Neural Network Control of a Frequency-Regulated Electric Drive of a Main Electric Locomotive

The paper presents the results of the development and investigations of an intelligent control system for an electric drive (ED) of the main-line electric locomotive DS3. It is shown that the use of methods of genetic algorithms for training and structural optimization of neural systems makes it possible to synthesize the control law excluding the self-oscillating process arising from the nonlinearity of the “friction pair” type load. The developed systems have a single easily realizable feedback on the speed of the motor which does not create difficulties in physical realization.