基于两步子空间划分的化工过程监测方法

发布时间：2018-07-09 18:28

本文选题：过程监测 + 两步子空间划分　；参考：《山东大学学报(工学版)》2017年05期

【摘要】：为了解决现代化工过程采集的数据维度高、分布复杂的问题,提出一种基于两步子空间(two step subspace division,TSSD)划分的化工过程监测方法。为了降低过程分析复杂度,将具有相似特性的变量划分为同一空间。考虑数据的复杂分布问题,将第一步得到的每个子空间划分为高斯空间与非高斯空间。利用主元分析(principal component analysis,PCA)和独立元分析(independent component analysis,ICA)方法建立检测模型并构造统计量。整合每个子空间的统计量并基于局部离群因子(local outlier factor,LOF)方法构建综合统计量。结果表明:TSSD方法对于16个故障均能取得最优的漏报率,尤其是故障10和故障16,漏报率分别为15.375%和6.75%,有效验证所提出的基于两步子空间划分的过程监测方法的优越性。
[Abstract]:In order to solve the problem of high dimension and complex distribution of data collected in modern chemical process, a chemical process monitoring method based on two-step subspace (two step subspace division is proposed. In order to reduce the complexity of process analysis, variables with similar characteristics are divided into the same space. Considering the complex distribution of data, each subspace obtained in the first step is divided into Gao Si space and non-Gao Si space. Principal component Analysis (principal component) and Independent component Analysis (independent component) were used to establish the detection model and construct statistics. The statistics of each subspace are integrated and the synthetic statistics are constructed based on the local outlier factor (local outlier factor) method. The results show that the optimal miss rate can be obtained for 16 faults, especially for fault 10 and fault 16. The missing rate is 15.375% and 6.75 respectively. The superiority of the proposed process monitoring method based on two-step subspace partition is verified effectively.
【作者单位】：华东理工大学化工过程先进控制和优化技术教育部重点实验室;
【基金】：国家自然科学基金资助项目(61374140);国家自然科学基金资助项目(61373173) 中央高校基本科研业务费专项资金资助项目(222201714031)
【分类号】：TP274;TQ02

【相似文献】