| 基于递归广义特征值分解的化工过程监控方法 |
| |
| 引用本文： | 高翔,刘飞. 基于递归广义特征值分解的化工过程监控方法[J]. 计算机与应用化学, 2007, 24(6): 803-806 |
| |
| 作者姓名： | 高翔 刘飞 |
| |
| 作者单位： | 江南大学自动化研究所,江苏,无锡,214122;江南大学自动化研究所,江苏,无锡,214122 |
| |
| 基金项目： | 国家高技术研究发展计划(863计划)课题(No.2006AA020204)，教育部新世纪优秀人才支持计划(NCET-05-0485).~~ |
| |
| 摘 要： | 根据独立元分析理论,提出了一种基于递归广义特征值分解的化工过程监控方法。广义特征值分解理论上等价于盲源分离领域中针对高阶统计量的联合对角化方法,而其递归算法又能很好地收敛至最优解。鉴于基于递归广义特征值分解的真CA方法在发掘数据集中隐含信息方面的优良性能,本文将其引入工业化工过程监控,并详述了具体步骤。在田纳西-伊斯曼过程中的应用验证了其有效性。
|
| 关 键 词： | 递归广义特征值分解 独立元分析 化工过程监控 |
| 文章编号： | 1001-4160（2007）06-803-806 |
| 修稿时间： | 2007-05-212007-06-18 |
Chemical process monitoring method using recursive generalized eigende-composition |
| |
| Gao Xiang,Liu Fei. Chemical process monitoring method using recursive generalized eigende-composition[J]. Computers and Applied Chemistry, 2007, 24(6): 803-806 |
| |
| Authors: | Gao Xiang Liu Fei |
| |
| Affiliation: | Institute of Automation, Southern Yangtze University, Wuxi, 214122, Jiangsu, China |
| |
| Abstract: | A new chemical process monitoring method based on ICA theory is proposed in this paper using recursive generalized eigendecomposition (RGED). Generalized eigendecomposition (GED) is essentially equivalent to the joint diagonalization of high order statistics that used in blind source separation. GED's recursive approach shows good ability to converge to the optimal solution. As the RGED ICA method indicates a good performance in extracting the most important information from data, this paper introduces it into industrial chemical process monitoring and discusses it in details. The application of Tennessee-Eastman (TE) process shows the validity of the proposed method. |
| |
| Keywords: | recursive generalized eigendecomposition independent component analysis chemical process monitoring |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
