This thesis describes Kalman filter for estimating the state variables which provide operator with important process status and support controller by supplying decontaminated input states. This paper also describes LQ (Linear Quadratic)-controller for satisfying the desired process performance.
The combined system with Kalman filter and LQ-controller is developed to alternate the classical three-element steam generator water level control system. The important element of this combined system is that a mathematical model is used in conjunction with the actual process data to optimize some performance criteria. The criteria involve important process variables and incorporate a sense of the desired process performance.
Comparative study and discussions are performed among the combined system and Modified PID-controller and Westinghouse controller, respectively.
Compared with the existing Westinghouse controller, this combined system shows more good response and shows considerably stable results on low power level. This combined system can be implemented in real time since LQ-controllerand Kalman filter have almost same algorithm.