A modified neural network controller configuration

In this paper two types of neural networks for generator voltage control are presented: a neural network with a single set of weights and another one with two sets. The selection of the appropriate set is depending on the nature of generator load: inductive or capacitive. Moreover, it has been demonstrated that the range of the values for the weights has a significant importance for the stability margin and voltage control performance. The performances of the proposed neural controller are compared with that of conventional controller.