Synchronous generator model identification using half-complex wavelet nonlinear ARX network

A method for identification of a synchronous generator is suggested in this paper. The method uses the theoretical relations of machine parameters and the new wavelet network structure combining a polynomial model with multi-resolution half-complex wavelet decomposition to find the state space model of the system. The proposed structure takes the advantage of both polynomial models and wavelet networks to model nonlinearities in the system. The proposed identification method is applied to a mathematic model of a 12 phase synchronous generator from experimental data collected on a PSCAD / MATLAB simulating machine. The results show that the identified model has very good accuracy and may be valid for a broad range of operating conditions.