Novel Hybrid Model to Predict Ribbon Solid Fractions in the Roller Compaction Process

Dec. 1, 2025, midnight - by Kiran Iyer , Senthil Kumarasamy , Pankaj Doshi , Allan Hutchins , Patrick Daugherity , Weili Yu , Ashwin Jain , Danielle Bubucis , Matthias Danner , Daniel Schwenk , Holger Schlack

Remove from favorites

Add to favorites

This work introduces a calibrated, data-efficient approach for scaling up roller compaction using only a minimal amount of laboratory-scale experimental data. The compression properties of three API-containing blends at various drug loadings were first characterized on a small-scale roller compaction simulator. These results were then used to build a predictive model that combines the Reynolds roller compaction framework with measured ribbon solid fractions, allowing accurate estimation of ribbon density based on process parameters such as roll force and roll gap. The model was validated by comparing predictions with ribbons produced on a Gerteis Minipactor at different scales, showing strong agreement across all tested formulations. The methodology was further extended into a hybrid regression model that incorporates API properties, formulation attributes, and Reynolds-model outputs to estimate solid fractions for partially characterized blends. Validation on a Gerteis 3-W-Polygran confirmed the model’s reliability. By enabling reliable digital Design of Experiments (DoE) and defining operating ranges with minimal experimental work, this approach provides an efficient digital tool to support roller compaction scale-up in fast-paced drug product development.

Read full article