Opening date: 15/07/2024
Closing date: Until Filled 

Research Area: Scientific Machine Learning for Fluid Mechanics applied to 
Bioreactor Modeling

Starting date: not later than March 2025
Duration: 3 years PhD position 

Funding: The PhD is fully funded by a Horizon Europe project to be started on 
Sept. 2024

Contract:  N3 level according to University of Zaragoza internal rules.

PhD program: Fluid Mechanics at University of Zaragoza (Spain)
Research group: GFN – Numerical Fluid Dynamics Group (UZ) in collaboration with 
the Sustainable Process Institute (UVa).

Eligibility:
• MSc in Engineering, Physics, Mathematics, Computational Science (or equivalent 
MSc degree).
• Provide a short CV including a full list of university grades.

For further information: salvador.izquierdo@unizar.es 

Project background: Bioreactor simulations play a crucial role in developing and 
optimizing bioprocesses. However, conventional simulation methods can become 
computationally expensive for large-scale bioreactors, hindering the scale-up 
process. This computational intensity stems from the multiscale and multiphysics 
characteristics inherent to bioreactors. This PhD thesis proposes the 
development of a framework based on composable Scientific Machine Learning 
(SciML) to accelerate bioreactor simulations. The framework leverages the 
strengths of both physics-based models (based on Computational Fluid Dynamics 
(CFD) simulations) and data-driven machine learning techniques. A hierarchy of 
Multifidelity and Multiscale physics-informed deep learning models will be 
developed to capture the essential dynamics of bioreactors. A flexible, 
composable architecture will be constructed to enable researchers to integrate 
various model components effortlessly for design and scaling analyses. The 
framework will be validated on real-world bioreactor data, demonstrating its 
accuracy and efficiency in predicting bioreactor behavior. Specifically, several 
gas-feed bioreactors will be studied: (i) ammonia biofiltration for nitrate 
production using a nitrifying bacteria biofilm; (ii) biogas fermenter for 
microbial protein production; and (iii) syngas fermenter for acetic acid 
production. This research has the potential to significantly reduce the 
computational cost of bioreactor simulations, facilitating faster scale-up of 
bioprocesses and accelerating advancements in biotechnology.