基于相关分析与最小二乘支持向量机的TE过程多模型建模

Multi-modeling of Tennessee Eastman Process Based on Correlation Analysis and Least Squares Support Vector Machine

摘要:

针对TE化工过程高度非线性、复杂性的特点,文章提出了一种基于相关分析和最小二乘支持向量机对TE过程进行多模型建模方法,以提高模型性能；首先对TE过程采用相关分析法划分为3个子系统,对每个子系统分别采用基于C-均值聚类的最小二乘支持向量机建模和基于K均值聚类的最小二乘支持向量机多模型建模；实验表明,基于K均值聚类的多模型建模能简化计算、提高模型精度、并且能更好的预测模型输出。

Abstract:

In order to deal with the nonlinear and complexity of TE,this paper proposes a multi-model modeling method for TE process based on CA and LSSVM which can overcome the disadvantages of single modeling. Firstly, the TE process is divided into three subsystems by CA. Then, LSSVM multi-modeling depended on the C-means and LSSVM multi-modeling depended on the k-means are applied to every subsystem respectively. At the same time, the experimental results also shows this algorithm can simplify the calculation, improve the accuracy of the model and forecast the model output better.