Abstract

Abstract:This paper analyzes and tests a proposed neural network (NN) to solve the DynamicEconomic Dispatch (DED) as part of the unit commitment problem. The proposed NN is afast and direct computation solver using a Hopfield model for solving the dynamic economicdispatch problem of thermal generators which is a dynamic optimization problem taking intoaccount the constraints imposed on system operation by generator ramping rate limits.Formulations for solving the ED and DED problems are explored. Through the application ofthese formulations, direct computation instead of iterations for solving the problem becomespossible. Not like the usual Hopfield network, which select the weighting factors of theenergy function by trials, the proposed network determines the corresponding factors bycalculations and employs a linear input-output model for the neurons. The effectiveness of thedeveloped neural network is identified through its application to the New England test system.Computational results manifest that the model has a lot of excellent performances.

The article was added to IASJ on 2013-02-04

389 Total full text downloads since the date of addition