One of the important requirements for operational planning of electrical utilities and also transactions of electrical power markets is the prediction of hourly load up to several days, known as short term load forecasting (STLF). Nowadays, intelligent methods, specially, artificial neural network (ANN) is the dominant method when it comes to STLF. The Levenberg-Marquardt (LM) algorithm
has been extensively used as training method for ANNs. In this work, a new momentum adjustment technique is implemented for
training ANN of Iran national power system (INPS) STLF. The performance is compared with conventional LM algorithm with other
momentum adjustment techniques. The new method of momentum adjustment for LM algorithm improves learning of ANN for STLF of INPS in the sense of error and time consumption.
Department of Power System Operation
Niroo Research Institute (NRI)