多操作空间强非线性系统自适应模型辨识的子空间方法

项目名称： 多操作空间强非线性系统自适应模型辨识的子空间方法

项目编号： No.61273188

项目类型： 面上项目

立项/批准年度： 2013

项目学科： 自动化技术、计算机技术

项目作者： 朱全民

作者单位： 中国石油大学（华东）

项目金额： 80万元

中文摘要： 石化生产中存在大量具有多个子空间的强非线性时变过程。对于这类过程，常采用自适应、模型预测等依赖于模型的控制方法,建模质量决定了控制系统性能。现有的线性模型指数遗忘最小二乘迭代算法不能有效地解决具有多操作子空间的非线性过程建模问题。本课题拟提出一套系统的非线性自适应模型子空间遗忘算法，能够根据系统输入输出信息辨识操作域的工作子空间，自适应的改变系统模型，为解决这类石化工业中的普遍问题提供一种行之有效的新方法。课题将以中国石油大学重质油国家重点实验室为依托开展实验研究。所得结论不仅能应用于自适应、模型预测等常见石油化工过程控制方法而且能为相关石化工艺与装备研究提供理论依据。开展这项研究，对阐明非线性石化过程建模机理、揭示具有多平衡点非线性过程建模一般规律，具有积极的科学和应用意义。

中文关键词： 系统辨识；递归参数估计；模型有效性检验；非线性过程建模；

英文摘要： There are many strong nonlinear time varying processes which have multiple subspaces in petrochemical production. For these classes of processes, adaptive control, model predictive control and other model based control approaches are often used to guarantee the desired performances. The modelling quality decides the performances of the controlled systems. The existing linear modle exponent forgetting least square iterative algorithm cannot deal with the modelling issues of multiple subspaces nonlinear processes. The proposed research will present a new and systematic nonlinear system adaptive model subspace forgetting algorithm. It can identify the working subpace of the operating filed according to the systems' input and output information and can change system model adaptively. Then this research can presented an effective new approach for the general issues in petrochemical production. The corresponding experiments will be carried out in the State Key Laboratory of Heavy Oli Processing of China University of Petroleum.The research results cannot only be applied to adaptive control, model predictive control and other model based petrochemical process control but also can be used as theoretical basis for the petrochemical technology and equipments. The proposed research has scientific significance and generic/e

英文关键词： system identification；recursive parameter estimation；model validity test；nonlinear process modelling；