![](bg7.png)
signal processing techniques are implemented
supplying input data to the Fuzzy-Neuro Expert
System that classify a possible fault.
The directional relay shows to be a robust
method to provide an indication of the fault
direction, beyond supplying angle information that is
used as input to the ES with satisfactory results.
The developed ATP/EMTP model allows the
simulation of diverse disturbances inside the
substation, which was used for compose Fuzzy Sets,
training the ANN and test the hybrid Fuzzy-Neuro
Expert System.
The Fuzzy-Neuro Expert System classifies the
fault in two levels of details. The Fuzzy System is
more generalist and identifies only the local of fault,
whereas, the ANN is qualified to indicate the related
circuit breaker. For this reason, the fact that the net
is very specialist, a level of classification error can
occur. In some ANN tests, the error is allied with the
wrong of circuit breaker and not with the local of
fault as bus, transformer, line or generators. It is
important to highlight that the ES input data are very
close and the classification process is not a trivial
task.
Several simulations of different values of epochs
was performed in network training process and the
best results were attainment with 150 epochs
converging to a MSE = 0.0057, ANN global test
error = 11.8% considering 6, 50 and 17 perceptrons
in the input, hidden and output layer respectively.
With this configuration the best results was obtained
and this structure was implemented in the fault
detection expert system.
The developed integrated system can become the
management maintenance activities more efficient.
Moreover, such system contributes for the increase
of the reliability, having as one of its benefits, the
reduction of involved time to detect and localize a
possible fault, optimizing the maintenance practices.
Currently, the system is being tested in the TGP
in the Southern Brazilian, but this methodology can
be used to detect and localize faults in similar energy
electric substations.
ACKNOWLEDGEMENTS
The authors wish to thank the American Energy
System (AES Uruguaiana) for the financial support
and Federal University of Rio Grande do Sul for the
facilities offered.
REFERENCES
Bonneville Power Administration - BPA “Alternative
Transients Program: ATP/EMTP”. Avalilable in:
http://www.emtp.org/.
Chen, W. H., Liu, C. W., Tsai M.S., 2000. “On-Line Fault
Diagnosis of Distribution Substation Using Hybrid
Cause-Effect Network and Fuzzy Rule-Based
Method,” IEEE Transaction on Power Delivery, vol.
15, no. 2.
Ebron, S., Lubkeman, D. L., White, M., 1990. “A neural
network approach to the detection of incipient faults
on power distribution feeders,” IEEE Trans. On
PWRD, vol. 5, no. 2, pp. 905–913.
Fukui, C., Kawakami, J., 1986. “An expert system for
fault section estimation using information from
protective relays and circuit breakers,” IEEE Trans.
On PWRD, vol. 1, no. 4, pp. 83–90.
Haykin, S., 1998. “Neural Networks: A Comprehensive
Foundation,” New York: Prentice Hall, 2nd ed.
IEEE Standart Common Format for Transient Data
Exchange (COMTRADE) for Power Systems, IEEE
Standard C37.11.1.1999, 1999.
Kezunovic, M., Rikalo; I., Fromen, C. W., Sevcik, D. R.,
1994. “Expert System Reasoning Streamlines
Disturbance Analysis,” IEEE Computer Application in
Power, vol. 7, pp. 15-19.
Kezunovic, M., 2004. “Intelligent applications in
substations: Disturbance analysis,” In IEEE PES
General Meeting, vol. 1, pp. 719-723, Denver, USA.
Phadke, A. G., Thorp, J.S., 1999. “Computer Relaying for
Power Systems,” England: Research Studies Press
Ltd.
Protopapas, C. A., Psaltiras, K. P., Machias, A. V., 1991.
“An expert system for substation fault diagnosis and
alarm processing,” IEEE Trans. On PWRD, vol. 6, no.
2, pp. 648–655.
Ranaweera, D.K., 1994. “Comparison of neural network
models for fault diagnosis of power systems,” Electric
Power Systems Research, pp. 99–104.
Suonan J. L., Xu Q. Q., Song G. B., Wu Y., 2004. “A
reliable directional relay based on positive sequence
compensated voltage and current components,” The
Institute of Electrical Engineers, Six Hills Way, pp.
104-107.
The Mathworks Inc. “Mathworks MatLab”. Available in:
http://www.mathworks.com/.
Tomsovic, K., Liu, C. C., Ackerman, P., Pope, S., 1987.
“An expert system as a dispatchers’ aid for the
isolation of line section faults,” IEEE Trans. On
PWRD, vol. 2, no. 3, pp. 736–742.
Yang, H. T., Chang, W. Y., Huang, C. L., 1994. “A new
neural networks approach to on-line fault section
estimation using information of protective relays and
circuit breakers,” IEEE Trans. On PWRD, vol. 9, no.
1, pp. 220–230.
Zadeh, L. A., 1965. "Fuzzy sets," Information and Control,
vol. 8, pp. 338-353.
A HYBRID EXPERT SYSTEM BASED ON NEURAL NETWORKS AND FUZZY LOGIC FOR FAULT
IDENTIFICATION IN ELECTRIC POWER SUBSTATIONS
441