Prony-RBFNN Approach for Tuning Power System Stabilizer

This paper presents a combined approach based on Prony analysis and radial basis function neural network for monitoring small signal stability and parameter tuning of power system stabilizer. Prony analysis method is used to estimate the modal components of low frequency oscillations associated with synchronous generators. In this method, the measured (simulated) time-domain signal is decomposed into damped sinusoids with four parameters per mode: amplitude, frequency, damping and phase angle. Once the local mode responsible for poor damping of the low frequency oscillations is identified, its damping factor and damped frequency are used to predict the parameters of the stabilizer using a radial basis function neural network. The tests results show that Prony analysis-neural network technique can be effectively applied in small signal stability analysis and power system stabilizer design. Key-Words:Power system stability and control, identification, Prony analysis, RBF, simulation