Application ofArtificial NeuralNetworks in Determining Critical ClearingTime inTransient Stability Studies
G. V. Rao · 2008
Thispaperdescribes a neuralnetworkbased adaptive pattern recognition approach bymakinga thorough analysis onapowersystem forestimation ofthecritical clearing time.A ninebussystemisconsidered forthepurposeof transient stability analysis. Faults atfive locations areassumed atdifferent instants. Critical clearing times forallfive faults at sixdifferent loading levels areobtained. Outofthirty cases, 24 casescorresponding tofourfaults havebeenusedfortraining theNeuralNetworkandremaining sixCCTscorresponding to thefifth fault atsixloading levels obtained byANN aswellas Modified Eularmethod. Thesameisrepeated forallfive faults. NueralNetworkdesigned with12inputneurons, 8 hidden neurons andoneoutput neuron. Backpropagation technique is usedtoadjust theweights. Analytical calculations arecompared withthevalues obtained byNeural Network.Results showthat ANN gives accurate results. IndexTerms- Backpropagation, Critical clearing time, Neural Network, Transient Stability.