Robust and adaptive tuning of power system stabilizers using artificial neural networks

Abstract

Tuning of power system stabilizers (PSS) over a wide range of operating conditions and load models is investigated using an artificial neural network (ANN). The neural nettwork is specially trained by an input-output set prepared by a novel approach based on genetic algorithms (GA). To enhance power system damping, it is desirable to adapt the PSS parameters in real-time based on generator operating conditions and load models. To do this, on-line measurements of generator loading conditions are chosen as the input signals to the neural network. The output of the neural network is the desired gain of the PSS that ensures the stabilization of the system for a wide range of load models connected to the power system. For training the neural network a set of operating conditions is chosen as the input. The desired output for any input is computed by simultaneous stabilization of the system over a wide range of load models using genetic algorithm. In this regard, the power system operating at a specified operating condition and various load models is treated as a finite set of plants. The problem of selecting the output parameters for every operating point which simultaneously stabilize this set of plants is converted to a simple optimization problem which is solved by a genetic algorithm and an eigenvalue-based objective function. The proposed method is applied to a test system and the validity is demonstrated through digital simulation.