Experimental studies with a neural network eased power system stabilizer

Employing an inverse input/output mapped artificial neural network (ANN) as a controller, an ANN based power system stabilizer (PSS) has been implemented and tested, in the laboratory environment. Experimental test results are presented in this paper. The ANN is trained off-line using the data generated by an adaptive PSS controlling the generating unit under typical disturbances. Test results show that the proposed ANN PSS exhibits very good performance in damping power system low frequency oscillations and greatly improves power system stability