大工业过程的稳态模型分散辨识及其强一致性分析  

Decentralized identification of steady-state models for large-scale industrial processes and its strong consistency

作　　者：刘知贵[1] 黄正良[1] 

机构地区：[1]西南交通大学计算机与通信工程学院

出　　处：《控制与决策》2005年第3期316-320,共5页Control and Decision

基　　金：国家自然科学基金项目(69674003).

摘　　要：针对动态线性大工业过程,提出了获得其可分稳态模型强一致性估计的分散辨识方法.该方法仅使用设定点的阶跃信号作为输入激励信号,并且每个子过程的输入输出稳态模型辨识是在相应的局部单元完成的,因而大大减少了对过程的干扰和信息的交换量.所提出的方法简洁,并且辨识精度高,仿真结果说明了该辨识方法的有效性和实用性.The decentralized identification technique for dynamic linear large-scale industrial processes is presented by which the strong consistency estimates of the steady-state models of the processes are obtained. The novel technique can only use step signals as input signals, and identification of input-output models of each sub processes is only implemented in corresponding to the local unit. Hence the disturbance to the processes and interchange of information of sub processes are decreased greatly. Simulation results show that the proposed technique is very efficient and practical.

关 键 词：大工业过程 分散辨识 强一致性 稳态模型 子过程 

分 类 号：TP11[自动化与计算机技术—控制理论与控制工程]