Contribution information
| Title | Data-based Design of Inferential Sensors for an Industrial Depropanizer Column with Data Pre-treatment Analysis |
| Status | Accepted |
| Final type | Poster |
| Final session | Industrial process design and control |
| Authors |
M., Mojto1,
K., Ľubušký2,
M., Fikar3,
R., Paulen4
1 Slovak University of Technology in Bratislava, Bratislava, Slovakia
2 SLOVNAFT, a.s., Vlčie hrdlo 1, 824 12 Bratislava, Slovakia, Bratislava, Slovakia
3 Slovak University of Technology in Bratislava, Bratislava, Slovakia
4 Slovak University of Technology in Bratislava, Bratislava, Slovakia
|
| Uploaded abstract | link |
| Brief content | Inferential (or soft) sensors infer rarely measured or completely unmeasured variables. The main challenge in designing an inferential sensor is to select a correct structure represented by sensor input variables. This work focuses on designing an inferential sensor for an industrial depropanizer column. The raw industrial data is pre-treated by the minimum covariance determinant (MCD) method. Subsequently, the inferential sensors are designed using several methods (OLSR, PCR, LASSO). Finally, the performance of designed sensors is compared with current inferential sensors in the refinery. |
| ID | 485 |