Decentralized Online Neuro-Identification of Turbogenerators in a Multi-Machine Power System

This paper proposes a new and a novel technique based on Artificial Neural Networks (ANNs) for nonlinear model of turbogenerators in a multi-machine power system. Only local measurements are required by each ANN in this new method, and hence it is called decentralized neuro-identificiation. Each turbogenerator in the power system is quipped with an ANN which is able to identify (or model) its particular turbogenerator from moment to moment This information can then be used by a second ANN at each generator to enable effective control of the nonlinear non-stationary process under all operating conditions. Simulation results are presented in this paper to show the potential of this new technique for designing fkture nonlinear ANN controllers