A Multi-Objective Power System Stabilizer

This paper proposes an integrated controller to regulate the terminal voltage of generators as well as to mitigate power system oscillations, instead of employing the combination of the conventional power system stabilizer and automatic voltage regulator. This intelligent controller is an online trained self-recurrent wavelet neural network controller (OTSRWNNC). To achieve the aforementioned objectives, two control errors are simultaneously minimized by updating the parameters of OTSRWNNC. In addition, the adaptive learning rates derived by the discrete Lyapunov theory are used to enhance the convergence speed of proposed controller. The proposed controller does not require any identifier to approximate the dynamic of controlled power system, because of its high learning ability. The performance of proposed controller is evaluated on a single-machine infinite-bus power system and two large power systems. Simulation results and comparative studies demonstrate the effectiveness and robustness of proposed controller in stabilizing power systems in a wide range of loading conditions and different disturbances.