Neuro-fuzzy power system stabiliser

A fuzzy logic power system stabiliser has been developed using speed and active power deviations as the controller input variables. The inference mechanism of the fuzzy logic controller is represented by a 7 × 7 decision table, i.e. 49 if-then rules. In order to use it under a wide range of operating conditions, its parameters have been tuned using a neural network. The tuned stabiliser has been tested by performing non-linear simulations using a synchronous machine-infinite bus model. It is shown that the neuro-fuzzy stabiliser is superior to a fixed parameter fuzzy logic power system stabiliser.