System Power
Resume:
IMPROVED TRANSMISSION LINE PROTECTION DURING CASCADING EVENTS
M. Kezunovic and C. Pang
Texas A&M University
United States
SUMMARY
Relay operations during cascading events play an important role for the power system
stability. Without knowing the whole system information, distance relay may make wrong
decision from the view of system reliability since it operates only according to local
measurements. This paper introduces an improved transmission line protection scheme during
cascading events. It utilizes the interactive method of system-wide and local monitoring and
analysis tools for on-line automated analysis of protective relay operation. The new system-
wide analysis tools based on Vulnerability Index (VI), Margin Index (MI), and Equivalent
Parallel Path Approximation (EPPA) allow the vulnerable transmission lines of the system
and vulnerable conditions to be identified. Then new local on-line monitoring and control
tools are capable of analyzing performance of relay operation and informing the centralized
tool about the outcomes that will be invoked to mitigate incorrect or undesirable relay action.
A case study is presented to demonstrate the improved transmission line protection scheme
using the new tools.
KEYWORDS
Transmission Line Protection, Security Analysis, Topological Processing, Power Flow, Distance
Relay, Cascading Event
M. Kezunovic and C. Pang are with the Department of Electrical and Computer Engineering,
Texas A&M University, College Station, TX 77843-3128, USA ( emails:
*******@***.****.***, *******@**.****.***)
I. INTRODUCTION
Power systems are getting more stressed under new market rules introduced by deregulation. The
reliability and security are the main issues of the current power systems. Although many factors will
affect the power system operation, cascading outage, especially the large-scale cascading outage, is a
key factor which may cause great economic loss and impact on people s life. Inadequate
understanding of unfolding events, relaying problems, bad weather conditions, human errors, etc. may
contribute to power system cascading outages [1]. Among those factors, relaying problems are the
major part to cause for cascading outage. According to the survey results, relay problems contribute to
75% of large area disturbances and cascading blackouts [2, 3]. The Western Coast System Blackout in
1996 [4] and Northeastern System Blackout in 2003 [1] clearly demonstrated the influence of relay
problems. It is extremely important to study the relay performance during cascading events.
Conventional solution to transmission line protection is to calculate the related relay settings based
on a short circuit study, and then let the relays make their decisions based on local measurements
without knowing or considering the entire system condition. In certain situations these decisions may
not be appropriate from the view of the entire system reliability consideration, which could cause large
area relay misoperations leading to system blackout. If the local information from substations cannot
be sent to the control center, the system operator may not be able to make decision to issue corrective
control, and hence may loose the precious possible chance to mitigate the cascading events. So better
understanding of the interaction between the relay operation and reliability of the power system
operation is needed, especially during cascading events. This may be achieved through better
monitoring of the interaction between power system operation and relay operation, and performing
corrective relaying action as a consequence.
This paper explores the means for on-line automated analysis of protective relay operations during
cascading events and presents an improved method for transmission line protection. The purpose is to
allow for automated corrective actions in the case of misoperations or unintended operations of
transmission line protection. By using power system static security analysis based on Vulnerability
Index (VI) and Margin Index (MI), the vulnerable transmission lines of the system and vulnerable
conditions are identified. The possible overload problems when those vulnerable conditions occur are
predicted by utilizing Equivalent Parallel Path Approximation (EPPA) based on the topology
processing technique. The security of the relays corresponding to those vulnerable transmission lines
will be increased. Possible relay operation correction may be executed. This improved transmission
line protection method is based on the interactive system/local monitoring and analysis scheme.
II. ARCHITECTURE OF THE NEW PROTECTION SCHEME
Correct operation of protective relays is critical to assuring a secure and reliable operation of power
systems. To our knowledge, no on-line analysis tools that will allow for automated monitoring of
protective relay operations to be able to perform corrective actions in the case protection dependability
and/or security are compromised are available today. The problem of the analysis reduces down to
two steps: analysis of the faults and analysis of relay operations. Why the two steps seem to be quite
straight forward, it is not as simple to develop on-line tools for the automated analysis. The complexity
of implementation and deployment is associated with the type of data made available for the analysis,
as well as the objective of the analysis.
The improved transmission line protection method requires coordinated system-wide and local
monitoring and control tools that will improve both system security and local protection. The
architecture of this interactive protection scheme is shown in Fig. 1. The approach for automated
analysis and correction of protective relay operation uses a steady state approach utilizing the power
flow method and topology processing method to identify vulnerable lines due to the stressed operation
conditions. Power flow method is used in static contingency analysis to identify overload and low
voltage problems due to transmission line outage. The security analysis tools based on Vulnerability
Index and Margin Index are used to identify the vulnerable transmission lines of the system and
1
vulnerable conditions.
Fig. 1. Overall interactive scheme for protection
For those vulnerable transmission lines, the possible relay operations can be verified by performing
dynamic analysis of relay behavior. Dynamic bus voltage phasors are calculated from the time-domain
transient stability analysis. They are used to calculate the apparent impedance seen by distance relay
and obtain the dynamic impedance trajectory [6]. If the apparent impedance falls into the relay
protection zone longer than the setting time, the relay will operate based on its design character. This
dynamic approach can identify vulnerable relay locations (lines) and vulnerable fault conditions.
In the mean time, by using Equivalent Parallel Path Approximation method, the possible overload
transmission lines will be rapidly located under the condition of vulnerable lines being switched out.
Once such a condition is identified, the system vulnerability tool invokes on-line monitoring tools
located at the substations. The local tools are capable of analyzing performance of local relay
operation and informing the centralized tool about the outcomes. They are also capable of performing
a corrective action to mitigate incorrect or undesirable relay operation as needed.
III. THE MAIN APPROACH
A. Power System Security Analysis
Several large area blackouts illustrate that due to lack of real time and precise information about
system events, the operators loose the chance to detect and mitigate the cascading outage. The reason
is that sometime the power system will be in a much different state from what was predicted in the
planning and other off-line studies. Many promising ideas for security analysis were proposed in the
past [7-9]. Recently, a new approach based on Vulnerability Index as well as Margin Index is
presented in [5, 10]. VI and MI provided precise vulnerability and margin information respectively for
both the individual elements and the whole system performance. According to the different equipment
groups in the power system, different equations and descriptions for generator level, bus level and
transmission line level analysis are discussed. The more detailed description for the generation level
and bus level analysis can be found in [5].
We can get the exact Vulnerability Index (VI) value for each branch under various system
conditions, which could be used to evaluate the security and performance for individual system
elements. The larger the VI, the more vulnerable the system or individual element..
The Margin Index (MI) of individual line distance relay is another very important parameter. It is
well know that the smaller the normalized apparent impendence seen by the distance relay at no fault
condition, the more likely the relay may misoperate since the apparent impendence may fall into the
distance relay zone 3 or even zone 2. So the margin index and vulnerable index are closely related the
value of the apparent impedance. The smaller the MI, the more vulnerable the distance relay decision.
2
The on-line monitoring tools located at the substations will be used to monitor those vulnerable
distance relays.
B. Fast Overload Transmission Line Search Method
The apparent impedance seen by distance relay may fall into the relay protection zone longer than
the setting time under overload conditions for transmission lines, which will cause relay misoperation.
Paper [11] presents a novel Network Contribution Factor (NCF) method to relieve the overload and
low voltage problems. Single-parameter/multi-parameter analysis and control can be achieved. NCF
method is demonstrated to be fast and accurate by the full AC load flow analysis. However, the
computation time for matrix inversion will increase dramatically with the increase of the power system
size. Here we provide a new concept of Equivalent Parallel Path Approximation for search for the
overload transmission lines rapidly.
We consider the power system as a linear network system for approximation. When one line is
tripped in the system network, for example Lij, we can take it into account by adding one admittance
between the node i and node j, which equal to the negative value of the branch admittance [12].
According to the Superposition Theorem, if a linear system is driven by more than one independent
power source, the total response is the sum of the individual responses. So the network status after one
line is tripped will be equal to the sum of the initial network and compensated network.
With the full AC power flow, the system initial status can be known. We focus on how to get the
compensated status for the system. From DC power flow, under the approximation conditions
xij, sin( i j ) i j, Vi V j 1.0 ), the approximation power flow for node i and j is:
( rij
i
Pij = j
(1)
x ij
Its DC network model can be easily obtained:
Vi V
I ij = j
(2)
rij
So if we could get the current distribution with the equivalent current source in node i and j for the
DC network, we will know the power flow distribution. It is very easy to get the exact current value
for each transmission line according to circuit theory. But its meaning is not significant under the DC
power flow for its approximation. Our purpose is to find the maximum current branches, which should
be the same branches carried the maximum distributed power flow. Here we present Equivalent
Parallel Path Approximation method to find the possible group lines which have the maximum power
flow.
For the DC network under single current source, the Loop Analysis method could offer the
analytical solutions if the independent loops can be identified. If the network is n-node b-branch
system, the number of independent loops will be [ b ( n 1) ] . Obviously, the total number of
loops which contain the current source is greater or equal to [ b ( n 1) ] . Those loops are defined
as equivalent paths, which carry some amount of current from node j to node i. In other words, the
whole non-source network can be decomposed into a group of parallel branches connected with node i
and j, which is shown in Fig. 2.
i
Non-sourse
Vi
Network
I ij
i p 2 i p1
i pm
j
Vj
Fig. 2. Equivalent Parallel Path Decomposition
For each Loop Analysis equation, the self-resistance is much larger than the mutual-resistance. We
3
will omit the mutual-resistance for approximation. If we want to find the maximum value current path,
we only need to locate the path with the minimum loop resistance. As a conclusion, we will locate the
possible overload on a transmission line under the conditions of one line being tripped by finding the
path with minimum total reactance.
Once the influence on each transmission line caused by one line outage is decided, the possible
overloaded lines could be identified by utilizing topological processing and equivalent parallel path
approximation. This approach is much faster since it has no matrix computation, which will benefit the
online application.
IV. CASE STUDY
We will take an example of the standard IEEE 39-bus New England System, as shown in Fig.3, to
demonstrate the proposed approach. The detailed data for full AC power flow running can be found at
[13]. The steady state results, including bus voltage, transmission line flows and generator outputs,
will be obtained.
Fig. 3. IEEE 39-bus New England System
We assigned all weights as 1, voltage variance as 0.075 p.u., PQ bus voltage magnitude limits as 1.0
p.u., transmission line transfer limits as 10.0 p.u., and line bus voltage angle difference limits as 40
degrees. The vulnerability index and margin index values for transmission lines are calculated, which
is shown in Table I. Different loading conditions are considered, including base load, 1.1 times the
base load, and 1.2 times the base load.
From Table I, we can see that the transmission line is more vulnerable with the load increasing. This
is consistent with practical experience.
The margin index for transmission line distance relay is calculated. The vulnerable lines will be
ranked by margin indices. Table II gives the top four most vulnerable lines under different loading
conditions. Similarly, relay is more vulnerable with the load increasing.
TABLE I
VULNERABILITY INDEX VALUES OF TRANSMISSION LINES AT DIFFERENT LOADING CONDITONS
Pl Ql Qc line_ ang relay line_off
Base load 2.856 0.132 0.021 0.182 0.192 0
1.1 load 3.457 0.180 0.012 0.225 0.233 0
1.2 load 4.116 0.244 0.008 0.273 0.282 0
4
TABLE II
MARGIN INDEX VALUES OF TOP 4 LINES AT DIFFERENT LOADING CONDITONS
Line No base load 1.1 base load 1.2 base load
37(B6-31) 4.204 3.491 2.866
38(B10-32) 4.798 4.123 3.521
42(B23-36) 4.782 4.055 3.444
45(B29-38) 5.711 4.977 4.341
According to the trajectory of impedance seen by the relay at bus 26 (B26) of L33 (B29-26), which
is shown in Fig. 4, the relay will misoperate. From t=1.0s to t=1.016s, which is during the second fault,
distance relay sees zone 2 fault; from t=1.626s to t=2.002s, the distance relay still can see zone 3 fault.
If the relay setting time is shorter than the time that the apparent impedance stays inside the zone 2 or
zone 3 circle, the distance relay will trip L 33 according to its algorithm. The system may lose stability
due to relay misoperation.
Fig. 4. Apparent impedance seen by the distance relay at B26 of L33
However, when L34 (B29-28) is tripped, the Equivalent Parallel Path Approximation method can
rapidly find the maximum power flow transferred from L 34 to other lines. The path with minimum
total reactance is (B29-B26-B28). The protection relays for those transmission lines have the
possibility of misoperation. The relay dynamic analysis also validated this result.
V. CONCLUSION
This paper presents an improved transmission line protection scheme based on the interactive
system-wide and local monitoring and analysis tools during cascading events. The means for on-line
automated analysis of protective relay operations is discussed. The static security analysis tools based
on Vulnerability Index and Margin Index identify the vulnerable transmission lines of the system and
vulnerable conditions. Equivalent Parallel Path Approximation based on topology processing
technique locates the possible overloaded transmission lines. With such vulnerability and security
information, local on-line monitoring and control tools are invoked for analyzing performance of local
relay operation and informing the centralized tool about the outcomes. Corrective action to mitigate
incorrect or undesirable relay action will be issued if necessary.
5
BIBLIOGRAPHY
[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 5,
2004, Available: http://www.nerc.com)
[2] A.G. Phadke, J.S. Thorp, Expose Hidden Failures to Prevent Cascading Outage (IEEE
Comput. Appl. Power, vol. 9, no. 3, July 1996, pages 20-23)
[3] NERC System Disturbances Reports, North American Electric Reliability Council (New
Jersey, 1984-1991)
[4] NERC Disturbance Analysis Working Group, Western Interconnection (WSCC)
System Disturbance August 10, 1996 (NERC 1996 System Disturbances Report, Aug
2002, Available: http://www.nerc.com)
[5] H. Song, M. Kezunovic, Static Analysis of Vulnerability and Security Margin of the
Power System (2006 IEEE PES T&D Conference, Dallas, May 2006)
[6] F. Dobraca, M. A. Pai, and P. Sauer, Relay Margins as a Tool for Dynamical Security
Analysis (Int. J. Electr. Power Energy Syst., vol. 12, no. 4, Oct. 1990, pages 226 234)
[7] G. C. Ejebe, B. F. Wollenberg, Automatic Contingency Selection (IEEE Trans. Power
Apparatus and Systems, vol. 98, No. 1, Jan/Feb 1979, pages 97-109)
[8] T. F. Halpin, R. Fischl, R. Fink, Analysis of Automatic Contingency Selection
Algorithms (IEEE Trans. Power Apparatus and Systems, vol. 103, May 1984, pages
938-945)
[9] Z. Bie, X. Wang, Evalution of Power System Cascading Outage (IEEE 2002 Power
System Technology Proceedings, Oct. 2002, vol. 1, pages 415-419)
[10] H. Song, M. Kezunovic, Static Security Analysis based on Vulnerability Index (VI)
and Network Contribution Factor (NCF) Method (IEEE PES T&D 2005 Asia Pacific,
Dalian, China, Aug. 2005)
[11] H. Song, M. Kezunovic, Relieving Overload and Improving Voltage by the Network
Contribution Factor (NCF) Method (2004 proceedings of the 36th Annual North
America Power Symposium, August 2004, pages 128-132)
[12] D. Zhou 2007 M.S. thesis (Dept. Electron. Eng., Xi an Jiaotong Univ., Xi an, China)
[13] M. A. Pai, Energy Function Analysis for Power System Stability (Kluwer Academic
Publishers, Norwell, MA, 1989, pages 222-227)
6
Contact this candidate