Application of cerebellar model articulation control in reheated-steam system

This paper proposes a cerebellar model articulation control (CMAC) used in reheated-steam system. CMAC is thought as a kind of association memory neural network It can be used to approach any nonlinear junction, dynamical modeling, etc. In this paper neural network was used online to identify the controlled plant and the weights of the adaptive PID controller are adjusted timely based-on the identification of the plant. So the controller has the adaptability and the system can be controlled effectively. The self-adaptive algorithm of the whole control system is proposed This approach is also applied to the reheated temperature plant with long time-delay, large inertia and time-variation in power plant. The simulation experiment demonstrates that the controller performs very well when disturb happens or plant parameter varies. It is of strong robustness and adaptive abilities and can be easily accomplished online.