Identification of synchronous generators using adaptive wavelet networks

Application of wavelet networks for the identification of a synchronous generator is described in this paper. Parameter adaptation laws are used to track the variations in the parameters, following changes in the generator operating conditions. The adaptation laws have been developed using a Lyapunov function. This guarantees the stability of the identification algorithm and also ensures the convergence of parameters and variables. The proposed method has been tested on a synchronous machine. Experimental results show good accuracy of the identified model and robustness of the algorithm following severe changes in the operating conditions.