磨矿分级过程多模型智能控制研究  

Multiple Models Intelligent Control for Ball Mill Grinding Circuits

作　　者：陈夕松[1] 翟军勇[1] 李奇[1] 费树岷[1] 

机构地区：[1]东南大学自动化学院,南京210096

出　　处：《系统仿真学报》2009年第11期3342-3345,共4页Journal of System Simulation

基　　金：国家自然科学基金(60574006);博士点基金(20070286039;200770286040)

摘　　要：提出了一种非线性多模型控制算法,在若干平衡点附近首先建立多个子模型及相应的控制器,再通过在线计算模型匹配度来适应模型参数的变化,被控对象的最终输入是各控制器输出的加权和。将其应用于磨矿分级非线性系统的控制研究,仿真结果表明,所提算法极大地改善磨矿分级系统的动态响应,并增强了系统的鲁棒性。A nonlinear multiple models control algorithm was proposed. First of all, sub-models and corresponding controllers were established at each equilibrium point of the system, and then through online computation each model matching degree was calculated to accommodate the change of parameters. Finally, the input of the process is a weighted sum of each controller＇s output. The proposed algorithm has been employed for the simulation study on grinding and classification process and the simulation results demonstrate that the system dynamics and robustness have been much improved.

关 键 词：多模型 智能控制 非线性 磨矿分级 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]