A Method of neural network internal model control in unstable time-lag process

The phenomenon of unstable time-lag process is usually familiar in the process industry, but it is hard to be controlled by the conventional method. In this paper a control method called double-loop feedback is put forward, first internal feedback stabilization is adopted, then neural network is used to form the internal model control system, finally it solves the problem of bias and instability between the model and the real process. Through the simulation, it is seen that the method has short adjusting time and high control accuracy, which shows the validity and superiority of neural network internal model control.