Abstract
In this paper, a novel dynamic-inner canonical correlation analysis (DiCCA) algorithm is proposed to extract dynamic components from high dimensional dynamic data. DiCCA extracts latent variables with descending dynamics, which are referred to as principal time series. Since DiCCA enables the principal time series to have maximal predictability, the most important dynamic features in the data are guaranteed to be extracted first. Therefore, usually a lower dimensional principal time series are able to provide good representation of the dynamic features, leading to the ease of interpretation and visualization. A case study on the Eastman plant-wide oscillating dataset demonstrates the effectiveness of the proposed method. Combined with Granger causality analysis, major oscillatory latent dynamics are analyzed, identified, and localized to equipment malfunctions.
| Original language | English |
|---|---|
| Pages (from-to) | 476-481 |
| Number of pages | 6 |
| Journal | IFAC-PapersOnLine |
| Volume | 51 |
| Issue number | 18 |
| Early online date | 8 Oct 2018 |
| DOIs | |
| Publication status | Published - 2018 |
| Externally published | Yes |
Bibliographical note
This work was supported in part by the Natural Science Foundation of China (61490704), the Fundamental Research Program of the Shenzhen Committee on Science and Innovations (20160207, 20170155), the Post-doctoral Fellowship Fund of the Chinese University of Hong Kong, Shenzhen, and the Texas-Wisconsin-California Control Consortium.Keywords
- dynamic data modeling
- Granger causality analysis
- latent dynamic model
- root cause diagnosis