论文部分内容阅读
化学机械抛光(chemical mechanical polishing,CMP)过程中由于柔性弹性隔膜的存在使得各腔室之间压力相互耦合,从而使得多区腔室的压力控制变得复杂。针对这一耦合现象,提出了一种将基于动态回归神经网络(dynamic recurrent neural network,DRNN)在线辨识与神经元解耦和分段变参数复合控制相结合的方案。利用DRNN的非线性映射能力以及神经元的在线实时动态解耦特性,获得对象的逆模型,消除了各区之间的耦合;采用分段变参数控制策略减少了由于初始时刻逆控制模型辨识不准而带来的不利影响和系统动荡,使得整个控制系统趋于稳定。实验结果表明:该方案不仅具有很好的在线辨识和解耦能力,同时较常规定参数比例积分微分(proportional integral derivative,PID)控制还具有自适应能力强、响应速度快、超调量小以及鲁棒性好等特点。
In the process of chemical mechanical polishing (CMP), the pressure of each chamber is coupled to each other due to the presence of the flexible elastic membrane, so that the pressure control of the multi-chamber chamber is complicated. In view of this coupling phenomenon, a scheme of combining on-line identification of dynamic recurrent neural network (DRNN) with compound control of neuronal decoupling and piecewise variable parameters is proposed. By using the nonlinear mapping ability of DRNN and the online real-time dynamic decoupling characteristic of neurons, the inverse model of the object can be obtained, and the coupling between different regions can be eliminated. Using the sub-variable parameter control strategy reduces the uncertainty of the inverse control model The adverse effects and system turmoil brought the entire control system stabilized. The experimental results show that the proposed scheme not only has good online identification and decoupling ability, but also has the advantages of adaptive ability, fast response, small overshoot and Robust and so on.