A Time Varying State Space Approach for Sugar Crystallization Process Modelling and Monitoring
| Authors: | Georgieva Petia, University of Aveiro, Portugal
Simoglou Alexandros, University of Newcastle, United Kingdom
Martin Elaine, University of Newcastle, United Kingdom
Morris Julian, University of Newcastle, United Kingdom
de Azevedo Sebastião Feyo, University of Porto, Portugal |
|---|
| Topic: | 6.1 Chemical Process Control |
|---|
| Session: | Knowledge Driven Batch Processes |
|---|
| Keywords: | Statistical process control, dynamic principal component analysisbatch sugar crystallization |
|---|
Abstract
The present paper contributes to the issues of batch process modelling andmonitoring by proposing a time-varying state space (TVSS) model for the evaporative sugarcrystallization industrial process. The study is focused on issues of on-line detection ofchanges in crystallization process operation, the early warning of process malfunctions andpotential production failures; problems that have not been directly addressed by existingstatistical monitoring schemes. The TVSS methodology is compared with current state-ofthe-art techniques and the results obtained demonstrate the superior performance of the TVSSapproach to successfully detect abnormal events and periods of bad operation.
