Artificial Intelligence: Summary and Hybrid Schemes - Digital Signal Processing in Power System Protection and Control - page 310

The cascade structure similar to the one from Fig. 15.2 has also been proposed
in [ 1 ] for dynamic security assessment of a power network. Signals from the
measurement as well as the ones obtained from state estimation and time-domain
simulations of selected contingencies are first processed by a neural network. The
performance indices delivered at the ANN output go further through a fuzzy
scheme, which finally outputs the predicted system state.
Another neuro-fuzzy scheme presented in [ 22 ] was proposed with the aim to
improve transmission line fault detection and classification. Here, the ART
neural network is applied to set the prototypes of trained clusters that are
further used in the process of fault detection and classification. The inputs of
the scheme are the data from power system simulation, data from substation
historical database as well as signal samples from real system delivered by
CVTs and CTs. The authors claim that the hybrid system developed is much
more

efficient that the other ones,also including single-technique protection

versions.
The other examples of hybrid schemes as applied for power system problems
include:

• Optimization of the generation companies' bidding strategy (fuzzy simulation
and neural network combined with GA) [ 3 ],
• New solution for economic power dispatch (genetic algorithm combined with
interior point methods harmony search) [ 14 ],
• Power quality analysis in distribution networks with a fuzzy-expert system
(combining the usefulness of fuzzy logic in interpreting the fuzzy inputs and the
expert system shell) [ 7 ],
• Power system fault diagnosis with fuzzy-expert system (COFES—incorporating
fuzzy symbol classification through an enhanced knowledge-base system) [ 12 ],
• Peak load forecasting with use of a fuzzy expert system (incorporating linguistic
fuzzy IF-THEN rules and expert's opinions) [ 5 ],
• On-line fault diagnosis on a transmission network with use of a fuzzy-expert
system (information received includes the open/closed states of circuit breakers
and the operational response of protection relays in conjunction with the
topology of the transmission network) [ 18 ],
• Alarm interpretation and fault diagnosis (generic neuro-expert system archi-
tecture that can overcome difficulties faced by stand-alone ES and ANN
schemes) [ 6 ],
• Transmission expansion planning using neuro-GA hybrid (neuro-computing is
hybridized with genetic algorithms) [ 21 ],
• Wind-solar generation power control (a genetic-based self-adaptive hierarchical
fuzzy controller is developed) [ 20 ],
• Digital power metering (fuzzy-based adaptive approach employing a genetic
algorithm) [ 9 ],
• Fuzzy power system stabilizer design (optimization with genetic procedure) [ 11 ],
• Monitoring and diagnostics of large power transformers (an adaptive neuro-
fuzzy system identification is applied) [ 17 ].