Font Size: a A A

Czochralski Crystal Growth Furnace System Of Model Predictive Control Method

Posted on:2011-10-14Degree:MasterType:Thesis
Country:ChinaCandidate:B T YanFull Text:PDF
GTID:2121360305970537Subject:Control theory and control engineering
Abstract/Summary:PDF Full Text Request
The main purpose of Crystal Growth is to grow out of uniform shape and internal quality meeting the requirements of the single crystal. In the Czochralski crystal growth process, the main method is regulating the pulling speed and heater temperature to track the target curves, in order to achieve the diameter growth of crystal. In this paper, it has been derived the dynamic characteristics theory of crystal growth.to obtain the crystal radius response equation.Combined with experimental experience,I have concluded that the pulling speed control diameter system is approximately an integrating plus time delay system. Then I have obtained system parameters through the design of identification experiments,which can be used to establish the parameter model of the pulling speed control diameter system.By studying the structure and principle of the traditional single-crystal furnace control, we know that the single-crystal furnace control system is essentially the system of tracking targets experience curve, combined with heater cascade control structure and feedforward control techniques to achieve the growth of crystal diameter. But the quality of the performance tracking target curve significantly affect the crystal growth process to attain a constant diameter. In this paper, the design has been made to obtain open-loop step response of the non-parametric model in the crystal furnace growth system, and has been applied to the dynamic matrix control (DMC) method to simulate tracking the given target curve, the performance are more better than conventional PID control. Considered that the dynamic matrix control using the step-response model as a predictive model has a shortcoming, which is the model can not be changed after obtained. Using the generalized predictive control (GPC)to simulate the long-time running Czochralski crystal growth process to test the ability of tracking the given curve. Simulation results show that the application of DMC and GPC in terms of anti-jamming capability, or in tracking the slope curve on behalf of pulling speed orders ability.are better than the tradition PID control.
Keywords/Search Tags:Czochralski crystallization process, feedforward trajectory design, parameter model, dynamic matrix control, generalized predictive control
PDF Full Text Request
Related items