Artificial neural network based modeling of governor-turbine system

This paper presents an artificial neural network based modeling of the governor-turbine system using measured governor test data. The study unit is a LNG fueled thermal unit utilized for the load-frequency regulation. The proposed model consists of three blocks. The first is the governor block which is modeled by using a conventional model including time-lag and dead-band. The second is the steam valve servo system, and the last is the turbine system including generator. Both the second and the third are modeled by using artificial neural networks in this paper. By using the proposed model, the dynamics of the governor-turbine system are modeled quite accurately. In addition, comparison studies have also been performed between the proposed and the conventional models