Pattern matching in historical batch data using PCA

The article seeks to answer the question: how can relevant information be extracted from huge historical databases? A pattern-matching methodology has been evaluated in a case study for a batch fermentation process. The proposed approach is both data driven and unsupervised. The new approach relies on PCA and a new similarity factor based on distance between the two datasets. The computational requirements are modest, allowing large databases to be processed in a relatively small amount of time.

[1]  A. J. Morris,et al.  An overview of multivariate statistical process control in continuous and batch process performance monitoring , 1996 .

[2]  Manabu Kano,et al.  Dissimilarity of Process Data for Statistical Process Monitoring , 2000 .

[3]  Stephen Grossberg,et al.  ART 2-A: An adaptive resonance algorithm for rapid category learning and recognition , 1991, Neural Networks.

[4]  Karlene A. Hoo,et al.  Process Data Analysis and Interpretation , 1999 .

[5]  Richard O. Duda,et al.  Pattern classification and scene analysis , 1974, A Wiley-Interscience publication.

[6]  J. C. Hale,et al.  Historical data recording for process computers , 1981 .

[7]  L Yerushalmi,et al.  Mathematical model of a batch acetone–butanol fermentation , 1986, Biotechnology and bioengineering.

[8]  Dale E. Seborg,et al.  Dynamic data rectification using the expectation maximization algorithm , 2000 .

[9]  Ali Cinar,et al.  Diagnosis of process disturbances by statistical distance and angle measures , 1997 .

[10]  C. Apte,et al.  Data mining: an industrial research perspective , 1997 .

[11]  J. Macgregor,et al.  Experiences with industrial applications of projection methods for multivariate statistical process control , 1996 .

[12]  Chonghun Han,et al.  Intelligent systems in process engineering : A review , 1996 .

[13]  Ali Cinar,et al.  Multivariate statistical methods for monitoring continuous processes: assessment of discrimination power of disturbance models and diagnosis of multiple disturbances , 1995 .

[14]  Dale E. Seborg,et al.  Pattern Matching in Multivariate Time Series Databases Using a Moving-Window Approach , 2002 .

[15]  Theodora Kourti,et al.  Multivariate SPC Methods for Process and Product Monitoring , 1996 .

[16]  Ahmet Palazoglu,et al.  Detection and classification of abnormal process situations using multidimensional wavelet domain hidden Markov trees , 2000 .

[17]  D. Seborg,et al.  Pattern Matching in Historical Data , 2002 .

[18]  S. Joe Qin,et al.  Joint diagnosis of process and sensor faults using principal component analysis , 1998 .

[19]  W. Krzanowski.  Between-Groups Comparison of Principal Components , 1979 .

[20]  Venkat Venkatasubramanian,et al.  A wavelet theory-based adaptive trend analysis system for process monitoring and diagnosis , 1997, Proceedings of the 1997 American Control Conference (Cat. No.97CH36041).

[21]  Ahmet Palazoglu,et al.  Classification of abnormal plant operation using multiple process variable trends , 2001 .

[22]  Christos Georgakis,et al.  Disturbance detection and isolation by dynamic principal component analysis , 1995 .