(660a) A Pharmaceutical Application of the Process Systems Engineering Workflow: Combining Data-Driven Approaches with Physics-Based Models for Fluid Bed Granulation.

This talk will cover the use of a commercial process simulator to create a model of a fluid bed dryer without the use of pre-existing libraries. The step-by-step process to go from the paper-based exercise of writing the equations, through the evaluation of degrees of freedom to systematically build the necessary sub-models and main model will be discussed. Although the global mass/energy concepts proposed by Ochsenbein[1] are followed, the drying calculations and driver forces for airflow take a different approach. The presentation will cover the experimental measurements required to obtain the material properties necessary for the model. Additionally, the talk will showcase how Python can be used to parse experimental data and generate the required code to facilitate the integration of data into the simulator.

The presentation will focus on the parameter estimation workflow, as multiple local solutions are typically found in this scenario. The methodology used for estimability[2] will be discussed, which led to the final selection of model parameters. Additionally, the presentation will cover the process of augmenting the model with an empirical model, making it usable with the available measurements. Specifically, a PLS model was built to relate process recipe values to model parameters and some product attributes not accounted for in the physics of the model. I will also show a demo for the use of pyphi to build a PLS model and automatically generate the necessary code to create a hybrid model in the process simulation engine. The presentation will conclude with examples on the use of the model to guide operational decisions and assess risks for tech transfer.

References

1. Ochsenbein, D.R., et al., Industrial application of heat-and mass balance model for fluid-bed granulation for technology transfer and design space exploration. International journal of pharmaceutics: X, 2019. 1: p. 100028.
2. Wu, S., et al., Selection of optimal parameter set using estimability analysis and MSE-based model-selection criterion. International Journal of Advanced Mechatronic Systems, 2011. 3(3): p. 188-197.