基于聚类算法的C-R模糊模型结构辨识

C-R Fuzzy Model Structure Identification Based on Clustering Algorithm

工业过程中多数系统呈现出非线性、时变性和多模态性等特征,往往难于用机理建模的方法建立它的模型,因此利用系统的输入和输出数据进行非机理建模是非常有意义的。对C-R模糊模型进行了改进,应用关系度聚类算法在线辨识出系统的模态,即系统C-R模型的模糊子空间的数目,提出了C-R模糊模型的聚类建模方法,仿真结果表明了该算法的有效性,节省了运算时间,简化了运算过程。

As the characteristic of most industrial process control systems are nonlinear, time-varying and multi-model, it is challenge to identify the model. The paper improve the clustering modeling method of the C R fuzzy and identifies the system mode, in oth- er words, the number of the fuzzy subspaces, with the clustering method based on relation degree. The simulation result shows that the identification mnethod is very effective.