Application Research On The Planning Of Feedforward Trajectory Through The Combination Of System Idendification And MPC

Because of the complicated characteristics, nonlinearity and time delay, of the single crytal furnace control system, we propose a simple and effective method. In silicon industrial systems, design a feed forward temperature trajectory, in order to realize the precise control of pulling speed. In the proposed method, the model predictive control strategy and system identification are used to incorporate the complex dynamic effect of the heat temperature on the pulling speed into the temperature trajectory design quantitatively. We will make a difference between the actual value and the set value of the pulling speed and temperature, respectively. And then linearized the results as the input of system identification to identify a dynamic matrix A.Then use the dynamic matrix A to design a MPC controller for the temperature control. Finally, the temperature trajectory will be used to control the pulling speed.In order to verify the feasibility and effectiveness and practical applicability of this method, the MATLAB/Simulink simulation study is carried out. The results show that it can effectively compensate the inaccuracy of temperature trajectory in CZ control system. The trimmed temperature trajectory matches plant dynamics quickly due to the prediction functionality in the MPC controller. The feedforward trajectories designed by the proposed method were implemented on 200 mm and 300 mm silicon crystal growers in the commercial Czochralshi process.