Available transfer capability determination using artificial neural network

• Main Author: Habibuddin, Mohd. Hafiz
• Format: Thesis
• Language: English
• Published: 2003
• Subjects: TK Electrical engineering. Electronics Nuclear engineering
• Online Access: http://eprints.utm.my/id/eprint/42613/1/MohdHafizHabibuddinFKE2003.pdf; http://eprints.utm.my/id/eprint/42613/; http://libraryopac.utm.my/client/en_AU/main/search/detailnonmodal/ent:$002f$002fSD_ILS$002f0$002fSD_ILS:334298/one?qu=Available+transfer+capability+determination+using+artificial+neural+network

Open access to the transmission systems places a new emphasis on the more intensive shared use of the interconnected networks reliably by utilities and independent power producers. Therefore, as a measure of the network capability for further commercial activity over and above already committed uses, the concept of available transfer capability (ATC) was proposed and defined by the Federal Energy Regulatory Commission (FERC) in 1995. This study proposes the use of an Artificial Neural Networks (ANN) to determine ATC in an interconnected power system. The ANN is a multilayer feedforward network employing LevenbergMarquardt training algorithm. Newton-Raphson load flow solution incorporating Continuation Power Flow (CPF) method was used to gather the training and test data. The inputs to the ANN are the load level and line flow in the power system. Only thermal limits are taken into consideration. The method was tested with 4 buses system and TNB Southern Region 25 buses system. Comparison with CPF method shows that the ANN is a feasible alternative method to determine ATC.