Identification of a turbogenerator system using adaptive time delay neural networks

This paper deals with identification of a multi-input multi-output (MIMO) nonlinear turbogenerator system by using dynamic neural networks. Using an adaptive time delay neural networks, a neuro-dynamic identifier is proposed. The capability of the proposed structure for representing the nonlinear input-output map of the turbogenerator system is shown analytically. The input-output data set is generated by PSCAD/EMTDC package for a turbogenerator system with default parameters. Simulation results reveal that the proposed neuro-dynamic structure is capable of successfully identifying a highly nonlinear turbogenerator system without any a priori information about the nonlinearities of the system and without any off-line training