Real power transfer capability calculations using multi-layerfeed-forward neural networks

摘要

This paper proposes a neural network solution methodology for thenproblem of real power transfer capability calculations. Based on thenoptimal power flow formulation of the problem, the inputs, for thenneural network are generator status, line status and load status and thenoutput is the transfer capability. The Quickprop algorithm is used innthe paper to train the neural network. A case study of the IEEE 30-busnsystem is presented demonstrating the feasibility of this approach. Thennew method will be useful for reliability assessment in the new utilitynenvironment