AI Hybridized Mechanistic Modeling of Chromatography
Sartorius Corporate Research has developed a novel framework for modeling chromatography. The approach combines a first-principle, mechanistic structure with deep neural networks to learn complex binding and diffusion characteristics from practically obtainable training data. The flexibility introduced by the AI components of the modelling framework make it applicable universally to a broad range of chromatographic technologies including membranes, monoliths, and traditional resins.
In this session, we will take a deep dive into the proposed model structure and explore the algorithm used to train the model. Model fitting results will be demonstrated for different datasets. Finally, the utility of the model will be demonstrated for a few key use cases. These include applications such as optimizing recipes for multi-column chromatography systems.
What You Will Learn:
- Understand the benefits of using AI hybridized mechanistic modeling of chromatography
- Learn about the proposed model structure and algorithm used to train the model
- Get insights into potential uses cases of the proposed approach