Real-time optimal excitation controller using neural network

A neural network based optimal excitation controller (NNOEC) is proposed in this paper. In this NNOEC, a BP neural network is used to adjust the optimal feedback gains according to the state variables of the generator. So the controller can automatically adapt the changed operating conditions of the system and always give optimal control. Simulations with the NNOEC and LOEC in single machine system and simulations with the NNOEC and AVR+PSS in three-machine system are conducted, where the simulations for the single machine system are carried out based on the Three Gorges 700 MW hydropower generator. Simulation results show that the designed NNOEC can provide good control performance under various operating points and different disturbances.