DCNN Based Saturation Model for Round Rotor Synchronous Generator

This paper presents a dynamic compensation neural network (DCNN) based technique to model saturation for a round rotor synchronous generator. The DCNN is based on the principle of dynamic error back-propagation to make the feedback compensation, which is benefit to reduce dynamic error and raise the modeling accuracy significantly. The structure of the generator model is introduced firstly. Then the structure and learning algorithm of this novel neural network model are given. The developed model is validated with study case not used in the training process and with large disturbance responses.