FAIR DATA Fund use case: BioBrAIn – a database of bioreactors
Authors: Héctor Maldonado de León, Victor Puig I Laborda, Mahdi Naderibeni and Cees Haringa
Brief description of the project
Despite recent advances, Computational Fluid Dynamics (CFD) simulations of bioreactors remain computationally demanding, especially for assessing full bioprocesses or optimizing design.
Reduced Order Models (ROMs) offer a solution, by reducing the complexity of the model while maintaining spatial detail affordably. Moreover, these models allow for integrating hydrodynamics and biokinetics at an affordable cost.
However, the use of reduced order models in bioreactor analysis is hindered by the steep learning curve of CFD software, time expense of creating a CFD database, and limited access to comprehensive databases essential for training and testing these models.
Our project focuses on openly release a database of well-documented CFD simulations to overcome these barriers, thus facilitating rapid ROM development and enabling researchers to benchmark and compare alternative dimensionality reduction methods.
The database is called BioBrAIn (Bioreactor Analysis and Intelligence), and contains CFD simulations for three reactor geometries and several operating conditions to democratize access to essential bioreactor data. This allows scientists and engineers to analyze bioprocesses more efficiently without requiring access to CFD and high-performance computing, facilitating developments in model and data-driven bioprocess design and optimization.
BioBrAIn is an initiative from the CASE group at the Biotechnology Department of the Faculty of Applied Sciences at TU Delft and colleagues from the Technical University of Denmark (DTU).
Key results:
- An open access database of well-documented CFD simulations of bioreactors
- A proposal for properly documenting CFD simulations in the biotechnology field
How FAIR Data Fund is supporting the project:
The FAIR Data Fund allowed us to hire a student assistant (Ansh Gupta, Chemical Engineering MSc. Student) to create the manual for interpreting and visualizing the data from the various simulations at hand. Furthermore, the 4TU.ResearchData team helped craft the documentation of the datasets and hosted the large datasets to be publicly available.
We are grateful to the team and 4TU.ResearchData for their support in advancing the development of surrogate models for bioreactors by making the data publicly available.
The collection of databases is available on the 4TU.ResearchData repository.
