Abstract
After introducing process data analytics using latent variable methods and machine learning, this paper briefly review the essence and objectives of latent variable methods to distill desirable components from a set of measured variables. These latent variable methods are then extended to modeling high dimensional time series data to extract the most dynamic latent time series, of which the current values are best predicted from the past values of the extracted latent variables. We show with an industrial case study how real process data are efficiently and effectively modeled using these dynamic methods. The extracted features reveal hidden information in the data that is valuable for understanding process variability.
| Original language | English |
|---|---|
| Pages (from-to) | 69-80 |
| Number of pages | 12 |
| Journal | Computers and Chemical Engineering |
| Volume | 114 |
| Early online date | 4 Nov 2017 |
| DOIs | |
| Publication status | Published - 9 Jun 2018 |
| Externally published | Yes |
Bibliographical note
Funding for the work in this paper from the following sources is gratefully acknowledged: the Natural Science Foundation of China under Grant 61490704, the Fundamental Research Program of the Shenzhen Committee on Science and Innovations (20160207, 20170155), and the Texas-Wisconsin-California Control Consortium. The authors are grateful to the Eastman Chemical process control team, in particular John Cox, for providing the industrial process data analyzed in this work.Keywords
- Dynamic PCA
- Dynamic PLS
- Dynamic CCA
- Dynamic latent variable modeling
- Process monitoring
- Dynamic feature extraction