation in line 9-39 due to
3-Φ fault and load outage (68-bus system)
SC Srivastava / IITK WAMCS QIP course 10 May 2019
Network Delay & Packet drop Compensation*➢ A New Time Delay Compensation (TDC) technique with packet
drop has been proposed.
➢ The network latency considered in the application of
synchrophasor assisted wide area control for the Static Var
Compensator (SVC).
➢ The power oscillation modes are estimated online by Modified
Extended Kalman Filter (MEKF) approach.
➢ Delay has been compensated by predicting the dynamics of the
delayed measurement signal.
➢ The performance of the proposed delay compensation scheme has
been tested on 39-bus and 68- bus systems.
➢An insertion sort algorithm has been used for chronological
sequence restoration of the data in case of packet disorder.
* Bibhu P. Padhy, S.C. Srivastava and Nishchal K. Verma, “A Network Delay Compensation Technique for Wide-Area SVC
Damping Controller in Power System,” IEEE PES Transmission & Distribution Conference & Exposition, Chicago, USA,
April 14-17, 2014. 60
Proposed Network Architecture
61
Power
System
m1(KT)
1 1m (KT - τ )
2 2m (KT - τ )
n nm (KT - τ )
2Buffer
1Sensor1
S (t)
nBuffer
1Buffer
nEKFnDelay Compensator
WADC
Satellite
GPS
Clock
rr rr
Other
PMUs
• • •
•••
•••
••
•M
U
X
BN
2EKF2Delay Compensator
1EKF1Delay Compensator
1PMU
1PDC
m2(KT)
2Sensor
2S (t)
2PMU
2PDC
mn(KT)
nSensorn
S (t)n
PMUn
PDC
rr rr• • •
rr rr• • •
Co
mm
un
ica
tio
n
Ne
two
rk•••
•••
•••
•••
Satellite
GPS Clock
nNPDC
2NPDC
1NPDC
Logic To
Calculate Time
Delay
dnΤ
d2T
d1T
Time Stamp of
Incoming Data Packets Control Center
Remote
Location
1y
2y
ny
1mode
2mode
nmode
1Sdc
2Sdc
nSdc
sV
SC Srivastava / IITK WAMCS QIP course 10 May 2019
Time Delay Compensation (TDC) Scheme
62
Incoming
Data
GPS Clock
Buffer
TOP
MEKF
Extract the Time
Stamp Time
stampT
GPST
dT
WADC
Power
System
Network Buffer
k k k kφ ,ω ,λ ,ζk k d
dc s
(λ -ζ T )
k k d
S (kT ) =
e sin(φ -ω T )
TDC
dc sS (kT )
sV
••• •••jy ky
3y 1y
Logic to Calculate
Time Delay
The delay compensated signal can be represented as
,3,
, ,1,
1
.sin( )i k T Td s
d s d s
Nx
dc k k T T i k T T
i
S y e x+
+ +
=
= = , ,4, ,4,(ln )
,2, , ,2,
1
.sin( )i k i k d i k
NA x T x
i k i k d i k
i
e x T x− −
=
= + +
, ,( )
, ,
1
.sin( )i k d i k
NT
i k d i k
i
e T
−
=
= +
SC Srivastava / IITK WAMCS QIP course 10 May 2019
39 Bus Test System The input/output signal selection using the coherency approach.
The remote input signals to the WADC are , and
63
G
G
G
G8
G1037
30 25
18
G139
G
31 11
12
17
26
27
10
32
G
16
13
14
15
G2
G3
28
G G9
29
G G5
34
20
19
G G4
33
24
21 22
G
G G7
36
23
G6
35
38
WADC
PMU 1
PMU 2
7
8
9
6
4
3
5
2
1
1( )tt 2 ( )tt
TSC TCR
SVC
Input 1 Input 2
Input 3
Output
PMU 3
16-19P 1-39
P26-27
P
SC Srivastava / IITK WAMCS QIP course 10 May 2019
Implementation Results
A random input delay of maximum value 1sec has been created in
the input channel-1( ) , 400 ms in the input channel-2( ),
and zero delay created in channel-3( ) as this is a local signal
Packet drop probability is considered as 4%
Test Cases:
A 3-phase fault was applied at bus-16 and bus-26 for 70ms.
A load outage at bus-16 and 28
3-phase fault at bus- 4 followed by outage
Bus connecting 21-22 line contingency
64
1-39P
16-19P
26-27P
4-14L
SC Srivastava / IITK WAMCS QIP course 10 May 2019
Simulation Results
9 11 13 15 17 19 21 23 25 27 29 31 33 35
-4
-2
0
2
4
x 10-3
Time(Seconds) (A)
W4-W
1 (
P.U
.)
9 11 13 15 17 19 21 23 25 27 29 31 33 35-5
0
5
x 10-3
Time(Seconds) (B)
W4-W
1 (
P.U
.)
Without WADC With Phasor Approach
Without WADC With Phasor Approach
10 15 20 25 30 350
0.5
1
Time(Seconds) (A)
(1-
)
10 15 20 25 30 350.38
0.4
Time(Seconds) (B)
Pac
ket
Del
ay(S
ec)
10 15 20 25 30 35-25
-15
-5
5
15
25
Time(Seconds) (C)
P
16
_1
9(p
.u.)
Delayed Signal PMU Output Predicted Signal
1 (A) Packet drop (B)
Random time delay (C)
Tracking of PMU signal
2. (A) 3-phase fault at
bus- 26 for 70ms (B) a 3-
phase fault at bus- 16 for
70ms
65
66
MATLAB as
Server
IPDC as Client
PMUs
Omicron
Amplifiers
Desktop PC
with RSCAD
NI PCI 6704
DAQ Card
SEL PMU
Arbiter PMU
Router
Satellite
GPS clock
3-phase
voltage and
current
signal
Ethernet
Based LAN
Control
signal
RTDS
Hardware Lab Setup used for Real-Time
Simulation of WADC+
+ Addressing
• Optimal input/output
selection
• Robust Control
• Time delay compensation
• Packet drop and packet
disorder compensation
SC Srivastava / IITK WAMCS QIP course 10 May 2019
67
29 33 37 41 45 49 530.52
0.53
0.54
0.55
0.56
0.57
Time(Seconds)
Del
ay
Delay with time for a 3 phase fault at bus-27 (39 bus NE system)
29 33 37 41 45 49 53
-1
-0.5
0
0.5
1
Time(Seconds)
W9-
W1
(P.U
.)
No WADC
Only Simulation
HIL Simulation
Oscillations in the speed deviation of the generator
WADC Simulation and Validation on RTDS
SC Srivastava / IITK WAMCS QIP course 10 May 2019
Conclusions Synchrophasor based WAMCP system will form an
important component of Smart Grid. It offers new paradigmof real time security monitoring and control.
Large number of PMUs are being deployed in thetransmission grid at strategic locations in various countries.
The successful implementation of WAMPC requiresdevelopment of suitable application tools.
PMUs must be embedded with proper dynamic phasorestimation algorithm, complying with the IEEE standards.
It can be effectively utilized for better situational awarenessmonitoring impending system instability, and improvinginter-area mode damping, important to avoid majordisturbances.
68SC Srivastava / IITK WAMCS QIP course 10 May 2019
Thank
You
69SC Srivastava / IITK WAMCS QIP course 10 May 2019