Machine learning for surrogate process models of bioproduction pathways.

Area of Science:

• Biotechnology and Biofuels
• Chemical Engineering
• Sustainable Energy

Background:

• Technoeconomic analysis (TEA) and life-cycle assessment (LCA) are crucial for biofuel and biochemical production.
• Conventional process simulation tools are computationally intensive and costly, limiting research accessibility.
• There is a need for more efficient modeling approaches in bioprocess development.

Purpose of the Study:

• To evaluate the potential of automated machine learning (AutoML) for developing surrogate models.
• To create accurate approximations of complex process simulation outputs.
• To reduce the computational expense associated with bioprocess modeling.

Main Methods:

• Developed surrogate models using an automated machine learning approach.
• Utilized conventional process simulation models as a basis for surrogate model training.
• Focused on established production pathways for high-value biofuels and bioproducts from biomass.

Main Results:

• Surrogate models accurately approximated cost, mass, and energy balance outputs.
• The AutoML approach significantly reduced computational expense compared to traditional simulations.
• Demonstrated the effectiveness of surrogate models for TEA and LCA applications.

Conclusions:

• Automated machine learning offers a viable and efficient alternative for bioprocess modeling.
• Surrogate models can accelerate research and development in biofuels and biochemicals.
• This approach enhances the reproducibility and accessibility of process economic and environmental assessments.