IDENTIFICATION OF SYNCHRONOUS GENERATORS USING “4SID” IDENTIFICATION METHOD AND NEURAL NETWOKS

Synchronous generators are multivariable systems and are well defined in a state space structure. Subspace state space (4SID) identification method is most suitable for identification of such systems. In this paper, the identification of a synchronous generator using this identification method is presented. To cover the nonlinearities, first the overall range of operation in the active (P) and reactive (Q) power plane is divided into subsections and a linear state space model is identified for the centre of each subsection. A neural network is then used to estimate the parameters for the operating conditions other than the centres of the subsections.