基于GMM的间歇过程故障检测

王静; 胡益; 侍洪波

doi:10.16383/j.aas.2015.c130680

基于GMM的间歇过程故障检测

doi: 10.16383/j.aas.2015.c130680

1.
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上海 200237

基金项目:

国家自然科学基金(61374140)资助

详细信息

作者简介:
王静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硕士研究生. 主要研究方向为间歇过程故障检测与诊断.E-mail: rain12027@163.com

通讯作者:
侍洪波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教授.2000 年获得华东理工大学控制理论与控制工程专业博士学位. 主要研究方向为工业过程建模, 控制与优化, 化工过程故障检测与诊断. E-mail: hbshi@ecust.edu.cn

计量
- 文章访问数: 1909
- HTML全文浏览量: 161
- PDF下载量: 1312
- 被引次数: 0
出版历程
- 收稿日期: 2013-07-16
- 修回日期: 2014-03-25
- 刊出日期: 2015-05-20

Fault Detection for Batch Processes Based on Gaussian Mixture Model

1.
Key Laboratory of Advanced Control and Optimization for Chemical Processes of Ministry of Education, East China University of Science and Technology, Shanghai 200237

Funds:

Supported by National Natural Science Foundation of China (61374140)

摘要

摘要: 对间歇过程的多操作阶段进行划分时,往往会被离群点和噪声干扰,影响建模的精确性,针对此问题提出一种新的方法:主元分析--多方向高斯混合模型(Principal component analysis-multiple Gaussian mixture model, PCA-MGMM)建模方法.首先用最短长度法对数据进行等长处理,融合不同展开方法相结合的处理方式消除数据预估问题;利用主元分析方法将数据转换到对故障较为敏感的低维子空间中,得到主元的同时消除了离群点和噪声的干扰;通过改进的高斯混合模型(Gaussian mixture model, GMM)算法对各阶段主元进行聚类,减少了运算量的同时自动得到最佳高斯成分和对应的统计分布参数;最后将局部指标融合为全局概率监控指标,实现了连续的在线监控.通过一个实际的半导体制造过程的仿真研究验证了所提方法的有效性.
- 间歇过程 /
- 多阶段操作 /
- 故障检测 /
- 高斯混合模型 /
- 全局概率指标
Abstract: A new novel of principal component analysis-multiple Gaussian mixture model (PCA-MGMM) method is proposed in this article to handle the problem about outliers and noise interference, which affects the accuracy of modeling when dividing multiple operation phases in batch processes. At first, a shortest length approach is used to solve the problem of unequal data, while a method of hybrid unfolding of a multi-way data matrix is used to eliminate data estimate problem. Secondly, using PCA sequentially to achieve a low-dimensional representation of the original data space, the operation not only gets the principal but also eliminates the outliers and noise interference. After that the modified algorithm of Gaussian mixture model (GMM) is adopted to automatically optimize the number of Gaussian components and estimate their statistical distribution parameters so as to reduce the computational complexity. Finally, the online monitoring is guaranteed to be continuous by using a global probability index integrated by local probability indices of each operation. The effectiveness and flexibility of the proposed method is validated through an empirical study on a real semiconductor process.
- Batch processes /
- multiple operation phases /
- fault detection /
- Gaussian mixture model (GMM) /
- global probability index

HTML全文

参考文献(17)

[1]	Cinar A, Parulekar S, Undey C, Birol G. Batch Fermentation: Modeling, Monitoring, and Control. New York: CRC Press, 2003.
[2]	[2] Lee J M, Yoo C K, Lee I B. Enhanced process monitoring of fed-batch penicillin cultivation using time-varying and multivariate statistical analysis. Journal of Biotechnology, 2004, 110(2): 119-136
[3]	[3] Nomikos P, MacGregor J F. Monitoring batch processes using multi-way principal component analysis. AIChE Journal, 1994, 40(8): 1361-1375
[4]	[4] Kourti T, Nomikos P, MacGregor J F. Analysis, monitoring and fault diagnosis of batch processes using multi-block and multiway PLS. Journal of Process Control, 1995, 5(4): 277-284
[5]	Zhang Jia, Sun Wei, Zhao Jin-Song, Sun Mei-Hong. Fault detection of batch process based on multi-phase multiway principal component analysis. Computers and Applied Chemistry, 2010, 27(3): 298-302(张佳, 孙巍, 赵劲松, 孙美红. 多段MPCA法监测间歇过程的故障. 计算机与应用化学, 2010, 27(3): 298-302)
[6]	Chang Yu-Qing, Wang Zhu, Tan Shuai, Wang Fu-Li, Yang Jie. Research on multistage-based MPCA modeling and monitoring method for batch processes. Acta Automatica Sinica, 2010, 36(9): 1312-1320(常玉清, 王姝, 谭帅, 王福利, 杨洁. 基于多时段MPCA模型的间歇过程监测方法研究. 自动化学报, 2010, 36(9): 1312-1320)
[7]	[7] Yu J, Qin S J. Multimode process monitoring with Bayesian inference-based finite Gaussian mixture models. AIChE Journal, 2008, 54(7): 1811-1829
[8]	[8] Yu J, Qin S J. Multiway Gaussian mixture model based multiphase batch process monitoring. Industrial and Engineering Chemistry Research, 2009, 48(18): 8585-8594
[9]	Zhao Chun-Hui, Wang Fu-Li, Yao Yuan, Gao Fu-Rong. Phase-based statistical modeling, online monitoring and quality prediction for batch processes. Acta Automatica Sinica, 2010, 36(3): 366-374(赵春晖, 王福利, 姚远, 高福荣. 基于时段的间歇过程统计建模、在线监测及质量预报. 自动化学报, 2010, 36(3): 366-374)
[10]	Rothwell S G, Martin E B, Morris A J. Comparison of methods for dealing with uneven length batches. In: Proceedings of the 1998 International Conference on Computer Application in Biotechnology. Osaka, Japan: Elsevier, 1998. 387-392
[11]	Zhou Dong-Hua, Li Gang, Li Yuan. Fault Diagnosis Technology of Data-driven in Industrial Process. Beijing: Science Press, 2011. 39-41 (周东华, 李钢, 李元. 数据驱动的工业过程故障诊断技术. 北京: 科学出版社, 2011. 39-41)
[12]	He Ming, Feng Bo-Qin, Ma Zhao-Feng, Fu Xiang-Hua. A unsupervised rough clustering method based on Gaussian mixture model. Journal of Harbin Institute of Technology, 2006, 38(2): 256-259, 322(何明, 冯博琴, 马兆丰, 傅向华. 一种基于高斯混合模型的无监督粗糙聚类方法. 哈尔滨工业大学学报, 2006, 38(2): 256-259, 322)
[13]	Figueiredo M A T, Jain A K. Unsupervised learning of finite mixture models. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2002, 24(3): 381-396
[14]	Choi S W, Park J H, Lee I B. Process monitoring using a Gaussian mixture model via principal component analysis and discriminant analysis. Computers and Chemical Engineering, 2004, 28(8): 1377-1387
[15]	Chiang L H, Russell E, Braatz R D. Fault Detection and Diagnosis in Industrial Systems. New York: Springer Verlag, 2001. 36-37
[16]	Xie X, Shi H B. Dynamic multimode process modeling and monitoring using adaptive Gaussian mixture models. Industrial and Engineering Chemistry Research, 2012, 51(15): 5497-5505
[17]	Barry M W, Neal B G, Stephanie W B, Daniel D W J, Gabriel G B. A comparison of principal component analysis, multiway principal component analysis, trilinear decomposition and parallel factor analysis for fault detection in a semiconductor etch process. Journal of Chemometrics, 1999, 13(3-4): 379-396