摘要
多变量积分过程的控制,一直是预测控制理论研究与应用过程中的难点问题.现有的研究成果更多的关注于算法的实现上,而很少关注理论依据.本文从积分过程的控制输入平衡关系出发,利用线性代数方程组解的相容性原理,得到了一个适用于判断多变量积分过程设定点是否可达的判据,可以作为算法能否实现多变量积分过程无静差控制的理论依据.同时分析了传统算法无法在存在模型失配情况下对积分过程进行优化与控制的原因,利用补偿因子重新设计反馈校正环节,使改进后的算法能够实现存在模型失配过程的优化与控制,并通过仿真验证了本文提出的结论.
The control of multivariable integrating process has always been a difficult issue in model predictive control for both the academic researches and applications.The existing researches pay more attention to the algorithm design rather than to the theoretical analyses.This paper starts from the balance of the control input for the integrating process,and applies the compatibility principle of linear algebraic equations to determine whether the set-points are reachable,establishing a theoretical basis for the double-layered predictive control.The poor control performance of model predictive control (MPC) in model mismatch is analyzed,and a compensation factor is employed for updating the algorithm so that it can be used to control the integral process with model-plant mismatch.Simulation results validate the effectiveness of the improved algorithm.
出处
《控制理论与应用》
EI
CAS
CSCD
北大核心
2014年第2期165-174,共10页
Control Theory & Applications
基金
国家自然科学基金资助项目(61374112
61174095
61004051)
中国科学院知识创新资助项目(KGCX2-EW-104)
浙江省自然科学基金资助项目(Y12F030052)
关键词
模型预测控制
多变量过程
积分控制
稳态分析
模型失配
不可测扰动
model predictive control (MPC)
multi-variable process
integrated control
steady-state analysis
model-plant mismatch
unmeasured disturbance