Nonlinear parameter neuro-estimation for optimal tuning of power system stabilizers

This paper describes nonlinear parameter estimation of non-smooth nonlinear device by using a feed-forward neural network (FFNN) embedded in a hybrid system modeling. The hybrid systems are modeled by the differential-algebraic-impulsive-switched (DAIS) structure. In a switched linear hybrid system, the FFNN is applied to identify full dynamics of an objective function J formed by the states. Moreover, the partial derivatives of function J with respect to the each state are approximated by the computation of the backpropagation through the FFNN. Then, this paper focuses on the FFNN based estimator for the non-smooth nonlinear dynamic behaviors due to saturation limiter of the power system stabilizer (PSS) in both a single machine infinite bus (SMIB) system and a multi-machine power system (MMPS).