Monitoring the Moisture Content in Pharmaceutical Batch Fluidized Bed Dryers Using Observer-Based Soft Sensors. Issue 2 (2020)

• Record Type: Journal Article
• Authors: Roseberry, Marc-Olivier; Gagnon, Francis; Desbiens, André; Bouchard, Jocelyn; Lapointe-Garant, Pierre-Philippe
• Is Part Of: IFAC-PapersOnLine. Volume 53:Issue 2(2020)
• Page Start: 12056
• Page End: 12061
• Publication Date: 2020
• Subjects: state estimation -- batch fluidized bed dryer -- moving horizon estimator -- offline measurement -- measurement delay; Automatic control -- Periodicals; 629.805
• DOI: 10.1016/j.ifacol.2020.12.747
• Languages: English
• ISSNs: 2405-8963

Abstract

Abstract: Tablet manufacturing in the pharmaceutical industry involves batch fluidized bed drying for particle moisture removal. This paper introduces five approaches for moisture content monitoring, relying either on a complex phenomenological model or its simplified version. The first two soft sensors consist of open-loop estimators, i.e. they simply simulate the models fed by the manipulated variables. Three closed-loop moving horizon estimators based on the simplified model are also proposed for improved robustness. In the first one, the measurements of the inlet gas and particle temperatures feed back the soft sensor. The last two closed-loop observers additionally can take into account infrequent delayed moisture content measurements, such as at-line loss on drying analysis. A validation of the soft sensors is performed with experimental data collected on a pilot scale fluidized bed dryer. Results show that the closed-loop observer with the delayed moisture content measurements still has an accuracy that is equivalent (and sometimes better) than the complex phenomenological model.