基于PWA融合模型的注塑过程保压段建模及控制策略  

Modeling and control strategy for packing-stage of injection molding process based on PWA fusion model

作　　者：黄耀波 刘佳新 徐祖华[1] 赵均[1] 邵之江[1] HUANG Yaobo;LIU Jiaxin;XU Zuhua;ZHAO Jun;SHAO Zhijiang(National Laboratory of Industrial Control Technology,College of Control Science and Engineering,Zhejiang University,Hangzhou 310027,Zhejiang,China)

机构地区：[1]浙江大学控制科学与工程学院,工业控制技术国家重点实验室,浙江杭州310027

出　　处：《化工学报》2020年第3期1103-1110,共8页CIESC Journal

基　　金：国家重点研发计划项目(2017YFB0603703);NSFC-浙江两化融合联合基金项目(U1509209,U1809207)。

摘　　要：注塑过程是典型的批次过程,由注射、保压、冷却等阶段构成。保压段是决定制品质量的一个重要阶段,且具有较大的非线性和时变性。分段仿射(piece-wise affine, PWA)模型能够有效描述保压段的特性,但是PWA模型硬切换方式会导致输出跳变问题,因此提出一种基于时间划分的PWA融合模型,在子模型切换过程中引入切换区间,通过线性加权方法进行模型融合,并采用基于分离最小二乘的辨识方法进行参数估计。在此基础上,设计了基于内模整定的多模型PID控制方法。注塑机实验结果表明了算法的有效性。The injection molding is a typical batch process and consists of three stages: filling, packing-holding and cooling. Packing-holding is an important stage to determine product quality and has obvious nonlinear and timevarying characteristics. Piece-wise affine(PWA) model can effectively describe the dynamics of packing-holding stage, but the hard switching of PWA model will lead to the output jump problem. Therefore, we propose a novel time-partitioned PWA fusion model, in which the switching regionis introduced between adjacent sub-models and model fusion is realized by linear weighting method. Then, an identification algorithm based on separable nonlinear least-squares is developed to estimate model parameters. Based on this, a multi-model PID control method based on internal model tuning is designed.Experimental results of injection molding machine show the effectiveness of the algorithm.

关 键 词：注塑过程 PWA融合模型 分离最小二乘 多模型PID控制 

分 类 号：TP13[自动化与计算机技术—控制理论与控制工程]