For the synchronous generator in the power plant, vessel and building emergency power source, the constant voltage control to load disturbance is basically the excitation control. The excitation controller must have capabilities to acquire the required transient response that able to overcome the inductive time lag of excitor and to damp out the inherent instability of the generator and prime mover. Many controllers such as PID, optimal control and self-tuning regulator have been applied and PSS(Power System Stabilizer) to promote stability was added in those controllers.
However, these methods linearized the non-linear model in the limited operating boundary, and/or estimated the dynamic characteristics of the plant. But these approaches require the powerful computing capability and the added feedback loops.
In this thesis, the dead-time compensator (Smith Predictor) is proposed to regulate the plant having long time lagging component. It is necessary for the compensator design to know the plant model accurately. But the variation of plant parameter and operating point can deteriorate the dynamic response. This problem is improved by dynamic tuning of compensator parameter, which is adjusted by fuzzy logic. The proposed controller is compared with PID controller and shows the effectiveness of the fuzzy-tuned compensator in the real experiment.
