基于模糊预测的MOCVD温度控制方法

The research on MOCVD temperature control based on the fuzzy predictive mechanism

为解决金属有机化合物化学气相淀积(MOCVD)设备温度控制的非线性、时变性以及大滞后等问题,给出了一种用模糊控制和预测控制相结合的复合控制方法.该模糊预测控制在模糊控制的基础上加入预测环节,通过增加相位超前减少时间滞后所带来的影响,准确预测温度变化趋势.与传统的控制方法相比较,具有算法简单,鲁棒性好的特点.仿真和试验表明,该方法有着较高的稳态精度和动态特性,在整个温度控制范围基本误差可达到1℃‰,有效改善了MOCVD系统温度的控制性能,对实际温度控制具有较好的指导意义.

This paper presents a new fuzzy predictive control method for the MOCVD device control system set to handle the non-linearity, time variability and large delay of plant controls. The designed controller consists of a fuzzy part and a predictor. The dynamics of the prediction term is adaptive to the system＇s time delay. Simulation results illustrate that this method is well adaptive to the alteration and improves the static precise and dynamic performance of the system, compared with its conventional counterpart. In addition, the basic error can reach 1℃‰ referred to input range.