并行算法纸张定量横向分布控制

Cross-direction Basis Weight Profile Control in Paper Machines Based on Parallel Algorithms

基于定量的横向耦合关系和纵向模型，建立了其横向分布的动态模型，设计了基于该模型的多变量广义预测控制器（ＧＰＣ），根据模型结构特征将控制算法进行分解，得到相应的ＧＰＣ并行算法，并给出其多处理机系统实现结构。计算机仿真结果表明，本文提出的建模与并行控制方法较好地改善了系统的的静、动态性能。

A novel dynamic model for cross-direction basis weight profile control is being proposed. The model takes into account the relationship between CD static behavior and the MD profile controlled with CARIMAX model. Based on such assumptions, a multi-variable generalized predictive controller (GPC)is designed. According to the structural characteristics of this model, a high dimensional GPC algorithm is decomposed into parallel algorithms. It is suggested that parallel control algorithms are to be implemented in a multiprocessor system. Computer simulation results indicate that system performance is improved.