基于无模型自适应控制的锌层厚度预估模型  

Coating Weight Predictive Model Based on Model-free Adaptive Control

作　　者：张岩[1] 邵富群[2] 吴鲲魁[3] 王军生[1] 

机构地区：[1]鞍钢股份有限公司技术中心,鞍山114009 [2]东北大学,沈阳110004 [3]辽宁科技大学,鞍山114021

出　　处：《系统仿真学报》2013年第5期1060-1064,共5页Journal of System Simulation

基　　金：国家自然科学基金(50604006);中央预算企业重大技术创新项目(2009-354)

摘　　要：针对锌层厚度控制过程中存在时变大滞后、非线性和多变量干扰而难以实现精确控制的问题,提出基于无模型自适应(MFA)结合Smith预估反馈控制方法。采用MFA预估算法可以实现Smith预估模型的参数偏差在线估计,并且不依赖预估模型的时间常数和延迟时间。从LabVIEW仿真数据结果可以看出,此控制策略对比增益自适应Smith控制算法,不仅能够有效地解决锌厚控制中系统高阶非线性问题,能有效克服不确定系统参数的扰动,并能有效处理系统时变大滞后问题。In order to solve the problem that accurate control is difficult to be realized for variable-time large delay, nonlinearity, multivariable disturbance in coating weight control process, coating weight control through model-free adaptive （MFA）with Smith predictive algorithm was proposed. By applying MFA estimation algorithm, parameters error of the Smith predictive model can be estimated on-line, and be independent of time parameter and delay time of the model. From LabVIEW simulation results, the control scheme resolves effectively strong non-linearity and high order in coating weight control, overcomes parameters distribution of uncertain system, and deals with variable-time rarge delay difficulty comparing with gain adaptive Smith r^redictive compensator.

关 键 词：锌层厚度控制 无模型自适应 Smith预估模型 LABVIEW仿真 

分 类 号：TP273.2[自动化与计算机技术—检测技术与自动化装置]