wide-area based intelligent and adaptive transmission system …silicon.ac.in/smart-2015/intelligent...
TRANSCRIPT
Wide-area based Intelligent and Adaptive Transmission System
Protection
S. R. Samantaray School of Electrical Science,
Indian Institute of Technology Bhubaneswar, Email- [email protected]
Smart-grid, Silicon Institite of Technology, Bhubaneswar
12/05/2015 1
Contents
• Introduction to Transmission system relaying
• Relaying attributes
• Challenges due to FACTs and Wind-integration
• PMU and WAMs
• Proposed Intelligent relaying schemes
• Proposed Adaptive Relay setting
• Conclusions
Smart-grid, Silicon Institite of Technology, Bhubaneswar
12/05/2015 2
Distance Relaying and its basic functions:
• Fault Detection • Fault Classification • Fault Location
Typical current and voltage waveforms
Evolution of Relays:
• Electromechanical relays - (1900-present)-1st Generation
• Solid state relays-(1970-1990)- 2ndGeneration
• Digital relays-micro-processor based relay (1982-Present)- 3rd Generation
• Intelligent Relays - (Adaptive Relays) Computer/ DSP/ FPGA based relays with intelligent algorithms- 4thGeneration
Smart-grid, Silicon Institite of Technology,
Bhubaneswar 12/05/2015 5
Attributes of the Relay: • Reliability: Reliability is generally understood to measure the degree of
certainty that a piece of equipment will perform as intended (a reliable relaying system must be dependable and secure)
• Dependability: Dependability is defined as the measure of the certainty that the relays will operate correctly for all the faults for which they are designed to operate.
• Security: Security is defined as the measure of the certainty that the relays will not operate incorrectly for any fault. As a relaying system becomes dependable, its tendency to become less secure increases.
• Speed: Speed of operation is the key indicator for relay performance and it provides the operating time of the relay :
Instantaneous Time delay High speed (50 milliseconds) Ultra high speed (4 milliseconds or less)
Stepped Distance Relay
A B C
Zone 1
Zone 2
Zone 3
F2
A B C
jX
R
Zone 1
Zone 2
Zone 3
Smart-grid, Silicon Institite of Technology, Bhubaneswar
12/05/2015 7
Possible Challenges
• Inclusion of FACTs devices in the modern transmission network.
• Integration of off-shore wind-farms in the transmission network.
• Distinguishing Faults from other conditions such as Power Swings.
Smart-grid, Silicon Institite of Technology, Bhubaneswar
12/05/2015 8
Smart-grid, Silicon Institite of Technology, Bhubaneswar
FACTs embedded in to the power system
TCSC
UPFC
12/05/2015 9
Smart-grid, Silicon Institite of Technology, Bhubaneswar
• The presence of the TCSC in fault loop not only affects the steady-state components but also the transient components .
• While the use of the UPFC improves the power transfer capability and stability of a power system, certain other problems emerge in the field of power system protection, in particular the transmission line protection, affecting greatly the reach of the distance relay.
• In the FACTS-based transmission line, if the fault does not include FACTS device, then the impedance calculation is like an ordinary transmission line, and when the fault includes FACTS, then the impedance calculation accounts for the impedances introduced by FACTS device.
• Thus, it is really challenging to build relays which can be intelligent enough to consider the aforementioned issued due to inclusion of FACTs.
Impact of FACTs devices
12/05/2015 10
Smart-grid, Silicon Institite of Technology, Bhubaneswar
Performance with STATCOM
Performance with UPFC
12/05/2015 11
Smart-grid, Silicon Institite of Technology, Bhubaneswar
• The difficulty that arises in integrating wind-farms is primarily due to uncontrollable wind speed as it continuously varies throughout a day resulting fluctuation in wind-farm output power.
• The output power of a generating unit has a nonlinear relationship with the wind speed and when such a farm is connected to the grid through a transmission line, the transmitted power and the relay end voltage (with respect to grid voltage) fluctuate continuously.
• Further, wind-farm generation capacity also greatly affects the tripping boundary of the distance relay. Thus, fixed setting approach in such an environment will lead to significant error in relaying decision.
• Wind-farm integration to the transmission line may also bring problems such as weak feed or weak source condition and some machines like doubly fed induction generator (DFIG) contributes only about 1.1 p.u (110% of full generation) after a few cycles of the fault inception, resulting into a very weak source
• Moreover, the integration of doubly fed induction generator (DFIG) based wind-farm has serious impacts on the existing distance and differential relaying schemes.
Impact of off-shore wind integration
12/05/2015 12
Smart-grid, Silicon Institite of Technology, Bhubaneswar
PMU and Wide-Area Measurement
12/05/2015 13
Need for Wide-Area Monitoring (WAM)
• Deregulation, competition and increase in complexity of today’s power networks have exacerbated power system stability issues including wide disturbances, which are not ably covered by existing protection and network control systems.
• As power grids get even more heavily loaded by sudden bulk power transfers, the system becomes very vulnerable and even minor equipment failures can result in cascade tripping and eventually, blackouts.
• To ensure system stability in a heavily loaded system, all or most installed components should remain in service and right actions must be taken quickly if the system has not recovered after a serious event.
• To cater to this requirement, the solution is to have real time monitoring. Such a wide area measurement system provides operators with real time knowledge of various instability issues and events as and when they occur.
• A typical wide area measurement or WAMS system is built up on a reliable communication system connecting power stations, network control centers and sub stations.
Smart-grid, Silicon Institite of Technology, Bhubaneswar
12/05/2015 14
Wide Area Monitoring System(WAMS)
GPS satellite
PMU
PMU PMU
PMU
Wide Area Monitoring System use a GPS satellite signal to time-synchronize from phasor measurement units (PMUs) at important nodes in the power system, send real-time phasor (angle and magnitude) data to a Control Centre.
The acquired phasor data provides dynamic information on power systems, which helps operators to initiate corrective actions to enhance the power system reliability.
Smart-grid, Silicon Institite of Technology, Bhubaneswar
12/05/2015 15
PMU based Wide Area Monitoring
Smart-grid, Silicon Institite of Technology, Bhubaneswar
12/05/2015 16
Smart-grid, Silicon Institite of Technology, Bhubaneswar
12/05/2015 17
PMU based WAMs
Phasor Measurement Units (PMU)
• Synchronized phasor measurements are traced from their origins in computer relaying to present applications in power system operation, protection, and control.
• The start of the modern EMS systems based upon state estimators can be said to have begun with the aftermath of the 1965 catastrophic failure of the North-Eastern power grid in North America.
• There was a great deal of research conducted in techniques for determining the state of the power system in real time based upon real-time measurements.
• Of course, there was not the possibility of achieving synchronized measurements in those days, and instead a technique was devised whereby measurements could be obtained by sequential scan and from them the state of the power system estimated by anon-linear state estimator.
• It was recognized that the state obtained in this manner at best described a quasi-steady state approximation to the actual state of the network.
Smart-grid, Silicon Institite of Technology, Bhubaneswar
12/05/2015 18
Phasor measurement
Phasor Measurement Units (PMU) uses the GPS to synchronize the sampling clocks, so that the calculated phasors would have a common reference.
Phasor representation of sinusoid Phasor estimation using DFT
Smart-grid, Silicon Institite of Technology, Bhubaneswar
12/05/2015 19
Smart-grid, Silicon Institite of Technology, Bhubaneswar
12/05/2015 20
Synchronized Measurements
Basic PMU connectivity to Power System
Smart-grid, Silicon Institite of Technology, Bhubaneswar
12/05/2015 21
PMU and GPS
Functional Block diagram of the PMU
GPS satellite transmission for achieving synchronization of sampling clocks in PMUs
Smart-grid, Silicon Institite of Technology, Bhubaneswar
12/05/2015 22
Real-Time PMUs and PDC
SEL PMUs
SEL PDC
Smart-grid, Silicon Institite of Technology, Bhubaneswar
12/05/2015 23
C37.118 Standard for PMU
Accuracy Index Total Vector Error (TVE)
Standard PMU Rates
Where and are measured values
and are theoretical values of the input signal
Smart-grid, Silicon Institite of Technology, Bhubaneswar
12/05/2015 24
C37.118 Compliance for PMU
The level-1 is intended as standard compliance level and evel-0 is provided for applications those can not be served by Level-1
Smart-grid, Silicon Institite of Technology, Bhubaneswar
12/05/2015 25
Smart-grid, Silicon Institite of Technology, Bhubaneswar
12/05/2015 26
Point to Point Communications: Protection application
Smart-grid, Silicon Institite of Technology, Bhubaneswar
Integrated Impedance based protection scheme
12/05/2015 27
Integrated Impedance based protection scheme for TCSC compensated Line
GTNET PMU Block of RSCAD library
Smart-grid, Silicon Institite of Technology, Bhubaneswar
12/05/2015 28
Proposed Dedicated Scheme for TCSC compensated Line
int_
( )
( )
sa ra
asa ra
U UZ
I I
int_
( )
( )
sb rb
bsb rb
U UZ
I I
int_
( )
( )
sc rc
csc rc
U UZ
I I
Integrated Impedance of each phase is defined as:
Smart-grid, Silicon Institite of Technology, Bhubaneswar
12/05/2015 29
Imaginary part of integrated impedance(IPII) During External Fault:
/ /
where 2/Y
2/Y / /
IPII img(2/Y) Large negative value
s r sg rg s sg r rg
sg rg
s r s r
s r s sg r rg
I I I I U Z U Z
Z Z
U U U U
I I U Z U Z
Smart-grid, Silicon Institite of Technology, Bhubaneswar
12/05/2015 30
Smart-grid, Silicon Institite of Technology, Bhubaneswar
Imaginary part of integrated impedance(IPII) During Internal Fault:
ACEs Er
sU rU
sgZ rgZ
lsZ TCSCZ
sZ rZ
sI rI
sgI rgIRf
fI
AC
lrZF
12/05/2015 31
Smart-grid, Silicon Institite of Technology, Bhubaneswar
12/05/2015 32
Imaginary part of integrated impedance(IPII) During Internal Faults:
Smart-grid, Silicon Institite of Technology, Bhubaneswar
12/05/2015 33
Observations: • In normal operating conditions (or in external fault condition), the
sign of the IPII (which reflects the impedance of the line capacitance) is negative with large absolute value.
• In case of internal faults (faults on the line to be protected), the sign of the IPII is mostly positive but may become negative with smaller absolute value for some fault conditions
• Thus, based on the sign and absolute value of IPII, the external and internal faults can be distinguished.
• However, the magnitude of IPII varies while subject to changes in operating parameters of the power system during fault conditions. Thus, setting a threshold on the magnitude of IPII will never work considering wide variations in faulted conditions.
Smart-grid, Silicon Institite of Technology, Bhubaneswar
12/05/2015 34
Variations in Parameters:
• Variation in fault resistance (RF) from 0 to 300Ω
• Variation in fault location: 20%, 30%, 50%,70%, 80%, and 95% of the total line length
• Variation in fault inception angle(FIA): 0,30,60,90
• Different types of fault: a-g, b-g, c-g, a-b, b-c, c-a, ab-g, bc-g, ca-g, a-b-c etc.
• Total cases simulated=6000
Smart-grid, Silicon Institite of Technology,
Bhubaneswar 12/05/2015 35
Smart-grid, Silicon Institite of Technology, Bhubaneswar
• To alleviate the above mentioned problems, the decision making process is further enhanced by cascading the data-mining algorithm such as DT.
• Data mining is defined as the process of discovering patterns in data and is a form of inductive learning.
• DTs are grown through a systematic process known as recursive binary partitioning; a “divide and conquer” approach where successive questions with yes /no answers are asked in order to partition the sample space.
Application of Data-mining
12/05/2015 36
Smart-grid, Silicon Institite of Technology, Bhubaneswar
Advantages of Data Mining Model (Decision Tree)
• It promote decision making
• It can handle high dimension data
• It does not required any domain knowledge or parameter setting, and is therefore suitable for exploratory knowledge discovery
• Learning and classification steps are simple and fast
• Good accuracy
• It perform well despite noisy or missing data (robustness).
• It convert result to a set of easily interpretable rules
• Simple to understand and implement.
12/05/2015 37
Data Mining Model Building for the proposed scheme:
Smart-grid, Silicon Institite of Technology, Bhubaneswar
• In the proposed scheme, imaginary parts of integrated impedance imgZint _ a, imgZint _ b and img Zint _ c are used as input to the DT.
• The target outputs (classes) are categorized as 0(normal or external fault), 1(a-g fault ), 2(b-g fault), 3(c-g fault), 4(a-b/a-b-g fault), 5(b-c/b-c-g fault), 6(c-a/c-a-g fault) and 7(a-b-c fault).
12/05/2015 38
Flowchart:
Smart-grid, Silicon Institite of Technology, Bhubaneswar
12/05/2015 39
Real-Time Digital Simulator (RTDS) Implementataion RTDS unit
PC interface
Ethernet communication
Smart-grid, Silicon Institite of Technology, Bhubaneswar
12/05/2015 40
Transmission system developed on RTDS
Smart-grid, Silicon Institite of Technology, Bhubaneswar
12/05/2015 41
DATA Collection….
Smart-grid, Silicon Institite of Technology, Bhubaneswar
12/05/2015 42
Decision Tree generated with (70-30)% training-testing ratio:
Smart-grid, Silicon Institite of Technology, Bhubaneswar
Img Zint_a< -115
Img Zint_b< -110 Img Zint_b< -266
Img Zint_c< -100 Img Zint_a>= -989 Img Zint_c< -500Img Zint_c< -254
0 (No Fault) 3(c-g Fault) 2(b-g Fault) 5(b-c Fault) 1(a-g Fault) 6(c-a Fault) 4 (a-b Fault) 7(a-b-c Fault)
Yes No
Yes No YesNo
Yes No No NoYesYes No Yes
12/05/2015 43
Confusion Matrix for 30% of test data set:
Smart-grid, Silicon Institite of Technology, Bhubaneswar
Predicted
Actual
0
(Normal)
1
(a-g)
2
(b-g)
3
(c-g)
4
(a-b/a-b-g)
5
(b-c/b-c-g)
6
(c-a/c-a-g)
7
(a-b-c)
0(Normal) 88 0 0 0 0 0 0 0
1(a-g) 0 432 0 0 0 0 0 0
2(b-g) 0 0 499 0 0 0 0 0
3(c-g) 0 0 0 453 0 0 0 0
4(a-b/a-b-g) 0 0 0 0 114 0 0 0
5(b-c/b-c-g) 0 0 0 0 0 118 0 0
6(c-a/c-a-g) 0 0 0 0 0 0 142 0
7(a-b-c) 0 0 0 0 0 0 0 128
12/05/2015 44
Simulation Results on RTDS
IPII magnitude of different phases during remote end internal a-g fault
Trip signals of different phases during remote end internal a-g fault
Fault Inception
(Sec)
10ms
(Sec)
10ms
Smart-grid, Silicon Institite of Technology, Bhubaneswar
12/05/2015 45
IPII magnitude of different phases during remote end internal a-c fault
Trip signals of different phases during remote end internal a-c fault
Fault Inception
Smart-grid, Silicon Institite of Technology, Bhubaneswar
Simulation Results on RTDS
12/05/2015 46
IPII magnitude of different phases during during remote end internal a-b-c fault
Trip signals of different phases during remote end internal a-b-c fault
Fault Inception
Fault inception
Trip Signal from DT
Trip Signal from DT
Trip Signal from DT
Smart-grid, Silicon Institite of Technology, Bhubaneswar
Simulation Results on RTDS
12/05/2015 47
Smart-grid, Silicon Institite of Technology, Bhubaneswar
Scheme D(%) (fault at10% of line)
D(%) (fault at60% of line)
D(%) (fault at90% of line)
Distance Relaying
100 62 10
Proposed Intelligent Relaying 100 100 100
Dependability Comparison between Distance Relaying and proposed relaying scheme for different fault locations
Scheme D(%) (RF=1Ω)
D(%) (RF=100Ω)
D(%) (RF=300Ω)
Current differential
100 85 45
Proposed Intelligent Relaying
100 100 100
Dependability Comparison between current differential scheme and proposed relaying scheme for different fault resistances
Scheme S(%) (Power Swing)
S(%) (External fault)
Integrated Impedance based Pilot protection
scheme[16]
50 92
Proposed Intelligent Relaying
100 100
Security comparison between integrated impedance based pilot protection scheme and proposed relaying scheme
Performance Statistics
12/05/2015 48
Smart-grid, Silicon Institite of Technology, Bhubaneswar
Fault Type
Fault Resistance
Fault Location
(%)
Operation
Proposed Scheme Distance Relaying Current Differential
Scheme
Trip Response Time
(ms)
Trip Response Time
Trip Response Time
A-G 1 10 Yes 15.25 Yes 18.35 Yes 16.15
B-G 10 95 Yes 15.15 No - Yes 16..05
C-G 200 50 Yes 15.05 Yes 18.16 No - A-B 20 20 Yes 14.35 Yes 19.25 Yes 15.64
B-C 30 90 Yes 15.03 No - Yes 15.64
C-A 250 25 Yes 15.23 Yes 18.63 No - A-B-C 300 95 Yes 15.05 No - No -
Performance Comparison for different types of faults in case of tested 9-Bus system
Scheme D (%)
30%
Compensation
D (%)
40%
Compensation
Bypass
Mode
(50cases)
Vernier
mode
(50cases)
Bypass
Mode
(50cases)
Vernier
mode
(50cases)
Proposed
Intelligent
Relaying
100 100 100 100
Effect of TCSC mode and compensation level on Dependability of the proposed relaying scheme
Performance Statistics
12/05/2015 49
Synchrophasors-Assisted IPII-Based Intelligent Relaying for Transmission Lines Including UPFC
AC
EsEr
sV rV
sgZ rgZ
lZ
sZ rZ
sIrI
sgI rgI
Rf
fI
AC
Ish
Zsh
AC
seVZse
UPFC Transmission Line
Equivalent circuit of the UPFC compensated transmission system for an external fault
Smart-grid, Silicon Institite of Technology, Bhubaneswar
IPII for external faults
12/05/2015 50
Smart-grid, Silicon Institite of Technology, Bhubaneswar
IPII is defined as follows
12/05/2015 51
Smart-grid, Silicon Institite of Technology, Bhubaneswar
IPII for internal faults
AC
EsEr
sV rV
sgZ rgZ
lZ
sZ rZ
sIrI
sgI rgI
Rf
fI
AC
Ish
Zsh
AC
seVZse
UPFC Transmission Line
Equivalent circuit of the UPFC compensated transmission system for an internal fault
12/05/2015 52
Smart-grid, Silicon Institite of Technology, Bhubaneswar
12/05/2015 53
Smart-grid, Silicon Institite of Technology, Bhubaneswar
UPFC
Substation-1 Substation-2
500kV
Transmission line
Bus-I Bus-II
EXTRACTION OF IPII OF EACH PHASE
FINAL RELAYING DECISION
TRAINED DT
USED FOR TESTING
CO
MM
UN
ICA
TIO
N
CH
AN
NE
L
CO
MM
UN
ICA
TIO
N
CH
AN
NE
L
PMU-I PMU-II
CTCT
PT
PT
To Circuit Breaker
EXTRACTION OF IPII OF
EACH PHASE
Training period of DT
TRAINED DT
Offline Process using Intel®
Core(TM)i5-2400 [email protected]
Online Process on RTDS platform using
Real-Time PB5 card
Data used as input file to
Rattle software package
To be used on-line
CB
Trip SignalProposed Scheme
12/05/2015 54
Smart-grid, Silicon Institite of Technology, Bhubaneswar
System Model on RTDS
• Variation in fault resistance (RF) from 0 to 300Ω • Variation in fault location: 20%, 30%, 50%,70%, 80%, and
95% of the total line length • Variation in fault inception angle(FIA): 0,30,60,90 • Different types of fault: a-g, b-g, c-g, a-b, b-c, c-a, ab-g,
bc-g, ca-g, a-b-c • UPFC mode of operations: Automatic power flow control
mode(APFC) and Bypass mode
12/05/2015 55
DT Building
Smart-grid, Silicon Institite of Technology, Bhubaneswar
12/05/2015 56
Trip signals of different phases during remote end internal a-g fault
Fig: IPII magnitude of different phases during remote end internal a-g
Smart-grid, Silicon Institite of Technology, Bhubaneswar
Performance assessment
12/05/2015 57
Trip signals of different phases during remote end internal a-c fault
IPII magnitude of different phases during remote end internal a-c fault
Smart-grid, Silicon Institite of Technology, Bhubaneswar
Performance assessment
12/05/2015 58
Trip signals of different phases during remote end internal a-b-c fault
IPII magnitude of different phases during remote end internal a-b-c fault
Smart-grid, Silicon Institite of Technology, Bhubaneswar
Performance assessment
12/05/2015 59
Smart-grid, Silicon Institite of Technology, Bhubaneswar
Dependability comparison
Performance assessment
12/05/2015 60
Smart-grid, Silicon Institite of Technology, Bhubaneswar
Performance with Different Fault Resistance
Effect of mode and compensation level
Performance assessment
12/05/2015 61
Smart-grid, Silicon Institite of Technology, Bhubaneswar
Comparison with existing relay
Performance assessment
12/05/2015 62
Smart-grid, Silicon Institite of Technology, Bhubaneswar
Phase Angle of the Positive sequence integrated impedance (PAPSI) based wide-area back-protection scheme
12/05/2015 63
Positive sequence integrated impedance (PSII) based wide-area back-protection scheme
Smart-grid, Silicon Institite of Technology, Bhubaneswar
12/05/2015 64
Positive sequence diagram for an internal fault
Wide-Area adaptive transmission system protection
Smart-grid, Silicon Institite of Technology, Bhubaneswar
12/05/2015 65
PAPSII for External fault
Positive sequence diagram for an external fault
In case of external faults, the positive sequence currents flowing into the protection zone is very less as compared to the line charging current which flow into the zone of protection (transmission line). The sign of the angle becomes negative as the current involved is the line charging current. This situation remains same for no-fault condition as is very small as compared to the fault current, resulting in similar PAPSII condition as external fault situation.
Smart-grid, Silicon Institite of Technology, Bhubaneswar
12/05/2015 66
Operating criteria
If -20<PAPSII<180 Then, there is an internal fault And if -180<PAPSII<-30 Then, there is an external fault/No-Fault If -20<PAPSII<-30 Then, it's a dead zone
0 deg
90 deg
-90 deg
180 deg
-30deg
Zones of
Internal Fault
Zones of External
Fault/No-Fault
-20deg
Dead Zone
Generator-3Load A
BUS-7
BUS-1
BUS-4
BUS-3BUS-9BUS-8BUS-2
Generator-1
Generator-2
BUS-5BUS-6
PMU PMU
PMU
PMU
PMU
PMU
Load B Load C
Validation on RTDS platform
Schematic of modified WSCC-9-bus system Time(s)
Threshold
Ma
gn
itu
de
of
PA
PS
II(d
eg
)
PAPSII values during voltage inversion following an a-g fault in line 7-8.
Smart-grid, Silicon Institite of Technology, Bhubaneswar
12/05/2015 67
PAPSII during unbalanced Fault with voltage Inversion:
Phase voltages before(upper figure) and after(lower figure) the capacitor following an a-g fault at 0.04 sec.
PAPSII values during voltage inversion following an a-g fault in line7-8.
Smart-grid, Silicon Institite of Technology, Bhubaneswar
12/05/2015 68
PAPSII during Unbalanced Fault with Current Inversion:
Phase currents at both ends of line 7-8 during current Inversion following an a-g fault at 0.04 sec.
PAPSII values during current inversion following an a-g fault in line 7-8.
Smart-grid, Silicon Institite of Technology, Bhubaneswar
12/05/2015 69
PAPSII during Load Encroachment:
Impedance trajectory of relay at bus-7 during load encroachment
Response of PAPSII during Load encroachment
Smart-grid, Silicon Institite of Technology, Bhubaneswar
12/05/2015 70
PAPSII during Results for Power Swing:
Response of PAPSII during stable power swing Impedance trajectory of relay at bus-7 during power swing
Smart-grid, Silicon Institite of Technology, Bhubaneswar
12/05/2015 71
Smart-grid, Silicon Institite of Technology, Bhubaneswar
Intelligent Differential Relaying Scheme
12/05/2015 72
Smart-grid, Silicon Institite of Technology, Bhubaneswar
• Wind-farm integration to the transmission line may also bring problems such as weak feed or weak source condition.
• The fault current in case of phase faults depends on the amount of generation at the instance of fault and the fault contribution characteristics of the machines.
• Some machines like doubly fed induction generator (DFIG) contributes only about 1.1 p.u (110% of full generation), after a few cycles of the fault inception, resulting into a very weak source.
• Further, the fault current contribution of wind turbines with crowbar protected DFIG affects the performance of existing current differential and pilot protection schemes .
• When both UPFC and Wind-farms are integrated together in the transmission lines, the system becomes more complicated and the performance of the conventional relaying scheme is greatly affected.
• Thus, there is a strong motivation in developing a dedicated relaying strategy for transmission line protection including UPFC and wind-farms together.
Motivation
12/05/2015 73
Decision tree-induced fuzzy rule-based differential relaying for transmission line including unified power flow controller and wind-farms
Smart-grid, Silicon Institite of Technology, Bhubaneswar
12/05/2015 74
Smart-grid, Silicon Institite of Technology, Bhubaneswar
Operating parameter variation
• Variation in fault resistance (Rf) from 0 to 100Ω • Variation in fault location: 20%, 30%, 50%,70%, 80%, 85%, and
90% of the total line length • Variation in fault inception angle(FIA): 0,30,60,90 • Variations in source impedance angle: 30 % from normal value. • Different types of fault: a-g, b-g, c-g, a-b, b-c, c-a, ab-g, bc-g, ca-
g, a-b-c • UPFC series injected voltage (Vse) varied for 0-15% of the
normal voltage • UPFC voltage phase angle(θse) varied from 0-360 • UPFC Control mode (Automatic power flow control mode and
bypass mode) • Variation in wind speed: 10m/s, 15m/s, 20m/s • Reverse power flow • Remote in-feed • Noisy Environment (Singal to Noise Ration: SNR 20dB)
12/05/2015 75
Smart-grid, Silicon Institite of Technology, Bhubaneswar
The System Studied: A 500kV, 50Hz power system: Single circuit Transmission Line with UPFC and Wind-farm). In this power system, there are two substations (sending end and receiving end), and one UPFC located at the mid-point of the transmission line (distributed model). Wind-farm is connected at the receiving end of the studied system. Hence, the system consists of two sources, UPFC and its associated components and a 400 km transmission line.
Line parameters:
12/05/2015 76
Smart-grid, Silicon Institite of Technology, Bhubaneswar
Relay
Substation-1
(Sending end)Substation-2
(Receiving end)
400 km,500kV
Transmission
line
Feature
Extraction at
Bus-1
Feature
Extraction at
Bus-2
CT CT
PT PT
Zse
Vse
Zsh
Vsh
UPFC
Wind-
Farm
Relay
Substation-1
(Sending end)
Substation-2
(Receiving end)Zse
Vse
Zsh
Vsh
UPFC Wind-
Farm
Single circuit transmission line with UPFC and Wind-farm
Double circuit transmission line with UPFC and Wind-farm
The System Studied
12/05/2015 77
Initial Features Used:
Smart-grid, Silicon Institite of Technology, Bhubaneswar
12/05/2015 78
Smart-grid, Silicon Institite of Technology, Bhubaneswar
X15< -312
0 (No Fault)
9(c-a-g Fault)
2(b-g Fault) 8(b-c-g Fault)
5(b-c Fault) 6(c-a Fault) 4 (a-b Fault)
10(a-b-c
Fault)
Yes No
1(a-g Fault)
7(a-b-g Fault)
5(b-c Fault)
YesYes
Yes
YesYes
Yes
Yes
YesYes
No
NoNo
NoNo
No
No
NoNo
X11>=1.6 X12>= 1.6
X12>=1.6X10>=1.6 X11< 1.6 X10>=1.6
X12>=1.5 X10>=1.6X11>=1.6
X12>=3
3 (c-gFault)
Yes
No
DT Building
12/05/2015 79
Smart-grid, Silicon Institite of Technology, Bhubaneswar
DT-Fuzzy Transformation
12/05/2015 80
Smart-grid, Silicon Institite of Technology, Bhubaneswar
Fuzzy Rule base
12/05/2015 81
Smart-grid, Silicon Institite of Technology, Bhubaneswar
(i) Dependability (D): Total number of fault cases predicted / Total number of actual fault cases.
(ii) Yield (Y): Total number of correct fault cases predicted / Total number of fault cases predicted. (iii) Security(S) = Total number of external faults predicted as external fault / Total number of external faults.
Performance Assessment
12/05/2015 82
Smart-grid, Silicon Institite of Technology, Bhubaneswar
Scheme Dependability (Fault at 10%
of the line)
Dependability (Fault at 80 %
of the line)
Dependability (Fault at 95% of the line)
DT with one end data
100 40 10
DT-Fuzzy 100 100 100
Scheme D (%) (fault at10% of
line)
D (%) (fault at60% of
line)
D (%) (fault at90% of
line) Distance
Relaying(Mho Characteristics)
100 62 12
Proposed Relaying Scheme
100 100 100
DT-induced Fuzzy rule base for fault classification of transmission line including UPFC and Wind-farm
Dependability Comparison between Conventional distance relaying and proposed DT-Fuzzy based relaying
Performance Assessment
12/05/2015 83
Smart-grid, Silicon Institite of Technology, Bhubaneswar
Features without noise
Actual class Predicted class Dependability (%)
Yield (%)
L-G L-G 100 100
L-L L-L 100 100
L-L-G L-L-G 100 100
L-L-L L-L-L 100 100
Features with SNR 20dB
L-G L-G 100 100
L-L L-L(49cases) + L-L-G(1cases)
100 98
L-L-G L-L-G 100 100
L-L-L L-L-L 100 100
Dependability and Yield comparison for different types of faults
Scheme D (%) S (%) Distance Relay (Mho
Characteristics) 65 75
Presented Scheme in [22]
72 72
Presented Scheme in[25]
83 88
Proposed Scheme 100 100
Dependability and Security comparison for different types of faults
12/05/2015 84
Smart-grid, Silicon Institite of Technology, Bhubaneswar
Types of fault Dependabilityof DT_Fuzzy(%)
Yield of DT_Fuzzy(%)
L-G 99.95 100
L-L 100 100
L-L-G 100 98.96
L-L-L 100 99.68
Performance Assessment for Remote Infeed Line
12/05/2015 85
Smart-grid, Silicon Institite of Technology, Bhubaneswar
Cross-Differential Protection scheme for Transmission Lines including UPFC
12/05/2015 86
Smart-grid, Silicon Institite of Technology, Bhubaneswar
N
πkni
eN-
k NT
nT,j-kwkTx
NT
njT,S
21
0
The expression for discrete ST becomes
The window function in the discrete domain is chosen as:
22
2
2
2r
c
NT
nbajTexp
πr
c
NT
nba
NT
njT,w
Fast discrete S-Transform (FDST) can be achieved by :
Appropriate choice of the frequency scaling is important for fast computation of the discrete S-Transform algorithm.
Fast discrete S-Transform
))n,k(W(FFT*)k(X)n,k(G
1
0
2N
k
)N/ikexp()n,k(G)n,k(G
FFT
Inverse FFT
)n,k(ie)n,k(A)n,k(S
12/05/2015 87
Smart-grid, Silicon Institite of Technology, Bhubaneswar
CUSUM-based cross-differential detection
Fault detection time (or sample point)
CUSUM for any signal is shown as follows
Cross Differential Energy
12/05/2015 88
Smart-grid, Silicon Institite of Technology, Bhubaneswar
System Studied
12/05/2015 89
Smart-grid, Silicon Institite of Technology, Bhubaneswar
CUSUM-based cross-differential fault detection in the parallel transmission system with UPFC, for ACG fault on Circuit-1: (a) currents in Circuit-1, (b) currents in Circuit-2, (c) |CS (I1)| - |CS (I2)| when fault distance = 70%, Rfault = 10 Ω, Vse = 6%, θse = 60°, (d) |CS (I2)| - |CS (I1)| when fault distance = 70%, Rfault = 10 Ω, Vse = 6%, θse = 60°.
Performance Assessment
12/05/2015 90
Smart-grid, Silicon Institite of Technology, Bhubaneswar
FDST and spectral energy based fault classification for ACG fault on Circuit-1, over a window of one cycle: (a) three-phase currents in Circuit-1, (b) FDST contours for phase-a current, (c) FDST contours for phase-b current, (d) FDST contours for phase-c current, (e) fault classification for phase-a using spectral energies,
Spectral energy based fault classification
12/05/2015 91
Smart-grid, Silicon Institite of Technology, Bhubaneswar
(f) fault classification for phase-b using spectral energies, (g) fault classification for phase-c using spectral energies.
Spectral energy based fault classification
12/05/2015 92
Smart-grid, Silicon Institite of Technology, Bhubaneswar
Fault Type
Condition-1 Condition-2
a b c a b c
AG 0.6136 0.0000 0.0000 0.1599 0.0009 0.0006
ABG 0.4026 0.2156 0.0002 0.1256 0.1186 0.0006
BC 0.0003 0.3567 0.2221 0.0000 0.2175 0.1863
ABCG 0.4589 1.1658 1.068 0.4134 0.3662 0.5569
Change in Energy during fault situation
Fault
Type
Rfault
(Ω)
Loca-
tion
(%)
Relay Unit Operation time
(ms) Faulted
phases a b c
AG 5 10 9.12 - - a2
BG 10 20 - 9.86 - b2
ABG 20 25 11.25 11.62 - a2, b2
BCG 15 15 - 11.49 10.53 b2, c2
ACG 5 20 9.58 - 9.58 a2, c2
AB 5 5 11.25 11.62 - a2, b2
BC 2 25 - 10.59 10.59 b2, c2
ABCG 8 30 11.25 11.49 10.57 a2, b2, c2
Response of the proposed algorithm for near end fault in Circuit-2
Performance Assessment
12/05/2015 93
Smart-grid, Silicon Institite of Technology, Bhubaneswar
Performance Assessment
Fault
Type
Rfault
(Ω)
Loca-
tion
(%)
Relay Unit Operation time
(ms) Faulted
phases a b c
BG 100 85 - 16.47 - b1
CG 50 75 - - 12.28 c1
BCG 150 95 - 18.45 16.31 b1, c1
CAG 40 80 12.40 - 12.40 c1, a1
AB 100 70 13.20 15.66 - a1, b1
BC 80 75 - 15.66 16.54 b1, c1
CA 120 65 15.79 - 12.70 c1, a1
ABCG 100 70 13.20 15.66 12.70 a1, b1, c1
Response of the proposed algorithm for far end fault in Circuit-1
12/05/2015 94
Smart-grid, Silicon Institite of Technology, Bhubaneswar
FDST and spectral energy based fault classification for BG fault on Circuit-1, over a window of one cycle: (a) three-phase currents in Circuit-1, (b) FDST contours for phase-a current, (c) FDST contours for phase-b current, (d) FDST contours for phase-c current, (e) fault classification for phase-a using spectral energies, (f) fault classification for phase-b using spectral energies, (g) fault classification for phase-c using spectral energies.
Performance Assessment
12/05/2015 95
Smart-grid, Silicon Institite of Technology, Bhubaneswar
Sl. No.
Vse
(in
%)
Rfault
(Ω)
Relay Unit Operation time
(ms) Faulted
phases a b c
1 0 10 12.18 13.95 11.52 a1, b1, c1
2 3 10 12.18 13.95 11.52 a1, b1, c1
3 5 10 11.94 12.57 10.03 a1, b1, c1
4 7 10 11.16 13.09 10.09 a1, b1, c1
5 10 10 11.70 12.56 10.49 a1, b1, c1
6 0 140 14.01 - 14.01 c1, a1
7 3 140 14.70 - 12.11 c1, a1
8 5 140 15.85 - 13.26 c1, a1
9 7 140 16.68 - 15.70 c1, a1
10 10 140 17.98 - 16.44 c1, a1
RESPONSE OF THE PROPOSED ALGORITHM FOR A FAR-END FAULT ON CIRCUIT-1
WITH VARIATION IN UPFC SERIES VOLTAGE
Sl. No. θse
(in °)
Rfault
(Ω)
Relay Unit Operation time
(ms) Faulted
phases a b c
1 0 5 - - 12.18 c1
2 45 5 - - 12.18 c1
3 90 5 - - 12.18 c1
4 180 5 - - 11.94 c1
5 270 5 - - 13.09 c1
6 360 5 - - 12.56 c1
7 0 130 - 10.12 14.57 b1, c1
8 45 130 - 10.12 14.57 b1, c1
9 90 130 - 11.69 15.73 b1, c1
10 180 130 - 15.85 15.85 b1, c1
11 270 130 - 15.60 16.88 b1, c1
12 360 130 - 16.23 17.78 b1, c1
RESPONSE OF THE PROPOSED ALGORITHM FOR A NEAR-END FAULT ON CIRCUIT-1 WITH VARIATION IN
UPFC SERIES VOLTAGE PHASE ANGLE
Performance Assessment UPFC parameter ariation
12/05/2015 96
Smart-grid, Silicon Institite of Technology, Bhubaneswar
0.5 0.505 0.51 0.515-0.1
-0.05
0
0.05
0.1
Time [s]
(E1
,a-E
2,a
),(E
2,a
-E1
,a)
SET = 0.0884
(E2,a
-E1,a
)
(E1,a
-E2,a
)No trip
signal issued
0.508 0.51 0.512 0.514 0.516 0.518 0.52 0.522-0.1
-0.05
0
0.05
0.1
Time [s]
(E1
,a-E
2,a
),(E
2,a
-E1
,a)
SET = 0.0884
(E2,b
-E1,b
)
(E1,b
-E2,b
)
No trip signal issued
0.56 0.561 0.562 0.563 0.564 0.565-0.1
-0.05
0
0.05
0.1
Time [s]
(E1
,c-E
2,c
),(E
2,c
-E1
,c) SE
T = 0.0884
(E2,c
- E1,c
)
(E1,c
- E2,c
)
No trip signal issued
Response to External Fault
FDST and spectral energy based fault classification for external ABCG fault in section B1-B3
12/05/2015 97
Smart-grid, Silicon Institite of Technology, Bhubaneswar
Sl. No. Vse
(in %)
Rfault
(Ω)
Relay Unit Operation time
(ms) Faulted
phases a b c
1 0 10 - - 13.12 c1
2 3 30 13.85 - - a1
3 5 50 14.24 13.45 - a1, b1
4 7 100 14.47 - 15.89 a1, c1
5 10 150 15.45 14.21 13.64 a1, b1, c1
θse
(in °)
Rfault
(Ω) Relay Unit Operation time (ms)
Faulted
phases
1 0 10 - 14.12 - b1
2 45 30 - 13.45 15.84 b1, c1
3 90 50 - 15.23 14.59 a1, c1
4 180 100 - 15.12 14.89 b1, c1
5 270 150 16.12 15.99 15.32 a1, b1, c1
RESPONSE OF THE PROPOSED RELAYING ALGORITHM FOR A FAR-END FAULT ON
CIRCUIT-1 ON RTDS PLATFORM
Performance Testing on RTDS platform
12/05/2015 98
Smart-grid, Silicon Institite of Technology, Bhubaneswar
Adaptive Distance Relay Setting
12/05/2015 99
Smart-grid, Silicon Institite of Technology, Bhubaneswar
Power system with UPFC and Wind-integration
Fault before UPFC
12/05/2015 100
Smart-grid, Silicon Institite of Technology, Bhubaneswar
Fault after UPFC
12/05/2015 101
Smart-grid, Silicon Institite of Technology, Bhubaneswar
Adaptive Relay Setting
12/05/2015 102
Smart-grid, Silicon Institite of Technology, Bhubaneswar
Adaptive Relay Setting
12/05/2015 103
Smart-grid, Silicon Institite of Technology, Bhubaneswar
Adaptive Relay Setting
12/05/2015 104
Smart-grid, Silicon Institite of Technology, Bhubaneswar
Adaptive Relay Setting
12/05/2015 105
Smart-grid, Silicon Institite of Technology, Bhubaneswar
Adaptive Relay Setting
12/05/2015 106
Smart-grid, Silicon Institite of Technology, Bhubaneswar
Performance testing during power swing
12/05/2015 107
Conclusions:
• Integrated impedance-based intelligent relaying is a new paradigm for EHV lines including TCSC and UPFC.
• It can identify effectively faulty phases and distinguish external faults from internal faults with high degree of dependability and security.
• DT-Fuzzy based differential relaying based on multiple parameters provides fast and accurate fault classification for line employing FACTs and Wind integration.
• Cross differential protection scheme provides reliable, fast and accurate relaying scheme based on time-frequency domain.
• Implementation of the above schemes on RTDS platform establishes the potential ability of each scheme for respective protection measures.
• The proposed adaptive relay setting is highly promising for line with FACTs and Wind-integration.
Smart-grid, Silicon Institite of Technology, Bhubaneswar
12/05/2015 108
Smart-grid, Silicon Institite of Technology, Bhubaneswar
The true sign of intelligence is not knowledge but imagination
Albert Einstein
12/05/2015 109
Smart-grid, Silicon Institite of Technology, Bhubaneswar
Thank You
12/05/2015 110