Czochralski Crystal Furnace Thermal Field Model Parameter Identification And The Model Predictive Control

The essential points of Czochralski crystal furnace crystal’s growth is to control the temperature in Czochralski crystal furnace thermal field. The control of thermal field temperature has a direct influence on the quality of the single crystal’s growth. The main purpose of our research is to provide the crystal with the most appropriate temperature by adjusting the heating power and the heating power based on the thermal field temperature and deviation of reference temperature.This paper mainly discusses the physical structure, technological process and control system of the single crystal furnace, as well as analyzes thermal equilibrium, heat flow pattern, heating power of the single crystal’s growth and the practical measured data of thermal field temperature. Then we master the changing of the heating power and thermal field temperature and establish a Single crystal furnace thermal field heating-warming model based on the pattern of thermal balance. Through MATLAB parameter optimization tools and based on the practical data and actual production experience, we identify and optimize the parameters of single crystal furnace thermal field heating-warming model and obtain the accurate Czochralski crystal furnace thermal field model.Czochralski single crystal furnace is a typical complex nonlinear system. This paper uses multiple model strategy and adopts the method of piecewise linearization to resolve the model of thermal field temperature in the main period of the crystal’s growth. With the small slant method, we linearize each model, compound piecewise models by adopting the strategy of model switching, and replace the original nonlinear model. Besides, we design and use the method of model predictive control, simulate the process of the crystal growth of the single crystal furnace, to control the process of adjusting the heater power by the deviation of output temperature and reference temperature. Through the simulation, we can see that the stability and adaptation of it is better than the normal PID control. I hope this kind of explore will be of great help to improve the automation of single crystal furnace technology.