estimator (typically an ANN) on simulation data in order to estimate the security margin.

The
estimator is expected to interpolate or generalize to similar unstudied cases. For on-line application,
a pattern matching or interpolation method rather than analytic approaches may be most appropriate.
Among the alternative methods, ANNs seems very promising (Sobajic and Pao, 1989; Pao and Sobajic,
1992; Mansour et al., 1997) because they have excellent generalization capabilities, superior noise
rejection, and fast execution (with most of the calculations occurring during the initial off-line training).
A recent report with survey results (Sauer et al., 2004) quite clearly showed that there is a major gap in
the operations security tools. This gap is the lack of an ability to evaluate stability margins in real time.
This report also included results from a project that focused on this gap and investigated the feasibility of
a new technique for bringing dynamic analysis into the operations environment. The work started with
two of the most time-consuming aspects of stability margin analysis: time-domain simulation and static
voltage margin computations. In a previous Power System Engineering Research Center (PSERC)
project (Tomsovic et al., 2003), it was shown that a system of estimators based on neural networks
could accurately and quickly estimate security margins for on-line application. This project produced a
number of contributions to the development of dynamic security analysis techniques.
. A comprehensive framework was developed for on-line estimation of security margins, calculated
based on current operating practices.
. The framework proposed families of estimators, each specialized for specific system limits and the
appropriate security criteria (i.e., static, dynamic, or voltage). The estimators can be combined to
provide an overall assessment of system operating conditions.
. A system of estimators was implemented and tested on a modified New England 39 bus system.
. On the basis of the insights from the New England system, a more sophisticated set of estimators
were implemented and tested on a 6000 bus model of the Western area system. The focus of this
study was the California–Oregon Intertie transfer limits.
. A number of software tools were developed to help automate the process of evaluating security
margins in off-line studies.
. The results show that it is possible to very accurately estimate security margins for large systems
on-line. The main limitation of the approach resides in the ability of time-consuming off-line
studies to accurately model system dynamics.

15.10 Off-Line DSA

In off-line DSA analysis, detailed time-domain stability analysis is performed for all credible contin-
gencies and a variety of operating conditions. In most cases, this off-line analysis is used to determine
limits of power transfers across important system interfaces. These limits then are used in an operating
environment that is hopefully not significantly different from those conditions considered. Since the
analysis is performed off-line, there is not a severe restriction on computation time and therefore
detailed analysis can be done for a wide range of conditions and contingencies. These studies include
numerical integration of the models discussed above for a certain proposed power transfer condition
and for a list of contingencies typically defined by a faulted location and specified fault-clearing time
(based on known relay settings).
The trajectories of the simulation are analyzed to see if voltage transients are acceptable, and to see if
transient stability is maintained for the specified fault-clearing time. If the results for one level of power
transfer are acceptable for all credible contingencies, the level of proposed power transfer is increased
and the analysis is repeated. This process continues until the level of power transfer reaches a point
where the system cannot survive all of the credible contingencies. The maximum allowable transfer level
is then fixed at the last acceptable level, or reduced by some small amount to provide a margin that
would account for changes in conditions when the actual limit is in force.

2006 by Taylor & Francis Group, LLC.

15.11 On-Line DSA
On-line DSA is used to supplement (or update) off-line DSA to consider current operating conditions. A
basic on-line DSA framework includes essentially two steps. The first involves a rapid screening process
to limit the number of contingencies that must be evaluated in detail. This rapid screening process might
consist of some direct method that avoids long numerical integration times (Pai, 1989; Fouad and Vittal,
1992; Pavella and Murthy, 1993; Chadalavada et al., 1997). In addition to giving fast stability evaluation,
these methods inherently include a mechanism for assessing the severity of a contingency. That is, if a
system is determined to be stable, the direct methods also provide an indication of ‘‘how stable’’ the
system is. This indication usually takes the form of an ‘‘energy margin.’’ For example, in reference to the
ball motion of Fig. 15.2, the maximum swing of the ball up the side of the energy well could be used to
quantify how ‘‘close’’ the ball was to leaving the well.
Most of these methods still require some numerical integration to simulate the impact of a major
disturbance and then predict stability or compute a margin to instability. Computation of the margin
usually requires the simulation to force the system into an instability, perhaps either by using a sustained
fault, or reapplication of the fault after scheduled clearing (Vaahedi et al., 1996).
This first step includes a decision process of which contingencies must be studied in greater detail.
Those that are judged to be ‘‘sufficiently stable’’ need not be studied further. Those that are considered
‘‘marginal’’ must be studied further. This process includes a ranking strategy that is usually based on the
energy margin computed in the direct method. Additional criteria involving artificial intelligence
approaches can also be used to aid the decision process (El-Kady et al., 1990).
The second step involves traditional time-domain simulation that includes extensive numerical
integration to reveal swing trajectories and voltage variations. This is performed on a small subset of
contingencies that were judged to be marginal according to the screening process of step one.
In on-line studies, the time for computation is a severe constraint in addition to the challenge of
interpretation and quantification of results. Typical performance goals for on-line DSA program are to
process 30 contingencies (each having 10 s of simulated time) for a 2000 bus, 250 generator system in
10 min (Ejebe, 1998).

15.12 Status and Summary

A recent survey of existing DSA tools provides a detailed description of the status of DSA tools (Vittal
et al., 2005). With the increase in transactions on the bulk power system there is a critical need to
determine transient security in an on-line setting and also perform preventive or corrective control if
the analysis indicates that the system is insecure. In recent years, the industry has seen the develop-
ment of large generation projects at concentrated areas of available fuel supplies. The stability
properties of the system have been drastically altered, while the new ‘‘nonutility’’ plants are not
cognizant of their impact on system stability. In this environment, stability issues may and will affect
available transfer capability. Stability problems may not happen frequently, but their impact, when
they do happen, can be enormous. Most of the time, off-line studies are performed to determine
conservative limits. In the new environment, the responsibility of monitoring system stability may
be vested with the regional transmission organization (RTO) and on-line stability monitoring
may be necessary.
This section deals with reviewing the current state-of-the-art in the area of on-line transient stability
assessment, evaluating promising new technologies, and identifying technical and computational
requirements for calculating transient stability limits and corrective and preventive control strategies
for cases that are transiently insecure.
Six on-line transient stability package vendors were identified by conducting a literature survey. A
detailed questionnaire that addressed several pertinent issues relating to on-line transient stability
assessment was prepared. All six vendors responded to the questionnaire. The responses received were

2006 by Taylor & Francis Group, LLC.

carefully analyzed. This analysis provided a detailed overview of the capabilities of available tools,
performance metrics, modeling features, and protective and corrective control measures.
An elaborate questionnaire was then prepared and sent to all PSERC member companies. This
questionnaire addressed specific needs in terms of required features, preferred performance, and control
capabilities. A detailed analysis of the received responses provided a clear picture of the desired features
and performance specifications of an on-line transient stability assessment tool.
A comparison of the analysis conducted on the vendor responses and the PSERC member company
responses identified areas and topics that needed further development and research. This information
will be useful in soliciting new research proposals and providing vendors a guide to the features that
need to be developed and implemented.
A literature survey was also conducted on new analytical developments in on-line transient stability
analysis. On the basis of this review, novel concepts based on quadratized models for power system
components were explored to investigate whether there would be a significant advantage gained in using
these models in terms of accuracy and computational burden.
DSA is concerned with the ability of an electric power system to survive a major disturbance. It must
assess the quality of the transient behavior as well as stability. DSA is performed both in off-line and on-
line environments and is computationally intensive due to the numerical integration involved in
evaluating the transient behavior of the system during major disturbances. Several recent and ongoing
projects have addressed the computational issue through screening techniques that provide rapid
analysis of stability outcomes and stability margins (Demaree et al., 1994; Meyer et al., 1997). Research
in this area is continuing as the need for DSA to evaluate available transfer capability (ATC) becomes
stronger in the restructured industry.

References
Anderson, P.M. and Bose, A., A probabilistic approach to power system stability analysis, IEEE
Transactions on Power Apparatus and Systems, PAS-102(8), 2430–2439, August 1983.
Anderson, P.M. and Fouad, A.A., Power System Control and Stability, Iowa State University Press, Ames,
IA, 1977, Reprinted by IEEE Press, 1993.
Chadalavada, V., Vittal, V., Ejebe, G.C., Irisarri, G.D., Tong, J., Pieper, G., and McMullen, M., An on-line
contingency filtering scheme for dyanmic security assessment, IEEE Transactions on Power
Systems, 12(1), 153–159, February 1997.
Chauhan, S. and Dava, M.P., Kohonen neural network classifier for voltage collapse margin estimation,
Electric Machines and Power Systems, 25(6), 607–619, July 1997.
Chen, L., Tomsovic, K., Bose, A., and Stuart, R., Estimating reactive margin for determining trans-
fer limits, Proceedings, 2000 IEEE Power Engineering Society Summer Meeting, 2000. IEEE,
Volume 1, July 2000.
Cihlar, T.C., Wear, J.H., Ewart, D.N., and Kirchmayer, L.K., Electric utility system security, Proceedings of
the American Power Conference, 31, 891–908, 1969.
Demaree, K., Athay, T., Chung, K., Mansour, Y., Vaahedi, E., Chang, A.Y., Corns, B.R., and Garett, B.W.,
An on-line dynamic security analysis system implementation, IEEE Transactions on Power Systems,
9(4), 1716–1722, November 1994.
Dommel H.W. and Sato, N., Fast transient stability solutions, IEEE Transactions on Power Apparatus and
Systems, 91, 1643–1650, July=August 1972.
Dy Liacco, T.E., The adaptive reliability control system, IEEE Transactions on Power Apparatus and
Systems, PAS-86(5), 517–531, May 1967.
Ejebe, G.G., On-line dynamic security assessment, Slides presented to the 1998 IEEE PES Working
Group on Dynamic Security Assessment Meeting, February 1998, Tampa, Florida (available from
the author).
El-Kady, M.A., Fouad, A.A., Liu, C.C., and Venkataraman, S., Use of expert systems in dynamic security
assessment of power systems, Proceedings 10th PSCC, Graz, Austria, 1990.

2006 by Taylor & Francis Group, LLC.

El-Keib, A.A. and Ma, X., Application of artificial neural networks in voltage stability assessment, IEEE
Transactions on Power System, 10(4), 1890–1896, November 1995.
Fouad, A.A. (Chairman. IEEE PES Working Group on DSA), Dynamic security assessment practices in
north america, IEEE Transactions on Power Systems, 3(3), 1310–1321, August 1988.
Fouad, A.A. and Vittal, V., Power System Transient Stability Analysis Using the Transient Energy Function
Method, Prentice-Hall, Englewood Cliffs, NJ, 1992.
Hakim, H. Application of pattern recognition in transient security assessment, Journal of Electrical
Machines and Power Systems, 20, 1–15, 1992.
IEEE Committee Report, Dynamic security assessment practices in North America, IEEE Transactions on
Power Systems, 3(3), 1310–1321, August 1988.
Kundur, P., Power System Stability and Control, McGraw Hill, New York, 1994.
Mansour, Y., Chang, A.Y., Tamby, J., Vaahedi, E., Corns, B.R., and El-Sharkawi, M.A., Large scale
dynamic security screening and ranking using neuron networks, IEEE Transactions on Power
Systems, 12(2), 954–960, May 1997.
Meyer, B. (Convener) and Nativel, G. (Secretary), New Trends and Requirements for Dynamic
Security Assessement, Cigre Task Force 38.02.13 Report, 21 rue d’Artois-R-75008 Paris, France,
December 1997.
Pai, M.A., Energy Function Analysis for Power System Stability, Kluwer, Boston, MA, 1989.
Pao, Y., and Sobajic, D.J., Combined use of unsupervised and supervised learning for dynamic security
assessment, IEEE Transactions on Power Systems, 7(2), 878–884, February 1992.
Patton, A.D. Assessment of the security of operating electric power system using probability methods,
Proceedings of IEEE, 62(7), July 1974.
Pavella, M. and Murthy, P.G., Transient Stability of Power Sytems: Theory and Practice, John Wiley,
New York, 1993.
Ribbens-Pavella M. and Evans, F.J., Direct methods for studying dynamics of large-scale electric power
systems—A survey, Automatica, 21(1), 1–21, 1985.
Sauer, P.W. and Pai, M.A., Power System Dynamics and Stability, Prentice-Hall, Upper Saddle River, NJ,
1998.
Sauer, P.W., Tomsovic, K., Dagle, J., Widergren, S., Nguyen, T., and Schienbein, L., Integrated Security
Analysis Final Report, CERTS Publication, July 2004.
Sobajic, D. and Pao, Y., Artificial neural net based dynamic security assessment for electric power
systems, IEEE Transactions on Power Systems, 4(1), 220–228, February 1989.
Taylor, C.W., Power System Voltage Stability, McGraw-Hill, New York, 1994.
Tomsovic, K., Bose, A., and Sauer, P., Integrated Security Analysis Final Report, PSERC Publication
03-06, Power Systems Engineering Research Center, May 2003.
Vaahedi, E., Mansour, Y., Chang, A.Y., Corns, B.R., and Tse, E.K., Enhanced ‘‘second kick’’ methods for
on-line dynamic security assessment, IEEE Transactions on Power Systems, 11(4), 1976–1982,
November 1996.
Vittal, V., Sauer, P., and Meliopoulos, S., On-Line Transient Stability Assessment Scoping Study Final Project
Report, PSERC Publication 05-04, Power Systems Engineering Research Center, February 2005.
Wehenkel, L., Automatic Learning Techniques in Power Systems, Kluwer Academic Publishers, Boston,
MA, 1998.

2006 by Taylor & Francis Group, LLC.