多变量系统的结构辨识及其在电加热炉上的应用

• 作 者: 杨卫东 舒迪前 刘宏才
• 机 构: 北京钢铁学院,北京钢铁学院,北京钢铁学院
• 出 处: 信息与控制
• 发表日期: 1983年03期
• 关键词: 结构辨识 多变量系统 电加热炉 噪声情况 向量 结构指标 矩阵 最小特征值 判别准则 入向量

本文在 Guidorzi 结构辨识方法的基础上,通过引入向量系的标准化、相关性指标和相对相关性指标等概念,对有噪声情况下多变量系统结构辨识提出一种判别准则。采用此准则对电加热炉系统的模型结构进行了辨识,并进行了相应的参数估计。计算结果表明:此判别准则较 Guidorzi 和文献〔2〕所给出的方法,易于确定有噪声情况下模型的结构参数,且基本不受数据数量级及长度 N 的影响。

Based on the Guidorzi structure identification method, by introducing some concepts of vector system such as standardization, correlation index and relative correlation index, a discriminant criterion is proposed for structure identification of multivariable systems under noisy conditions. Using this criterion, the model structure of the electric heating furnace system is identified and the corresponding parameter estimation is carried out. The results show that compared with the method given by Guidorzi and [2], this discriminant criterion can easily determine the structural parameters of the model under noisy conditions, and is not influenced by the order of magnitude and the length N of the data.