多变量ARX-Laguerre函数PID预测控制算法及其应用  

Multivariable ARX-Laguerre function PID predictive control algorithm and its application

作　　者：郭伟[1,2] 邵帅 李涛[1,2] 张鹏程 GUO Wei;SHAO Shuai;LI Tao;ZHANG Pengcheng(Jiangsu Collaborative Innovation Center on Atmospheric Environment and Equipment Technology, Nanjing University of Information Science and Technology, Nanjing 210044, China;School of Automation, Nanjing University of Information Science and Technology, Nanjing 210044, China)

机构地区：[1]南京信息工程大学江苏省大气环境与装备技术协同创新中心,南京210044 [2]南京信息工程大学自动化学院,南京210044

出　　处：《中国科技论文》2020年第7期767-773,共7页China Sciencepaper

基　　金：国家自然科学基金资助项目(61573189,61973168)。

摘　　要：针对重油催化裂化分馏塔控制系统满足稳定设计却难以优化系统性能的问题,提出了一种基于多变量ARX-Laguerre函数模型的PID预测控制(multivariable ARX-Laguerre function model predictive control combined with proportion integration differentiation,MALMPCPID)算法。以增量式的多变量ARX-Laguerre函数为预测模型,通过带有遗忘因子的最小二乘递归方法在线辨识Laguerre系数矩阵,并将滚动优化的性能指标改写成PID参数形式,以提高系统的动态性能。将所提的MALMPCPID算法应用于Shell公司重油催化裂化分馏塔进行仿真实验并与现有算法进行对比。结果表明,所提算法对多变量强耦合过程具有良好的控制效果,而且快速性和解耦能力得到了大大提高。According to the problem that it was difficult to optimize the system performance when the design of stabilization was satisfied in resid fluid catalytic cracking unit fractionator,a PID predictive control algorithm based on MIMO ARX-Laguerre function model is proposed.In this algorithm,the incremental multivariable ARX-Laguerre function model was used as the prediction model,and the Laguerre coefficient matrix was identified online by the least square recursive method with forgetting factor.In order to improve the dynamic performance of the system,the performance index of rolling optimization was rewritten into PID parameter.The proposed MALMPCPID algorithm was applied to simulation experiment of Shell heavy oil catalytic cracking fractionator and compared with the existing algorithm.The results show that the proposed algorithm has good control effect for multivariable strong coupling process,and the rapidity and decoupling ability are greatly improved.

关 键 词：自动控制理论 重油催化裂化分馏塔 模型预测控制 多变量ARX-Laguerre函数模型 比例积分微分控制 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]