基于神经网络的造纸过程定量水分系统预测函数控制

摘要

针对造纸过程中的定量水分控制系统具有大时滞、大惯性、强耦合及多变量等特性,提出了一种新的多变量系统的预测函数控制方法.首先给出了PID神经网络解耦控制方法,将多变量系统变成单变量系统;然后根据预测函数控制原理对单变量系统设计控制器,并结合Smith预估思想,有效地克服了时滞对控制系统稳定性及控制品质的影响;最后通过仿真,证明该控制方法使系统具有在线计算量小、跟踪快速和精度高等特点,能够实现复杂多变量系统的有效控制.%Considering the features of basis weight and moisture control system in papermaking process , such as large time delay, big inertia, strong coupling and multivariable , it proposes a new predictive functional control method for multivariable systems .In order to change multivariable system into a single variable system , it pres-ents PID neural network decoupling control method .Based on the basic principle of predictive functional control and these decoupled single variable systems , it designs a controller .Combining Smith prediction , this control can overcome the delay influence on stability and performance index effectively .The simulation results show that the proposed method with less calculation online , fast tracking and high control accuracy can realize effective control for complex multivariable systems .