针对化学机械研磨(CMP)过程非线性、时变和产品质量不易在线测量的特性,提出了一种基于T-S模糊模型的CMP过程智能run-to-run (R2R)预测控制器FIPR2R；通过G-K聚类算法和最小二乘法对CMP过程的T-S模糊预测模型离线辨识,解决了复杂CMP过程难以建立精确数学模型的难题和提高了模型预测精度;通过双指数加权移动平均(dEWMA)中对过程扰动及漂移进行估计的方法实现反馈校正和基于克隆选择算法的滚动优化求取最优控制律;提高了控制精度;性能分析结果表明,FIPR2R控制器的控制性能优于dEWMA方法,有效抑制了过程扰动和漂移的影响.%For chemical mechanical polishing (CMP ) process characteristics of nonlinear , time -varying and not being in-situ meas-ured easily, a CMP process intelligent run-to-run (R2R) predictive controller named FIPR2R is proposed. CMP process T- S fuzzy pre-dictive model off-line identified by G-K clustering algorithms and least squares method solves difficult problem of constructing accurate mathematical model of complicated CMP and reduces model error. Feedback correction is achieved from disturbances and drifts estimated by dEWMA method and optimal control law is derived from receding optimization based on clonal selection algorithm, thus control precision is improved. Performance analysis results illustrate that FIPR2R controller is better than dEWMA on control performance, process disturb-ances and drifts is suppressed significantly, and RMSE of material removal rate (MRR) brought down is 5. 2% compared to dEWMA.
展开▼
