大工业过程稳态模型的分散辨识

Decentralized Identification of Steady-state Models for Large-scale Industrial Processes

针对大工业过程，利用优化过程中设定点的阶跃信号。采用分散辨识的方法，获得了大工业过程的稳态模型，并给出了其可辨识的充分条件。同传统的辨识方法相比，该技术具有精度高、计算简单、信息传递量少以及对系统干扰小等特点其，仿真研究显示了分散辨识技术的有效性和实用性。

In this paper, under mild conditions, the steady-state models for large-scale industrial processesare obtained by using the step signals of set-points in the procedure of optimizing control, and the sufficient condi-tions are obtained. The identification technique has advances of high accurately simple count, and less interferenceto the indusrical processes, as well as less information transmission with contrast to classical identification tech-niques.The validity and practicabity of this complete decentralized identification are demonstrated by sedation re-sults.