Application of neural network prediction in process control

In process control, there exists obvious time-delay phenomenon, which brings phase-delay to the system and reduces the stability of the system. There also exists unknown time-varying state which has no mathematical model in the system. PID controllers can't control delay and nonlinear state effectively, although they are applied to the control system design widely. In this paper, neural network characteristic of adaptive learning was used to build accurate model and compensate for time-delay and nonlinearity of the system. Forecast based on the feedback control algorithm was made to solve the problems about time delay and nonlinearity of the water tank level control. Accurate control was achieved.